“Radicalizing Enactivism: Basic Minds Without Content” A Scaffolded Review

When people talk about the possibility of foreknowledge of the future they always forget the fact of the prediction of one's own voluntary movements. —Ludwig Wittgenstein.

Very few philosophers will be persuaded by this important book by Dan Hutto and Erik Myin, most obviously because it undermines what for many will be held as irrefutable: the idea that minds are necessarily dependent upon inner content. By content is meant representations of one form or another: data structures, signals etc. that are supposedly transduced, encoded, interpreted, recombined, judged, parsed or otherwise processed according to information-processing models that despite decades of fine-grained EEG, MRI, fMRI, DTI and PET research remain entirely undiscovered in the Byzantine labyrinths of the brain.

For Hutto and Myin the assumption that cognition involves content is thoroughly unjustified in all but the most ‘scaffolded’ of cases: cases in which minds are supported by socially evolved practices of linguistic representation-use. In the case of more basic minds, they argue, content is neither necessary nor logically possible.

In a philosophical landscape dominated by representationalist thinking, all of which is perilously dependent upon the assumption that cognition involves content, this book fulfils its radical intention of exposing the many ways in which the content view runs thoroughly into the sand.

There isn’t room here to detail the many incisive arguments levelled against the content view by Hutto and Myin, but suffice it to say that the majority of the book is devoted to a point-by-point critique of the principal content-dependent theories, including several that share much in common with the authors’ brand of enactivism. No doubt critics will complain that the authors overemphasise their counterarguments at the expense of a comprehensive alternative thesis. Yet, if this book were nothing more than a thorough disavowal of the doctrine of content, its contribution to the field would be significant. As it is, Hutto and Myin have, at the very least, cleared the ground and provided some vital tools for further exploration and this is both welcome and, some would say, long overdue.*

Radical as this book is though, I suspect that Hutto and Myin are either withholding a more uncompromising version of their argument or else their position falls somewhat short of full blown radicalism. Their retention of content in the case of scaffolded minds in particular concedes more than is strictly necessary to their adversaries. Perhaps they wish to offer an olive branch to soft representationalists in the hope that they might be turned. But the fact that minds can be supported by an extensive range of performative practices, procedures and artefacts is no reason to suppose that scaffold users have somehow evolved into inner content bearers. The mere suggestion that this is the case does damage to what is otherwise a strong argument against the content view. 

Undoubtedly brain states have a functional correspondence to behaviours ('covariance conditions' as the jargon goes), but this is no reason to assume that these brain states literally constitute representations, no matter what the stage of their evolutionary development. This is the conclusion that we should draw from radical enactivism and it is strange therefore that Hutto and Myin are in the least credulous of contentful representation, even in the case of scaffolded minds. Scaffolds are external resources after all and the role they play is not one in which inner states become surrogate scaffolds.

The capacity to think using contentful representations is an example of a late-developing, scaffolded, and socially supported achievement. It originates from and exists, in part, in virtue of social practices that make use of external public resources, such as pen, paper, signs, and symbols.

It may be the case that Hutto and Myin are simply being economical here with their list of publicly available resources. No doubt they wish to establish a clear distinction between the resources and social practices utilised by language users and those used by other creatures. However, in doing so, they suggest an evolutionary break where none need be imputed. So, whilst I agree that social practices are centrally implicated in the emergence of mind, I am not at all convinced that only late-developing symbolic capacities qualify for consideration as constitutive of thought — conceptual thought, yes, but not all thought. So scaffolded thought need not - indeed should not - be regarded as exclusively linguistic in origin. Even language had to evolve from more basic roots. There are many immediately available behavioural (enactive) resources and communicative capacities that creatures commonly employ that deserve serious consideration as sufficient for the emergence of mind long prior to the discovery of the various processes and procedures of symbol manipulation and language. Gestures, dance, facial expressions, mock performances—including play, vocalisations, mimicry, exaggeration, deceit, distraction, diversion, concealment, shamming, and numerous other patterns and kinds of intelligent behaviour are all important contenders for the constitution of basic thought, or at the very least intentional directedness.

Intentional directedness - as it turns out - is a major challenge, not only for Hutto and Myin’s theory but for all principled non-representationalist theories. Representationalist thinkers, in contrast, can simply invoke inner representations of future states of affairs and claim these as the causal basis of intentional action. This is a very convenient theoretical strategy, but if the mounting arguments against representationalism — of which Hutto and Myin's should be regarded as canonical — are correct, then the representationalists ploy is recklessly misconceived.

What is it then, we might ask of the radical enactivist, for an eagle to act intentionally and to anticipate future events before they unfold? The challenge here is to explain how 'nonverbals' are capable of predicting future states of affairs with a high degree of accuracy. How are they capable of tracking moving objects when obscured? Can we credit such creatures with the ability to envision future circumstances and if not, what is it about human skills that lead us to suppose that only we are possessed of the ability to visualise — certainly not our verbal skills?

Hutto and Myin evidently discern little in the way of a challenge arising as a consequence their otherwise well argued and justified rejection of inner representations. They write: ‘The simplest life forms are capable of intentionally directed responding.’ but precisely what predictive causal influences are involved here they give only the merest suggestion when they offer: ‘informationally sensitive responses to natural signs.’ They claim that this is 'austere talk' and that it avoids the assumption of 'meaning' and 'representation' that they find frequently betrayed in the work of Thomson (2007) for example. I’m not at all sure though that the policy of offloading the attribution of content onto external ‘signs’ and ‘information’ is anywhere near as  austere as is required: advantageously sensitive responses to natural stimuli would be genuinely austere but would still leave the question of intentional directedness completely untouched.

A more convincing route to resolving this question is already available to Hutto and Myin, yet it is so thoroughly partitioned off in their theory that it seems unlikely that they would be willing to reconfigure their thinking to accommodate it. The scaffolding that they claim is only involved in late developing achievements of thinking is, as we have already seen, more extensive than their thesis currently allows. If this is true, and the evidence for these capacities is widespread amongst social organisms, then we already have good reason to believe that these enactive capacities themselves might be significantly implicated in intentional directedness. I will return to this point presently.

In an earlier publication, Hutto (2008) discusses the question of intentionality directly and he argues that the 'dances' guiding the behaviour of honey bees are 'contentless' and therefore non-representational. He claims that these dances are 'Local Indexical Guides' and that the bees are 'informationally sensitive to natural signs.' His preferred terminology is presumably intended to be as neutral as possible regarding it's representational implications yet he could hardly have chosen more prototypical cases of representation than indexical guidance or informational signification. He also states that bee dances incorporate: 'two distinct aspects: one carries information about the distance of the nectar from the hive and the other carries information about the direction in which it is located.' [My emphasis] This is curious because just few pages earlier he writes: 'It is easy to be misled on this score by free and easy talk of information "being carried" by signals and states.'

Either creatures communicate with one another or they don't, and if they do, then the only resources available to mediate this communication are public representations. So, for example, indexes are representations in which one thing is used to direct attention to another thing: a pointing finger [index] or footprint are obvious instances. But, as the saying goes: 'The finger pointing at the moon is not the moon.' Indexes indicate other things and other locations, but bees do not literally stand outside the hive pointing in the direction of the relevant flora. If I see a flower and point to it, there is no doubt (at least for other humans skilled in representational practices) that I indicate the flower. But if I point to the flower through the solid wall of a hive, the indexical relationship suddenly becomes significantly less obvious. So, if bees do use indexical guides, they cannot be of the direct indicative kind — even intelligent animals like dogs or chimpanzees have trouble responding appropriately to a pointing finger (Tomasello 2006). Similarly, if we examine the concept of guidance we will find that this too turns out to be a paradigm case of representation with more complexity in its simplest forms than Hutto seems prepared to acknowledge. Guidance is a performative form of exemplification in which we physically enact the relevant route or procedure: we show the way. Deliberate guidance of this kind is actually fairly rare in nature, whereas non-physically enacted guides (like maps for instance) are very rare indeed (an intentionally produced  trail would probably qualify though). Bees do not leave trails lingering on the air for one another, nor do they personally guide one another to their floral caches. So if we wish to illuminate what appears to be an extraordinarily economical yet sophisticated form of public representation exemplified in the display behaviour of bees, then we have no alternative than to explore their socially evolved capacities for representation production**. No inner representations need be imputed — no knowing that, just embedded, embodied and enactive know-how.

I suspect that Hutto's denial of the representational behaviour of bees is motivated by his eagerness to dispense with teleosemantic accounts of cognition — accounts that characterise mental states in linguistic terms. Whilst I am entirely sympathetic with his ends, I think his means are gained at the cost of an appropriately nuanced understanding of the relationships and differences between nonverbal and fully verbal representational practices. Nobody doubts that bees have the capacity to respond to their successful foraging trips by behaving in ways that lead other bees to forage similarly. What is at issue is whether the information (direction and distance) is internalised in representational form or whether there might be a less extravagant way to conceive of it. Hutto and Myin are absolutely right to pursue this latter line of enquiry, yet Hutto's ‘Indexical Guides’ seem unlikely to withstand the occamists razor when compared with the intuitive, if utterly mistaken, conceptual simplicity of inner representations. A more substantial non-representational theory of intentional directedness is urgently required.

How might this be achieved? I suggest once again that an adjusted version of Hutto and Myin's Scaffolded Minds Hypothesis is all that is necessary. Where a sharp distinction needs to be drawn though, is between organisms that are capable of producing representations (like bees and human beings) and those that are not (like viruses and trees). Representation-producing organisms provide good cause to suspect that they might be capable of anticipating some future states of affairs, i.e. of representing them in token form – as is the case with bees. Such capacities would qualify therefore as viable causal drivers towards currently unfulfilled future states of affairs. In the case of other organisms, we haven’t yet embarked upon a study of what causally influential dispositions-to-represent may be mediating their actions but this would seem to be a field rich with untapped potential. What we can say though, is that the capacity to produce representations of future states of affairs should be considered instrumental in the attribution of intentional action as opposed to mere purposeful responsiveness.

Although I have only sketched the vaguest outline here of what is a far more extensive enactive theory of intentional directedness, it is nonetheless closely consistent with the main body of Hutto and Myin’s important theoretical work. Minds are by their very nature scaffolded and without such scaffolding, mindedness would be inconceivable.

In an article for the Notre Dame Philosophical Review, Tom Roberts claims that Hutto and Myin’s most radical idea is that “basic minds are not brain-bound; they are not defined by representational transactions; they are fully and constitutively world-involving.” I would challenge Roberts on this observation. The book is called 'Radicalizing Enactivism' after all, not 'Radical Enactivism'. It is a rallying cry, not a stipulation of terms; a set of tools, not an academic ornament. It's radicalism derives from the advantages it confers to those who put it to use: enacting its radicalism.

Basic minds – indeed all minds - are constituted by what they are capable of representing: of precisely the public representational transactions they are capable of engaging in. Beyond these capacities to represent their causal engagements with the world, all organisms – ourselves included – are merely evolved purposeful responders. It just so happens that humans are massively disposed to represent their causal encounters. And what we are not capable of representing we can’t claim to perceive. The same applies, it might be said, to enormously important but largely ignored or misunderstood radical theories of mind.


For further discussion of Dan Hutto's theories you may be interested to read the next post  here.

*William Ramsey’s ‘Representation Reconsidered’ (2009) is another important contribution in this regard.
** The question of the representational nature of bee dances remains controversial. Adrian Wenner in particular is critical of the evidence provided regarding bee "language". His own research favours scent carried by successful foraging bees which then triggers fellow bees to search for the same scent. Nonetheless there remains reason to suppose that a simple form of representation may be at work - at least in the direction that the dances are conducted which correlates with the direction of food sources. More evidence is required to settle the issue.

Philosophical Behaviourism - An Incomplete Project

Our earlier history, the part without writing or artifacts, can only be deduced from what we do and what our primate relatives do – the hieroglyphics of the chimpanzee’s whimper and the baby’s smile. -Alison Jolly (1937-2014)

During the heyday of psychological behaviourism a partially related school of philosophy emerged in Oxford, inspired by the later ideas of Austrian-British philosopher Ludwig Wittgenstein. Philosophical behaviourism (also known as Analytical or Logical behaviourism) had no concern for the scientific study of stimuli and conditioned responses but instead sought to explain mental concepts through an analysis of the ways we use language: through linguistic behaviour. Although this emphasis on language-use was a central preoccupation amongst these 'Ordinary Language' philosophers, the implications for a broader understanding of the relation between skilled action and mind are not difficult to detect. Gilbert Ryle especially, sought to clarify that our 'inner world' - and all of the dealings we are inclined to say that we conduct there - is a manifestation of our various interactions with the physical world itself, not some occult realm that somehow brings the world to light for us. So, to extend Ryle's argument a little: the mind does not endow us with capacities but rather is a consequence of them, of actions we are capable of performing.

It might be worthwhile examining why Ryle never made this last point explicitly. Most obviously, if he had, he would have faced an immediate difficulty. If it is true that the mind is a consequence of our physical capacities, then what possible skill could correspond to - or rather constitute - our imagining a colour? Perhaps our capacity to name a colour could be sufficient? But if this is the case then consciousness is only possible if we are first capable of language. Surprisingly, many theorists persist in directly, indirectly and inadvertently arguing that language is indeed essential to consciousness (though Ryle was not among these 'intellectualist' thinkers). A more plausible contender then - as a skill constitutive of mind - might be thought to be embodied in the capacity for sensory discrimination. However, this explanatory route turns out to be extravagant in precisely the same way that the language-only route seems overly restrictive. If all that is required for the emergence of mind is a capacity to discriminate one thing from another, then we have no alternative than to count even some of the most primitive organisms as minded - if only in the most elementary ways.

It seems likely that Ryle was aware of these difficulties, yet I suspect that his (and others') overriding focus upon language may also have contributed to the lack of exploration of alternative explanations in this area. Going by the majority of philosophical theories developed and pursued during the 20th Century, you could easily be forgiven for thinking that the only field of interest worthy of serious investigation was to be found amongst the many sayings, signs, semes, signifiers and syntactical structures of language. And whilst the 'linguistic turn' in philosophy continues to exert a major influence upon a wide variety of fields of enquiry throughout the world, comparatively little attention has been paid to skills of nonverbal communication. Even with the recent upsurge of interest in theories of Embodied Cognition, there remains a general disregard for the evolutionary emergence of nonverbal skills of representation. 

Yet, if we fail to concern ourselves with the evolution of these rudimentary capacities we leave an unassailable mountain for anyone seeking to explain the evolutionary emergence of language. In spite of an absence of evidence and a dearth of reasoned argument, it is currently almost universally assumed that the symbolic practices comprising language sprang into existence unaided by other more rudimentary forms of communicative behaviour. 

With a more expansive view of representational practices than is currently on offer, it becomes significantly more feasible to discern the embryonic origins of language and even to formulate a speculative solution to Ryle's dilemma. If we propose capacities of nonverbal representation as significantly constitutive of mind, we avoid the excesses both of overgeneralisation and of restrictive limitation and situate the attribution of mind at a more plausible stage between basic organismic behaviour and the highly sophisticated symbol manipulations of which language users are capable. 

Innumerable animals are capable of basic forms of mimicry and countless creatures communicate through the use of sounds, gestures, scents, colour changes etc. We need no persuasion that our linguistic skills must have had more rudimentary origins. Symbolic practices must have, at the very least, evolved alongside non-symbolic skills, if not as a more sophisticated consequence of them.

The shift is one from a focus on “things”, such as representations, to a concern with “activities”, such as the act of representing. [...] The task is to understand a variety of representational practices, and wherein they are representational. The means we employ in doing so will be various: historical analysis of their emergence, sociological analysis of the conditions under which they operate, experimental psychological analysis of representational gaps and gluts, anthropological analysis of practices of symbolization, evolutionary analysis of social environments and our sensitivity to them. (Robert A. Wilson 2010)

Philosophers have barely even scratched the surface of these enquiries into the nature, variety and extent of representational practices, of behaviours and acts of non-verbal representing. But what should be clear is that animals do indeed have minds - if only simple minds - because mind is not dependent upon language but upon representational capacities that are functionally and procedurally distinct from language and that must have evolved long before these skills of symbol manipulation of which we are so fond but which so often cloud our understanding of other creatures and ourselves.

Negative Causation: a brief refutation

For a handful of philosophers the notion of absence gives rise to concerns over the nature of causation. Jonathan Schaffer of Rutgers University, for instance, argues that 'negative causation', as he calls it, is a 'genuine form of causation' in which 'causes need not be physically connected to their effects.'

At first blush, the concept of negative causation seems to fit the facts quite well. Pull a ladder from under someone and the absence of support precipitates a fall. Sever someone's wages and the absence of income will likely cause significant disruption. What is blindness other than an absence of vision, or ignorance other than an absence of knowledge?

Whenever we discuss causation we tend to talk in terms of influence and the exchange of forces. There is nothing contentious about this. Schaffer provides a long list of instances where the removal of some impediment triggers a consequence or chain of events and he argues that in each case we have indisputable evidence of causation through absence. He gathers together many plausible examples from a broad range of fields, yet one crucial point goes unmentioned. If causation arises in a negative form then it should, like its positive sibling, come in varying degrees or states. Photographic negatives consist of tonal gradations from transparent to opaque; negative numbers mirror their positive counterparts and even the theory of negative energy is not limited to a single state.

Another helpful way to appreciate the conceptual error is to recast negative causation as zero causation. Interestingly, as soon as we do so, it becomes obvious just where the fault lies: zero causation is zero influence. 

We don't need a theory of negative causation. Absence works perfectly well to designate the lack of causal influence. And the only reason we might be tempted to say that absence exerts influence is because, in the absence of influence, other forces come into play that would otherwise be held in check.

Photographic Absence: a fuss about nothing

'There is no such thing as an empty space or an empty time. There is always something to see, something to hear. In fact, try as we may to make a silence, we cannot.' ― John Cage

Philosopher Mikael Pettersson, explores the causal status of shadows and absences in photography. Despite the apparent absurdity of the issues he raises, the implications are profound.

For a conference presentation at the Institute For Advanced Study at Durham University, Pettersson began by reference to the widely held view that photography is a prototypical causal medium wholly dependent upon the known regularities of the universe for its efficacy. If it so happened that photography were an unreliable medium, if every exposure were a struggle with uncertainty and every image were a tapestry of the indeterminate, there would be little value in these objects that so conveniently find their way into our lives. Indeed, if photography lacked causality, it seems unlikely that vision itself would be possible.

But if photography is a causal medium, Pettersson asks, how can non-entities like shadows and other kinds of absence commonly depicted in photographs have causal influence and, moreover, what is their causal foundation?

It is tempting at this stage to throw our hands in the air and to declare the preposterousness of the enquiry. Absences, and this necessarily includes the complete absence of illumination, lack causal influence because there is literally nothing to do any causing. The box of photographic paper that sits on my desk, awaiting exposure to light, gains it's value by virtue of its complete and unalterable insensitivity to the absence of light. We discern shadows in images due to the causal presence of light and its varying intensities precisely because the absence of light has no influence whatsoever. Photographic traces that depict shadows are the result of electrochemical processes designed to replace unexposed portions of the image with chemicals that absorb light, thus simulating darkness.

But to pursue this line of response to Pettersson's enquiries would be to overlook one of the most important enigmas of human, and to an indeterminate degree also creaturely, concern. Possibly the most profound event in any life is the encounter with the death of another. Absences cause—or at least seem to cause—some of the most intense experiences we are ever likely to face. So the question over the possible causal influence of absence is by no means a trivial one. It is arguably the most significant question of all.

Another conference delegate, Vivian Mizrahi, questioned Pettersson's emphasis on photography. Why, she asked, does the same not apply to other forms of image making? For Pettersson, the widely accepted causal directness of photography, the fact that it is not mediated by the brain—as is the case with drawing or painting, seems to provide stronger grounds for attributing a causal role to absence rather than any mind-dependent intermediary. Nonetheless, minds do play an inescapable role in the reception and interpretation (or 'seeing-in') of images. This was the concluding point to which Pettersson turned his attention, though not without giving due consideration to several of the more plausible but ultimately unsatisfactory acausal theories of influence in absentia.

Is there any evidence of the causal influence of absence? It might be argued that the absence of food in our stomachs causes us to seek sustenance; or that a lack of time causes lateness; or that a hole on a beach causes sand to infill the void; or that the absence of aerodynamic lift causes an airplane to fall from the sky. All of these instances, and many more that we could enumerate, are seemingly plausible examples of the causal influence of absence. However, on inspection, many turn out to be cases of commonplace causality. A decrease in blood glucose, triggers hunger; lateness is caused by an over abundance of things to be done and falling is caused by the release of potential energy. 

However, the last example mentioned above is far from clear-cut. Some theorists would claim that the release of potential energy is due to an absence of physical support. There seems to be growing theoretical uncertainty over whether such instances are the result of an oversimplification in description or whether there genuinely are cases of negative causation. Personally, I'm not sure - though I suspect that the fault is linguistic in origin. If exhaustive detail were necessary to describe the causal history of even the most basic state of affairs, the chain of causes would lead inexorably to the Big Bang. Abstraction in description is therefore both necessary and inevitable. At some point even the most exhaustive account falls silent. The concept of absence then, like the concept of zero, is a tool that has proven to have enormous efficacy for us. Yet unlike the concept of zero, which is only ever acquired through education, there is evidence to suggest that human infants display an awareness of absence from a very early age and many animals also behave in ways that suggest a capacity to register absence.

Do these glimmerings constitute the rudimentary evidence of conceptual capacities, of proto-linguistic thought? Perhaps, but we should be wary. If we wish to provide a coherent causal theory of absence we will first need to explain the causal relations that lead to attributions of absence.

I have written previously of another philosopher, Anya Farennikova, who takes the view that we literally perceive absence. Farennikova contends that we possess mental states in which absences are represented. There is a serious problem with this theory though, and it is a difficulty reflected in the issues Pettersson brings to light. To assume that an inner representation represents absence requires this absence to have causal influence, thus violating the laws of causality.

The alternative is to view absence not as a causal entity of any kind but as the name we give to the common mismatch we find between what we expect and what we actually perceive. So, when we say that an infant or animal exhibits awareness of absence, what we really mean is that they are capable of forming expectations and of being surprised when these  do not apply.

Absence is an enormously powerful and convenient concept, but the fact that we can treat such abstractions as concrete entities is no cause to invoke causality where none is possible. As far as images go, even the most explicit absence can only be recognised as such if we are capable of a contrasting expectation: of describing, delineating or selecting a substitute of that which is absent, of the actual presence envisaged.

Or, as Carl Sagan said (admittedly in a completely different context): "Absence of evidence is not evidence of absence."

Triggers, Signals and Intentionality

I was recently recommended to read Ruth Millikan's work on biological intentionality. Intentionality is widely thought to be the result of anticipatory capacities of one sort or another, but quite how these capacities manifest themselves on a neurological level is a fiercely debated subject upon which Millikan takes a radical but I think entirely mistaken line. Millikan takes the view, not only that brain states are representational, but that they are essentially semantic: ‘Biosemantic.’

In the next few paragraphs I aim to show why biosemantics fails to achieve its goal of naturalising representational content, and I aim to do this by way of a discussion of the evolutionary emergence of the most basic form of signalling commonly observed in nature (i.e. rabbits thumping the ground, the tail splashing of beavers and numerous other rudimentary alarm signals).

Organisms need to be sensitive to changes in their environment and to be differentially responsive to available stimuli, i.e. capable of adjusting their  responsiveness according to prevailing circumstances. If a certain advantageous or disadvantageous circumstance is commonly preceded by other regularly occurring stimuli, then it is of significant advantage for organisms to be capable of responding effectively to these antecedent environmental triggers.

Environmental triggers, of whatever kind, should never be confused with signals. The scent that leads an organism to a source of food is not a literal signal propagated by the food, nonetheless it is a potentially detectible property in the environment propagated by the food. Strictly speaking, a passing shadow or rustling in the undergrowth etc. are not signals, signs or indicators of impending danger, because they are not deliberately produced, i.e. they are not intentional. They are simply detectable characteristics of the environment that occasionally precede threatening events. As we will see, this fact about the necessary intentionality of signalling is one of the principal weaknesses in mainstream theories of intentionality like Millikan’s. If intentionality is dependent upon representations and representations are dependent upon intention then there is no way to break into the circle.

So, what do we know about the emergence of signalling in a biological context? Or, more to the point, how can a publicly perceptible behavioural trigger (a startle response, say) evolve into an intentional signalling behaviour of the rabbit ground-thumping variety?

In order to answer this question we first need to recognise that vulnerable organisms benefit  from social coexistence because this provides a safer environment in which the probability of attack is greatly reduced. Additionally, when any one individual is attacked, the ensuing commotion has the potential to trigger evasive responses on the part of neighbouring individuals. Any individuals failing to detect and respond to such disturbances will be vulnerable to further attacks and will consequently be less likely to survive.

In order for a behavioural trigger to become a signal then, the following conditions need to be met:

1. A group of organisms must be under selective pressure.
2. These organisms must behave in regular and conspicuous ways when attacked, thereby producing a stimulus with the potential to be used as a signal.
3. Consumers must be capable of detecting the stimulus. 
4. Behaviours triggered by the signal must be advantageous to both producer and consumers in the majority of instances*. 
5. Producers must become capable of producing the signal independent of its standard causes if the signal is to have efficacy over and above that provided by ordinary behaviour.

As can be seen, the steps necessary for even the most basic form of signaling are complex and demand very different circumstances than those that pertain amongst the cell structures of the brain. Quite how an analogous form of signaling, of the kind that Millikan and others impute, could evolve through the interaction of brain cells awaits even the most basic demonstration.

What needs no demonstration though, is the fact that many creatures produce publicly perceptible signals. All that needs to be recognised is that it is the capacity to produce such representations that is instrumental in cases of intentionality, not some alleged but so far entirely undetectable representations flitting around inside our heads.

*Individuals may benefit by “crying wolf” in certain circumstances but the efficacy of this strategy will be limited in the long term. Similarly, withholding a signal may have short term advantages and would explain why many animals respond to alarm calls by increased attention instead of shelter-seeking behaviour.

Realism and Reality

Despite the fact that more distant objects are projected at a smaller scale onto the retina than closer objects, it is crucial that we perceive the size of distant sources of food, predators etc. as accurately as possible. If our ancestors had perceived proximate objects as larger than distant objects, their chances of survival would have been severely limited. One of the major evolutionary obstacles for the development of visual processing therefore, must have been to overcome the fact that distant objects are projected onto the eye in this way. What we see when we look at distant fruit is distant fruit, not tiny little tidbits. Nonetheless, when we draw distant fruit we have to render them at a smaller scale than closer fruit. It is nothing less than extraordinary that this technique works at all, let alone that it leads us to say that pictorial images look realistic. And it is no wonder also that it took us so long to discover one of the most important strategies for producing such images: perspective.

According to the influential educational psychologist Jean Piaget (1896-1980), young children do not draw “what they see”, they draw what they know. In Piaget’s view the child is not capable of “pure observation” but instead “he sees the world as if he had previously constructed it with his own mind.” Piaget’s theory has its origins in the philosophical doctrine of Subjective Idealism, a view that conceives of perception as a mental construct, an inner analogue of the external world - a world to which we have no direct access. Traces of this doctrine inform the work of many theorists and educationalists, both past and present, including John Ruskin and Georges-Henri Luquet whose work greatly influenced Piaget. Both Luquet and Piaget claimed that children’s drawings develop from “intellectual realism” (i.e. what children know) towards “visual realism” (i.e. what they see).

So, according to this theory, when we draw with visual realism, we draw what we see. But does this really stand up to scrutiny? Visual realism and reality (what we actually perceive) are by no means the same and it doesn’t follow therefore that there is a simple correlation between what we see and visual realism. If there were, we probably wouldn’t need to distinguish between the real and the unreal, between the real and the realistic, between reality and realism.

We take it as given that everything we see is real and we reserve words like ‘realism’ and ‘realistic’ for representations. So, to say that photographs are realistic is to say very little about what we actually see.

It should be obvious that we have evolved to see the world – as far as is humanly possible – exactly as it is. And whilst pictures look like the world, the two are not interchangeable. Our terminology has evolved in its own ways to reflect this fact.

No matter how realistic an image might be and no matter how susceptible we are, in certain rare circumstances, to mistake images for reality, there is never some point at which pictorial realism gradually or suddenly becomes full blown reality. There is no special lens, no magical painterly potion, no mystical technique that will ever transform a depiction into reality. Realism is forever barred entry into the kingdom of the real.

For a child learning to draw, it is a significant challenge to acquire the many skills necessary to convert three-dimensional experience into two-dimensional depictions. Fortunately children are surrounded by examples (pictures) that show them that the feat is possible. But one of the most significant things that stands in their way is a formidable evolutionary background in which the perception of a distant predator has always been the perception of a distant predator, not the perception of a cute little predator looking at us with hunger in its eyes.

Difficulties for the Philosophy of Illusion

Illusion:

1.     a deceptive or misleading appearance

2.     a false or misleading impression, idea, belief or understanding

3.     a false perception of an object or experience due to the mind misinterpreting the evidence relayed to it by the senses.

The concept of illusion is an ancient one, yet its venerable pedigree and widespread popularity are no guarantee of its value as a tool for uncovering the nature of perception. It is a handy concept for sure, but perhaps we should be wary of convenience, especially as a route to insight.

Examples of illusion are not difficult to come by - many are the stock-in-trade of conjurers and ‘illusionists’, but where some illusions might be distinguished, especially from the tricks and entertaining feats of stage-artists, is in the degree to which they might be regarded as having something to reveal about the workings of perception. When a conjurer uses slight of hand to “deceive the eye”, we do not suppose that this has anything useful to tell us about our sensory capacities. The idea that, with sufficient skill and dexterity, the hand can move more expertly than is easily perceived is unsurprising. Optical illusions, on the other hand, produce puzzling responses or anomalous visual artefacts that call for more sophisticated explanations.

Philosophers – knowing that their theories often stand or fall on the evidence of scientific enquiry – frequently refer to optical illusions in order to substantiate their claims. However, during the 1960’s, an important body of evidence emerged that cast significant doubt on many of these claims, yet has gone largely unacknowledged in philosophical circles.

Müller-Lyer Illusion

In a 1966 study undertaken by Segal et al into cross cultural variations in susceptibility to optical illusions the researchers found significant variance between differing communities and age groups across the globe. Some groups, for instance, reported little or no difference between the apparent lengths of the lines of the famous Müller-Lyer diagram. An earlier study by Hudson (1960), of culturally isolated South African children, encountered very similar findings. Both studies attributed their results to a lack of habitual exposure to pictures amongst the communities studied. Hudson dubbed this lack of familiarity: ‘pictorial illiteracy’. In fact, even children well schooled in language and arithmetic skills (but lacking pictorial literacy) were not susceptible to what is commonly described as the ‘pictorial illusion of depth’ and were therefore unsusceptible to the depth cues that many optical illusions exploit.

In 2006 Robert N. McCauley and Joseph Henrich write:

For those who experience it, the illusion may persist, but susceptibility to the Müller-Lyer illusion is neither uniform nor universal. Moreover, a plausible argument can be made that through most of our species’ history most human beings were probably not susceptible to the illusion.

If this is correct, and corroborating evidence can be provided from the art historical record to support it - then there is good cause to doubt the relevance of illusion in the explanation of perception. Moreover, if the following evidence is anything to go by, these doubts are in greater need to explanation than ever.

In numerous well documented studies, it has been shown that when people reach to grab three-dimensional versions of optical illusions, their grip aperture (the distance between finger and thumb) is unaffected by the illusion. So, whilst we may be inclined to say that one part of an optical illusion appears to be larger than the other, our ability to physically interact with these illusions is unaffected.

From an evolutionary point of view, it is of the utmost importance that we do not confuse a distant object for a small object, especially if the distant object has significance for our potential to survive. It is extraordinarily fortunate in fact, that the capacity to recognise and use perspectival images (in which distant objects are depicted at disproportionate scales than nearby objects) has not been entirely overridden by the evolution of our perceptual skills. If the research of Hudson, Segal et al is correct, then it would seem that this capacity to derive depth cues from perspectival images is a learnt skill and is not an immediately available part of our genetically acquired perceptual repertoire. And McCauley and Henrich are surely right when they speculate that our susceptibility to illusions must be a relatively recent consequence of the increasingly widespread use of pictorial imagery. What better explanation do we have for the widespread indifference amongst animals to our attempts to interest them with images?

"The Innocent Eye Test", Mark Tansey, 1981

I hope the evidence presented here makes it clear that the standard view of illusion - as deceptive, misleading or false - is thoroughly inadequate as a tool for the investigation of the nature of perception.  If our theories don’t fit the evidence then it is time to change our theories. I suggest that we start with the theory of illusion.