Intellectualism Refuted (Part III: The Unending Contents of Content)

In a paper entitled "What Makes Perceptual Content Non-conceptual?" (1998), Sean Dorrance Kelly examines Gareth Evans' (1982) arguments for a non-conceptual understanding of perception. Kelly lists these as follows:

"a) that perceptual content is the same for humans and animals,

b) that perceptual content is belief-independent,

c) that perceptual content is, sometimes at least, irreducibly articulated in terms of dispositions by the perceiver to act upon the object being perceived, and

d) that perceptual content is more finely grained than the concepts in terms of which we classify our thoughts."

Before I discuss each of the above (in this and following posts) I'd first like to say a few words about what Kelly, Evans and others term "perceptual content." In my view, to speak of content in this way is to beg the question: content of what form? If the answer is "representational content", as Kelly and Evans make plain, then we already have an unexamined assumption on the table before we have even begun any serious analysis. It is my contention that a thorough analysis of social practices of representation lies at the heart of any principled explanation of perception, and any unfounded assumption about what representations are—and more importantly: how they function—is both premature and potentially obstructive to insight. For this reason, when discussing perception I intend to speak only of perception and to reserve the term "perceptual content" specifically for the discussion of  theories that make explicit use of the term.

In (a) Evans' takes the view that humans and animals share the same "informational states" and thus the same perceptual content. He writes: "The informational states which a subject acquires through perception are non-conceptual, or nonconceptualised. Judgements based upon such states necessarily involve conceptualisation." Although I agree that perception is non-conceptual and that judgements necessarily involve conceptualisation, I find Evans' claim that perceptual content is the same for humans and animals to be lacking in argumentative force. All he provides is a rather flimsy assertion that perception is non-conceptual. Furthermore, as as John McDowell (1994) points out: "the word 'content' plays just the role in Evans' account that is played in that position by the fraudulent word 'conceptual.'" Unlike Evans, McDowell (who edited and published Evans' book after his death in 1980),  conceives of perception as fundamentally conceptual. Nonetheless, as Daniel Hutto (1998) points out, McDowell's commitment to conceptual content, like that of other intellectualists, creates a "heavy burden when it comes to explaining the origin and development of concepts." 

The problem is readily located in the entailments of the content view. If we conceive of brains as containers then it will be necessary to provide both a theory and evidence of what these containers actually contain and how this content is generated, distributed, stored, retrieved and interpreted. The alternative is to conceive of brains as embodied organs that have evolved in response to challenging environmental influences and which have therefore developed highly sophisticated dispositions to respond to a variety of causal encounters. On this view, "content" is simply a convenient but misleading characterisation of these complex dispositions to respond. When someone asks your name, you are disposed to say your name — no inner representation is necessary. Nonetheless, despite the explanatory power of this view, the mainstream position throughout much of the cognitive sciences and philosophy holds that brains are containers and that their content is fundamentally representational. A variety of competing representational theories are proposed but in all cases the basic idea is that brains obviously cannot construct their own 1:1 copy of the world with which to do their work. Instead, it is assumed that brains generate representational states which correspond in informationally significant ways with the encountered world. A minimal version of this approach construes content as a form of "covariation" between internal (brain) and external (world) events — the idea being that as a perceived event occurs, corresponding changes occur in the brain and as the event varies, so too do the corresponding brain states. These states are thus regarded as representational states or, as Mark Cain (2013) puts it: "according to the causal covariation theory, the LOT [Language Of Thought] symbol HORSE means horse because tokens of that symbol are caused by, and only by, horses." Hutto and Myin (2013) are unconvinced: 

"If information is nothing but covariance then it is not any kind of content—at least, it is not content defined, even in part, in terms of truth-bearing properties. The number of a tree's rings can vary with the age of the tree; however, this doesn't entail that the first state of affairs says or conveys anything true about the second, or vice versa. The same goes for states that happen to be inside agents and which reliably correspond with external states of affairs—these too, in and of themselves, don't 'say' or 'mean' anything just in virtue of instantiating covariance relations."

The idea that mental content "means" or is "about" the things it is directed towards is a central article of faith amongst many, if not all, representationalists, but as Kathleen Atkins (1996) points out:

"In an important sense we do not really know what "aboutness" is. Certainly, at the outset, a vague realism about the directedness of mental/neural events is adopted: representations are "tied" to objects and properties and hence (there being no good reason to suppose otherwise) bear some kind of relation to them. But if we do not know exactly what it means to regard a particular as a particular, to see this thing as being of a certain type, this place as the same place, and so on—hence what kinds of capacities or abilities are involved in having representations that are about those things—then we do not know, in any substantive sense, in what that relationship consists. We only trust that it is."

Theorists are right to enquire into the relationships between neural events and the perceived world, but whether those neural events are "about" the world is simply too uncertain to merit the credibility it currently commands, especially if this credibility is simply taken on trust. 

On the issue of whether we know what aboutness is, perhaps we don't in the context of alleged inner representations but what we do know is that in the publicly perceptible world, aboutness is very well understood. Aboutness is the meaning we attribute to symbolic representations of numerous kinds and forms, all of which rely on rules of use for their efficacy. For something to be about something else — for it to signal, refer, indicate, signify, token, connote, or denote, is for it to be generated and interpreted within an already intelligent system. Without such a system or rule to give it meaning, a wink is nothing more than a contraction of the eyelids (Ryle 1968). Without a system or rule, a symbol or signal cannot reliably stand-in (i.e. represent) anything (Bickhard and Turveen 1995). It merely is what it is. 

Moreover, acts of symbolic communication are intentional and herein lies the greatest challenge in proposing a coherent account of mental content. Because if an inner representation is needed to drive an intentional action, then a further intentional representation is needed to initiate this intentional representation and this regress of representations is without end! An alternative naturalistic theory is therefore needed that avoids this regress, not to mention the equally illogical assumption that some form of symbolic language underlies language itself (Fodor 1975, Fodor and Pylyshyn, 1988). As Gilbert Ryle put it to Daniel Dennett in a letter in 1976: "Fodor beats Locke in the intricacy of his 'wires-and-pulleys', when what was chiefly wrong with Locke was the (intermittent) intricacy of his 'wires-and-pulleys'!"

In the next post I will explore how the challenge of intentional action can be met without recourse to inner representations. Such an account will also prove helpful in understanding how we might reasonably attribute intentional action to animals. In so doing it may also be possible to give some substance to Evans' intuition over the common foundations of animal and human perception.

Intellectualism Refuted (part II: Efficacy or Truth?)

"Natural selection does not care about truth; it cares about reproductive success." —Steven Stich, 1990

In his 2011 book "Know How", Jason Stanley argues that to know how to do something is to know a fact that answers the question: "How can I do this?" According to this view, know-how is based on knowledge of truths and to this extent it is a sub-species of knowledge that something is the case, or what philosophers call "propositional knowledge." Propositional knowledge is a somewhat controversial idea because many philosophers assume that animals also possess certain kinds of propositional knowledge despite the fact that many otherwise intelligent creatures have significant difficulty understanding basic abstract ideas let alone truths. Perhaps this is why psychologists prefer to talk of "procedural knowledge" (knowledge how) and "declarative knowledge" (knowledge that) and to reserve the latter for the discussion of strictly linguistic capacities. 

Other philosophers contend that whilst non-human animals may not be capable of propositional skills they do possess beliefs and desires that are not linguistically enabled and that deserve to be considered as "propositional attitudes" (more on this along the way). Stanley makes no claims about non-human capacities in this book so it is uncertain whether he regards his theory as being applicable to our animal cousins. If he does, then he faces a considerable challenge if he wishes to explain how propositional knowledge could emerge prior to know-how.

Stanley describes himself as an "intellectualist" and as such he rejects Gilbert Ryle's arguments against intellectualism put forward in Ryle's 1949 book "The Concept of Mind." The arguments Stanley mounts against Ryle are by no means inconsequential and were Ryle still around he might well have his work cut out defending what until now have been some very serviceable arguments, the most famous of which has become known as "Ryle's regress." Basically Ryle's idea is that if intellectualism is true, then every intelligent act would need to be preceded by an intelligent thought and this thought itself would have to be initiated by another intelligent thought and so on ad infinitum. Ryle uses this regress to refute what he calls the "doctrine of intellectualism" and to argue for the logical possibility of unpremeditated acts of intelligent know-how. Stanley argues that if intelligent acts of know-how are possible in the absence of prior acts of intellection then so are acts of intellection themselves and consequently Ryle's regress fails to convince. Stanley explores several of Ryle's further variations on the same basic thesis, all of which he contends, similarly fail in their refutation of intellectualism.

So what should the anti-intellectualist do in the face of such counter arguments? Is Ryle's regress the only defence we have against the claim that intelligent action is conceptual to the core? If Stanley is correct, Ryle's arguments represent the only serious challenge to intellectualism — unless Stanley is picking a conveniently absent adversary in order to create the impression that he has anti-intellectualism on the run. In 1982 Gareth Evans' book "Varieties of Reference" made several frequently cited arguments against intellectualism, none of which Stanley mentions, even though he quotes Evans widely on other matters.

Stanley's position relies on the supposition that to know how to do something intelligently amounts to knowing a truth. My contention (pace Stich 1990) is that where efficacious action is concerned, truth is, at best, of secondary importance. If being able to swim is knowing a truth, as Stanley claims, then what purpose does this truth actually serve? The concept of truth is irrelevant when compared with the competences from which the ability is comprised. Stanley would presumably contend that each atom of competence is also an atom of truth. But if knowing how to do things like swimming were simultaneously the acquisition of constituent truths, then teaching by description would be the only educational method necessary. Demonstration and practice would have nothing further to impart. However, skill in describing a procedure does not bestow a skill in the performance of the procedure described. A knowledge of linguistics does not make an orator. And a witness who provides a clear description for a identikit image need have no expertise in drawing.

In a response to a related 2001 paper by Stanley and Williamson, Ellen Fridland (forthcoming) writes: "knowledge of rules and facts is not identical to the practical knowledge of how to put those facts and rules into practice effectively." In a footnote she continues:

Of course, Stanley and Williamson would agree with this. They would say that the knowledge has to be represented in a particular way.  My claim is that such a way of representing would itself necessarily be a kind of (nonpropositional) knowledge.

Indeed, if the knowledge must be represented in a particular nonpropositional way then this representation cannot be symbolic—on pain of circularity. If we require a non propositional representation, as Fridland suggests Stanley and Williamson would concede, then the choice will necessarily be limited to one or other of the following: a demonstration, performance, exemplification, illustration, image, diagram etc. None of these is easily  translated from even the most rudimentary propositional truth because abstract concepts simply do not lend themselves to nonverbal representation.

Truths are typically statements or stories. A true statement is a statement of truth but a true likeness is not a likeness of truth.  Truths are linguistic items that we take  to represent objects and events objectively. But if it could be shown that a widely employed and extremely useful class of nonverbal representations are never objective then the intellectualist's position would be in need of major revision. I think such a counter-example is ready-to-hand in the commonplace practice of representational simulation (Brook 1969, 1997, 2014). In order to avoid the incredulity of theorists who do not share my convictions concerning the perceiver-dependence of simulating representation, rather than arguing this line directly I will instead take a more indirect route via the more widely acknowledged arguments raised by Evans (1982). I intend to show that in each case Evans' arguments against intellectualism are best understood in the context of skills of non-verbal representation (including simulation) production and use, a fuller understanding of which will challenge not only Stanley's view but intellectualism per se.

Intellectualism Refuted (Part I)

"Let us not forget this: when 'I raise my arm', my arm goes up. And the problem arises: what is left over if I subtract the fact that my arm goes up from the fact that I raise my arm?" —Ludwig Wittgenstein

In the next several posts I intend to explore the philosophical doctrine of intellectualism and to contrast this with a principled alternative that avoids many of the shortcomings that accompany the intellectualist view. Intellectualism takes many forms, but in each case the foundational assumption holds that intelligent action, in all its varied manifestations, is the result of one or more forms of conceptual thought, judgement, computation, reason, ratiocination or intellection. So, when animals act intelligently on the basis of anticipated future states of affairs, these actions are taken by intellectualists to be the product of logical operations conducted at some level in the brain.

Brains are extremely complex organs and neuroscience, being a relatively youthful field of enquiry, hasn't yet discovered any clear and unequivocal evidence of these postulated logical processes. It would probably be fair to say then, that much intellectualist thinking is simply the consequence of widespread uncertainty about what kinds of mental states or processes might legitimately qualify as the inner correlates of the many sophisticated things that animals (including human animals) are capable of doing and is therefore simply a convenient way of saying that brains process inputs that lead eventually to behavioural outputs of one kind or another. Likewise, when people speak of "perceptual judgements," very often, all that is meant is that perception is a skilful capacity to discriminate between similar objects or attributes etc. If intellectualism were merely a matter of convenience of this kind, then there really wouldn't be anything worth debating. As it is though, the assumptions and commitments of intellectualism are far reaching, not only for perceptual theory but for a wide range of fields throughout the cognitive sciences, philosophy and potentially medicine also.

If the stakes are this high, then it obviously matters a great deal that theorists get it right about whether brains compute information relayed via neurological data representations or whether, as a growing minority of theorists and researchers argue, this explanatory route is untenable. Theorists like Dan Hutto and Eric Myin maintain that brains are best understood as embodied, embedded and enactive organs of response and intentional action that do not—indeed cannot—produce their own inner representations. The vicious logical regresses that emerge as soon as such intention-directing representations are invoked lead to the unavoidable conclusion that such representations must be an impossibility. Theories that reject representationalism therefore deserve to be taken very seriously. So what is needed in order to propel anti representationalism forward is a coherent explanation of how intelligent predictive behaviours could emerge and evolve whilst at the same time avoiding any of the errors of intellectualism. As I progress through this discussion I will explore how predictive behaviours can be explained in the absence of the conceptual powers that are such a indispensable part of the intellectualist's story.

My views on the subject of intellectualism are motivated by the conviction that conceptual skills have arrived on the evolutionary scene only relatively recently with the emergence of language and that these skills are supplementary to other vitally important non-conceptual skills that are widely misunderstood and overlooked. Like language, these skills are representational but unlike language they do not permit the contemplation of abstractions of the kind that are a central feature of verbal communication. It is my view that these non-conceptual skills pre date our linguistic capacities and continue to be skills that we  regularly deploy in our intentional interactions with the world. When we struggle to find words to express our intentions, it is often precisely these skills that we are labouring to translate into linguistic form, not another equally semantic but indirectly accessible "Language of Thought" or "Mentalese" as is hypothesised by many intellectualist thinkers.

Despite its intuitive appeal, the Language of Thought hypothesis is far from an established fact about cognitive processing. What is beyond dispute though, is that organisms possess both genetically inherited as well as environmentally conditioned predispositions to respond to their causal encounters with the world. To have learned something is therefore to develop a disposition to behave in a particular way. Very often the best way of demonstrating that you have learnt how to do something is simply by doing it, not by describing it. A verbal description of a procedure, no matter how finely nuanced, is no proof of competence. For this reason I take our skills of performance and exemplification to be both logically as well as evolutionarily prior to our skills of description, but in each case it should be noted that a dispositional theory of representational capacities is fundamental to an understanding of all intelligent action. On this basis I also take the view that the capacity to envisage future states of affairs is by no means a fundamentally conceptual aptitude. If we wish to understand how animals are capable of anticipating the immediate future—of how predators are often remarkably skilful in predicting the behaviour of their prey—then there are really only two explanatory routes available to us: conceptual or non-conceptual. I intend to show that the conceptual route is both logically and evolutionarily infeasible.

What’s So Wrong With The Selfie?

Paleolithic handprints, Argentina (Courtesy of Wikimedia Commons)

With the help of devices like smartphones and tablet computers it has never been easier to produce photographs, and with the rise in social media it has never been easier to share them. But is this availability of media a neutral influence? Does it empower or corrupt? Does it enable us to express ourselves in new ways and to discover new things about the world or does it cultivate narcissism of a previously unknown kind or degree?

It is certainly no bad thing that the means to produce depictive representations is no longer the exclusive preserve of a privileged few. But with the wider availability of media has come an unprecedented proliferation of representations clamouring for our consideration. Of course there are many ways to skim, filter or ignore this information, so there is certainly no requirement that we treat each call to our attention with equal regard. In fact, to bemoan the proliferation of content is perhaps an admission of a lack of selectivity. Nonetheless,  distractions can sometimes be an important source of unexpected discovery — why else would we be so susceptible to them in the first place if they were only ever a disadvantage?

But perhaps the problems lie elsewhere — not with the quantity or kind of information on offer but with its quality. A good example of this problem is to be found in the cultural phenomenon of the "selfie"? Selfies are what would once have been called "self portraits", but the shift in terminology is telling. Dropping the word "portrait" and appending the suffix "ie" not only trivialises the practice, it infantilises it. It takes a term most closely associated with artworks and replaces it with the language of the nursery. This alone wouldn't be so bad — defenders of the selfie might well argue that the attraction of the term is precisely its lack of pretension and the break it makes with the pomp of the artworld. The desire is as understandable as it is naive.

Much as depictions are capable of prompting sophisticated thoughts and associations, they are not fundamentally conceptual objects. Pictorial images do not rely on our skills in the use of abstractions for their efficacy, which is presumably why infants have no difficulty recognising them. Self-hood, on the other hand, is an abstract concept — it does not lend itself to being recorded on film or sensor. Self-hood can only be represented by allusion, analogy, reference, or signification — all of which are fundamentally symbolic strategies for the communication of meaning. In other words, a picture of someone can only be about them if it makes use of one or more of these codes, associations, references etc. and these are inextricably enabled and informed by culture. A picture of a person is not intrinsically about them. So a picture can never be of the self — which is why portraits need not actually depict the person portrayed. Such is the power of allusion.

So where do these thoughts bring us as regards selfies? For the most part, selfies seek to be light-hearted, uncomplicated depictions with as few pretensions to the world of portraiture as possible. Selfies rarely, if ever, seek to articulate meanings. They seek to record the presence of their makers and to share simple pleasures and conviviality of a playful kind. To this extent selfies are the most naive form of depiction. 

Whilst selfies seek to rid themselves of artworld references they are nonetheless steeped in unintended meanings. Culture not only enables artists to articulate and allude to meanings but it also enables everyone to 'read' images; to interpret them via the many codes and clues of which culture is formed. Selfies expose all kinds of subtle, and sometimes not-so-subtle, information about the people depicted, from their sexual orientation to their social class, from their political views to their emotional state and sometimes even the state of their relationships. Even the quantity of images produced by some individuals can be strongly suggestive about the maker's need for attention and approval. And whilst our interpretations may not always be accurate we are nonetheless unavoidably drawn to treat these images as information, as indicators about the drives, intentions and preoccupations of others.

In their infantilising rejection of content, the champions of the selfie have lost touch with the very thing that gives portraits meaning. Ironically, selfies are almost exclusively focused on appearances and have practically nothing interesting to say about the self.