Featured Video Play Icon

The Mental Affordance Hypothesis

Dr Tom McClelland
University of Warwick


Affordances are opportunities for action. A teapot, for example, has the property of being grippable. When a subject grips the teapot, she exploits the teapot’s affordance. The concept of affordances, introduced by the ecological psychologist J.J. Gibson (1966), has been applied extensively across a range of disciplines. Throughout the considerable literature on affordances the afforded actions that theorists discuss are, with only a few exceptions, bodily actions such as gripping, walking or eating. This paper presents the hypothesis that we are also sensitive to affordances for mental actions such as attending, imagining and calculating. Although this Mental Affordance Hypothesis is ultimately answerable to the empirical evidence, a variety of phenomenological and theoretical considerations strongly suggest that we are appropriately sensitive to opportunities for mental action. 


Target Presentation in PDF

Target Presentation from Tom McClelland (Warwick)

Is there already a concept of mental affordances in the literature? In design theory, Hartson (2003) introduces a concept of cognitive affordances, but these are understood as features that aid an agent’s understanding of the use of an artefact rather than as opportunities for mental action. Also in design theory, Zhang & Patel (2006) define cognitive affordances as those affordances that depend on background knowledge, such as post-boxes only affording posting to an agent with an understanding of the postal system. Perhaps one could argue that the act of posting a letter is thus mental in some sense, but mental action clearly isn’t Zhang & Patel’s primary concern here. In artificial intelligence, Raubal & Moratz (2008) present an artificial agent that responds to affordances to perform the mental act of deliberating about which bodily affordance to act upon. Sloman similarly posits ‘deliberative affordances’ for human subjects (2008). In both cases though, the only mental act considered is deliberation, so there is no real exploration of the broader class of mental affordances. In the philosophical literature on affordances, Proust (2016) posits ‘cognitive affordance-sensings’ but her concern is specifically with meta-cognitive feelings, and her characterisation of these affordances as non-conceptual appraisals of one’s situation diverges from the conception of affordances that I will offer in this paper. Scarantino mentions the possibility of mental affordances in passing (2003, pp. 960-961), and Rietveld & Kiverstein explore affordances for ‘high-level’ actions that one might characterise as mental (2014). The concept of mental affordances is given a preliminary treatment by the author (McClelland 2015) but is not developed in any detail. In summation, although there are murmurings about affordances for mental action, nobody has unpacked exactly what it means for there to be mental affordances or made an explicit case for their existence.[1] The current paper is intended to remedy this.

The paper proceeds in four stages. In the first section I explore the somewhat nebulous concept of affordances and identify the key conditions of affordance-possession. In the second section, I use these conditions to offer a precise formulation of the Mental Affordance Hypothesis. In the third section I introduce a series of cases that plausibly satisfy those conditions: affording attention; affording the imagining of a bodily action; and affording counting. In then conclude by sketching a mental affordance research program that would reinforce my case for mental affordances and establish the applications they have to a range of theoretical issues.

1. What are Affordances?

1.1. Affordances as Opportunities for Action

Gibson introduced the term ‘affordance’ in his 1966 work The Senses Considered as Perceptual Systems. His understanding of the concept evolved throughout his career, and his most fully developed account of affordances can be found in his final work The Ecological Approach to Visual Perception (1979). In this book, he introduces the concept as follows:

The affordances of the environment are what it offers the animal, what it provides or furnishes, either for good or ill. The verb to afford is found in the dictionary, but the noun affordance is not. I have made it up. I mean by it something that refers to both the environment and the animal in a way that no existing term does. (1979, p. 127)

Although the notion of affordances remains a key concept in the ecological school of psychology (e.g. Michaels 2003) it has also been taken up across a wide range of other disciplines including: cognitive psychology (e.g. Tucker & Ellis 1998); neuroscience (e.g. Cisek & Kalaska 2010); music (e.g. Krueger 2011); anthropology (e.g. Ingold 2011); design theory (e.g. Norman 1999) and artificial intelligence (e.g. Horton et al. 2012), not to mention phenomenology (e.g. Dreyfus 2002) and philosophy of perception (e.g. Siegel 2014). Although the proliferation of the concept of affordances gives us a vast and diverse body of research from which to draw, it comes at a cost: the term ‘affordance’ has become unmanageably polysemous, with different theorists using the term in different ways to suit their varied purposes (Hartson 2003; Michaels 2003; and Scarantino 2003). Our first task, then, is to pin down an appropriate conception of affordances.

At the heart of the concept of affordances is the notion of opportunities for action. An opportunity to perform an action is a situation in which it is possible for a subject to deploy some ability they possess. Consider a tree’s property of being climbable by me. Affordances are relativised to particular subjects, thus what is climbable for me will differ from what is climbable for you, which will in turn differ from what is climbable for a squirrel. For a tree to be climbable for me is for it to stand in a certain relation to my ability to climb: it must be something toward which I can successfully deploy that ability (see Nanay 2010, pp. 430-432). Some trees will stand in this relation to my climbing ability (e.g. sturdy trees with plenty of branches) and other trees will not (e.g. weak trees with too few branches). The tree’s climbability and my ability to climb it are complementary dispositions: they are a mutually dependent pair of dispositional properties, much like a sugar cube’s disposition to dissolve in my tea and my tea’s disposition to dissolve the sugar cube (Turvey 1992). The foregoing suggests the following condition on some object or situation x affording φ-ing for S:

The Opportunity Condition: x affords φ-ing for S only if x constitutes an opportunity for S to φ.

A couple of clarificatory remarks are in order. First, the entity x that constitutes an opportunity for action needn’t be an object. A scenario on the street might present an opportunity for me to cross the road, but here it is the overall situation that presents the opportunity to φ rather than some particular object (see Siegel 2014). Also, an entity constitutes an opportunity for action for a subject independently of how the subject takes it to be (Gibson 1979, p. 139). A tree might be climbable by me even if I fail to notice its climbability, and might be unclimbable by me even when I mistakenly take it to be climbable (see Gaver 1991).

The Opportunity Condition is almost universally treated as a necessary condition of affordance-possession.[2] It should not, however, be regarded as a sufficient condition. Positing affordances is meant to tell us something substantive about how agents successfully engage with their environment, yet positing opportunities to act tells us nothing at all. After all, it is hardly an insight that there are situations in which it is possible for creatures to exercise their abilities. Any theoretically informative conception of affordances must supplement the Opportunity Condition with a more demanding condition (or conditions) on affordance-possession. There are two overlapping strands in the affordance literature that introduce conditions concerning the way in which subjects are sensitive to opportunities for action. In the first strand, the further condition of affordance-possession is that affordances are perceptible. In the second strand, the further condition of affordance-possession is that perception of the object or situation potentiates the afforded action i.e. automatically readies S to perform that action. I will use the term ‘sensitivity’ to encompass both of these ways in which a subject might be responsive to an affordance. I now turn to consider each of these possible conditions in turn.

1.2. Affordances as Perceptible Opportunities for Action

Affordances are commonly characterised as perceptible opportunities for action (see Gibson 1979; Michaels 2003; Dotov et al. 2012). On this view, recognising the climbability of a tree is not a matter of seeing a set of qualities then inferring that an object with those qualities is climbable by us. Instead, we can simply see its climbability. This gives us the following condition:

The Perceptibility Condition: x affords φ-ing for S only if x’s property of constituting an opportunity for S to φ is perceptible by S.

This condition does not bring with it any commitments regarding the nature of our perceptual relation to affordances. Gibson’s understanding of affordances is bound up with a number of claims about affordance perception, including the claims that: affordances are perceived directly rather than represented; affordances are perceived through ‘optic flow’ information without any need for internal processing, and; that in ordinary perception we are aware exclusively of affordances. The Perceptibility Condition, however, should be read in a way that is compatible with views that oppose Gibson on these matters. It is consistent with: a representational view of affordance perception (e.g. Prosser, 2011; Siegel 2014); with affordance perception involving internal processes that disambiguate ambiguous sensory inputs (Christensen & Bicknell forthcoming); and with ordinary perceptual awareness being characterised not just by affordances but by objects and qualities (Nanay 2010; Christensen & Bicknell forthcoming). The condition is also neutral on whether our perception of affordances is visual or some other mode of perception (though for convenience I will often talk of seeing affordances).

Why think we perceive opportunities for action? This is a thesis that should be judged on its explanatory value. Although the explanatory applications of this thesis are too numerous to explore here, there are some that should be highlighted. One application concerns the adaptive value of such perceptual capacities: if opportunities for action had to be inferred using our capacity-limited cognitive resources, our responsiveness to salient opportunities would be impeded in situations where our cognitive resources are over-stretched. Another key application concerns skill acquisition. The transition from unskilled to skilled action plausibly involves the acquisition of certain perceptual capacities. For example, where a novice driver has to infer that a situation requires braking, a skilled driver can simply see this feature of her situation.[3] Finally and, for my purposes, most importantly, the thesis that we perceive opportunities is often motivated phenomenologically. Many report, for instance, that they perceptually experience the teapot not just as white and as smooth but as grippable. Some of the most vivid experiences are those in which we are presented with a particularly strong affordance, such as a frosted cake strongly affording eating (Siegel 2014). This important aspect of our phenomenology is best explained in terms of our perceiving opportunities for action.

1.3. Affordances as Potentiating Opportunities for Action

The Perceptibility Condition is present – whether explicitly or implicitly – in a great deal of the affordance literature. There is, however, a different condition on affordance-possession that underwrites a substantial body of research in cognitive neuroscience. Here what matters is a kind of motoric sensitivity to opportunities for action. On this conception, an object or situation x affords φ-ing for S only if S perceiving x potentiates S actually φ-ing.[4]

A range of studies suggest that when we see an item that presents an opportunity to perform a particular action, the motor pathways responsible for performing that action (or parts of that action) are potentiated. In a study by Tucker & Ellis (1998), subjects were required to identify the orientation of a presented object by pushing a button with their left hand to indicate that the object is upside-down or pushing a button with their right hand to indicate that it is the correct way up (or vice versa in other trials). This study revealed an ‘interference effect’. Where the presented item is a teapot, if the handle of the teapot is facing toward the subject’s right-hand, this slows down their response time when the trial requires a left-handed button-push and improves their response time when the trial requires a right-handed button-push.[5] Ellis & Tucker explain this interference in terms of the perception of the teapot potentiating a right-hand grasp. The readying of this right-handed movement makes a right-handed button push faster, but impedes a left-handed button push. Crucially, they propose that perceiving the teapot potentiates the appropriate grip automatically: since gripping the teapot is irrelevant to the task, the potentiation of the gripping motion must occur independently of the agent’s intentions. They interpret this phenomenon as manifestation of the intimate connection between perceptual and motor systems in the brain. They explain that ‘[t]he visual system is highly integrated with the motor system to the extent that no clear divide exists between what one could call purely visual processing and purely motor processing.’ (1998, p. 830) Consequently, such motoric responses to perceived stimuli should be unsurprising.

We gain an interesting clue into how potentiation works from the phenomenon of utilization behaviour (Brazzelli & Spinnler, 1998). This is a condition, caused by brain damage to the frontal lobe (Besnard et al. 2011), in which subjects are compelled to ‘utilize’ items that they see. When presented with an apple, subjects eat the apple regardless of whether they are hungry. When presented with a toothbrush, they brush their teeth even in inappropriate contexts such as a doctor’s office. When presented with pens, they draw with them even if there’s no paper to draw on. This condition has been interpreted in terms of subjects being unable to suppress the motor processes automatically triggered by their environment (e.g. Rietveld 2012; Cisek & Kalaska 2010). Perception of an apple automatically potentiates the motor process responsible for eating. In a typical subject, if eating the apple is an unsuitable response then that motor process is suppressed (whether consciously or unconsciously). But for subjects with relevant frontal lobe damage, such suppression is not possible so they are compelled to eat. These considerations point us toward the following condition on affordance-possession:

The Potentiation Condition: x affords φ-ing for S only if S’s perception of x potentiates the process responsible for S φ-ing.

In order to apply this condition fruitfully, we must be specific about what it means for an action to be potentiated. The foregoing discussion highlights four key features of potentiation. I suggest that these features should be encoded in an operationalised definition of potentiation. S’s perception of x potentiates S φ-ing just in case S’s perception of x causes S to be in a state such that: if S does not suppress φ-ing then S will actually φ; if S actually φs in response to x then S’s response time will be more rapid than it would have been had S not been in that state; if S performs an act relevantly congruent with φ-ing then their performance of that act will be more rapid than it would have been if they were not in that state; if S performs an act relevantly incongruent with φ-ing then their performance of that act will be more slow than it would have been had S not been in that state.

Why think that stimuli automatically potentiate actions? Although the experimental and pathological data discussed offer considerable support for this thesis, it gains further support from its wider explanatory applications. As with the thesis that opportunities for action are perceptible, it has explanatory applications regarding the adaptive value of potentiation and regarding the role of potentiation in skill acquisition. It also has important applications to our agential phenomenology. The agential experience of catching an on-coming ball is best characterised as one of permitting an action to occur rather than of initiating an action spontaneously. The notion of potentiation helps explain this phenomenology: the on-coming ball potentiates the act of catching the ball, so in order to catch the ball the subject need only allow this process to unfold rather than having to initiate the act of catching.

2. What Are Mental Affordances?

In light of the foregoing, how should we frame the hypothesis that there are affordances for mental action? At the very least, we need the condition that x affords a mental action only if it constitutes an opportunity for mental action. Given our discussion of the thinness of the Opportunity Condition, it should be no surprise that it can be satisfied fairly straightforwardly. Mental actions include (though are by no means limited to) attending, imagining, remembering, expecting, evaluating, deciding, calculating and judging. There is considerable debate over the extent to which we have agency over our mental processes (see Soteriou & O’Brien 2009). Although this debate has important implications for what does and does not count as a mental act, it is worth noting that the view that there are no mental actions does not have a serious following. As such I will help myself to the claim that there are such things as mental actions. The division between mental and bodily actions is likely to be blurry. The mental act of deliberating about something may involve the overt bodily act of talking to yourself or (as we will touch on later) perhaps a covert ‘off-line’ bodily act of talking in your head. Nevertheless, such fuzziness should not lead us to doubt that there are such things as mental acts.

Opportunities for mental action are situations in which it is possible for an agent to exercise a particular mental capacity. I take it to be uncontroversial that there are opportunities for mental action. A stimulating documentary presents an opportunity to reflect, an old photo album presents an opportunity to reminisce, a place of worship presents an opportunity to contemplate and a fantasy novel presents an opportunity to imagine.[6] With a little work, a wealth of such examples could be provided. This leads me to conclude that the Opportunity Condition is easily satisfied. However, we have seen already that the Opportunity Condition is not plausibly a sufficient condition of affordance-possession. It seems we must adopt at least one of the further conditions that specify the way in which subjects are sensitive to these opportunities for action, but which one? I propose adopting both further conditions. If difficulties emerge for the claim that opportunities for mental action are both perceptible and potentiate the afforded mental action, we are free to retreat to the more moderate position that we are only sensitive to such opportunities in one of these respects. However, until such difficulties emerge we should aspire to find mental affordances that satisfy both extra conditions. Another reason for adopting both further conditions is that there is an attractive position according to which potentiation and affordance perception are two sides of the same coin. On this view, when gripping is potentiated by a grippable object, that motor-state itself constitutes a perceptual representation of the object’s grippability. This view is motivated by the thought that the state in question is assessable for accuracy: it is veridical when the stimulus is grippable and non-veridical when the object is not. Although it would be inappropriate to assume such a view without further argument, it does give us further reason not to pick one condition at the expense of the other. The foregoing yields the following account of mental affordance-possession. An object or situation x affords a mental act φ iff:

  1. The Opportunity Condition: x constitutes an opportunity for S to perform the mental act of φ-ing.
  2. The Perceptibility Condition: x’s property of constituting an opportunity for S to φ is perceptible by S.
  3. The Potentiation Condition: S’s perception of x potentiates S φ-ing.

The Mental Affordance Hypothesis is simply the claim that there exist at least some mental affordances satisfying all three of the conditions above.[7]

A number of a priori considerations obstruct the case for the Mental Affordance Hypothesis. The first worry is that the Mental Affordance Hypothesis predicts an implausible proliferation of affordances. Opportunities to grip are constrained by the presence of grippable objects in your perceptual environment. Opportunities to attend, in contrast, are ubiquitous. Being attendable is a far less demanding property than being grippable, so any perceptual environment will be saturated with attending-affordances. The same goes for opportunities to imagine, to count or to reflect upon. The Mental Affordance Hypothesis predicts we are sensitive to this plethora of affordances: our perceptual experiences would be shot through with myriad mental affordances, and each of these mental acts will be potentiated. The implausibility of such a commitment casts doubt on the hypothesis.

I suggest that there is a parallel problem regarding bodily affordances, and that the solution to this problem can be imported to address the challenge to mental affordances. Although opportunities for mental action are ubiquitous, this does not mean that all those opportunities are perceived, or that all those opportunities potentiate the relevant mental act. Our sensitivity to opportunities is tuned to relevant opportunities. In a field full of footballs, we might perceive the kickability of a particularly well-placed football, and the kicking of that ball could be potentiated, but that won’t be the case for all the footballs we perceive. Similarly, we will only be sensitive to those opportunities to perform a mental act that are particularly relevant. An aptness-filter is built into affordance sensitivity, so responding to an affordance is limited only to those cases where action is likely to be apt. This is not to say that our responses to affordances are wholly sensitive to aptness. Recall, teapot gripping is potentiated even though it is irrelevant to Tucker & Ellis’s task, and utilization behaviour patients characteristically perform inappropriate actions. Nevertheless, an aptness-filter is still evident in both cases: not every opportunity for left- or right-handed actions in the subjects’ perceptual environment interferes with their performance, and not every opportunity for action in the perceptual environment is acted on by the utilization behaviour patient.[8]

Other a priori obstacles to mental affordances target particular conditions. Regarding the Perceptibility Condition, one might worry that opportunities for mental action are too sophisticated to be represented perceptually (Nanay 2010, p. 432). It is one thing to perceive a ball as to-be-caught but quite another to perceive a situation as to-be-deliberated-about. This parallels worries in wider debates about perceptual content: one might accept that a perceptual state can represent a tree as green, and perhaps even represent it as having a certain characteristic Gestalt, but be resistant to the claim that it represents the tree as a pine tree. Perhaps perceptual content is too basic to encompass opportunities for mental action. Regarding the Potentiation Condition, one might worry that only motor processes are the kind of process that can be potentiated by stimuli. To the extent that mental action is non-motoric, it cannot be potentiated. Although I was careful not to define potentiation motorically, it might still be suggested that only motoric states could satisfy the operationalised conditions that I presented. Neither of these worries are knock-down objections to the Mental Affordance Hypothesis. They do, however, highlight the kind of challenge we will face when trying to find well-supported cases of mental affordance-possession.

What method should be used to establish whether some candidate satisfies the three conditions? I will use phenomenological observations as a point of departure. Although such observations have often played a key role in supporting claims about our sensitivity to affordances, I will endeavour not to rely on phenomenological observations too heavily. The first reason for this is that phenomenological observations are too easily disputed, and such disputes are notoriously difficult to resolve. The second reason is that phenomenology is an imperfect reflection of our underlying mental processes: one can perceive an opportunity to grip and have a gripping response potentiated without it showing up in one’s experience, and one can have an experience as of grippability and the potentiation of gripping without the relevant underlying psychological states. Despite these limitations, it would be a mistake to disregard phenomenological considerations entirely (see Koffka 1935). Defeasible evidence is still evidence, and I will supplement these observations with theoretical and empirical considerations. That said, the claim that there are mental affordances is ultimately answerable to the empirical data, so my argument constitutes only a preliminary case for mental affordances that, I hope, will motivate the relevant empirical investigation.


3. Are There Mental Affordances?

Case I: Affording Attention

Consider the following situation. You are working at your desk but outside a builder has a radio on too loud. Although you successfully keep your attention trained on your work, the music outside is a continual distraction. I suggest that this scenario is best described in terms of the noise affording focal attention.

Phenomenologically, the music is experienced as demanding our attention. The music invites us to perform a certain act, namely that of focally attending to it. Crucially though, this is an invitation we are free to ignore, and it is possible to succeed in keeping our focal attention trained on our work. This is not to say that we don’t attend to the music at all (indeed, a case could be made for thinking that wholly unattended stimuli are unexperienced, which would be at odds with the phenomenology of the scenario). Rather, we attend to the music only peripherally and resist the invitation to direct our focal attention toward it.

Does the music satisfy the three conditions of affordance-possession? The music presents an opportunity for us to perform a certain action – the act of focally attending to it – so the Opportunity Condition is plausibly satisfied. Furthermore, our representation of the music as to-be-attended is more plausibly construed as perceptual than non-perceptual. Although there are no uncontroversial criteria for distinguishing perceptual from non-perceptual states, two features strongly associated with perceptual processes are that they are non-inferential and doxastically impenetrable. One’s representation of the noise as to-be-attended certainly seems to be non-inferential – it is something we recognise directly rather than with the help of mediating premises. Furthermore, one’s representation of the noise as to-be-attended seems to be doxastically impenetrable. If it were doxastically penetrable, the belief that our attention is better directed at our work would stop us from representing the music as to-be-attended. What we find, however, is that the noise continues to make a demand on our focal attention regardless of our beliefs.

Our auditory perception of the music also plausibly potentiates the act of focally attending. This is not a situation in which attending to the stimulus is obligatory – in the scenario described, we succeed in keeping our focal attention directed on our work. But nor is it a situation in which our focal attention is unaffected by the stimulus – we are in a state such that we would focally attend to the music if we stopped deliberately directing our attention at our work. This situation is plausibly understood in terms of the stimulus potentiating our focal attention. The stimulus automatically readies our focal attention to be directed toward it, but we are able to suppress this automatic signal and direct our focal attention elsewhere.

We have initial reasons to believe that the music satisfies all three conditions of affordance-possession. The only remaining question is whether attending is a mental act. Attention has both an overt aspect and a covert aspect. Overt attention is the bodily activity of directing one’s sense organs toward a particular stimulus, property or region. Covert attention is the mental activity of concentrating on a particular perceived stimulus, property or region. These two activities are dissociable: one can direct one’s covert attention toward things other than the target of one’s overt attention. That said, the two activities typically coincide. The focus of our gaze, for instance, is normally the focus of our concentration. Crucially, when we deliberately attend to a stimulus we don’t typically perform two acts – one bodily and one mental. Rather, attending involves a single act of will, albeit an act that has both bodily and mental results. I suggest that attention should be regarded as a hybrid act that is both bodily and mental in nature. Insofar as it is partly mental, affordances to attend will thus qualify as mental affordances.

The fact that attention has this hybrid nature suggests an interesting argument for mental affordances. If one is on board with the idea of affordances for bodily action, then there’s no principled reason to resist affordances for the bodily action of overtly attending. But if one is going to countenance affordances to attend, one should posit affordances for the hybrid act of (overtly and covertly) attending rather than just for the bodily act of overtly attending. After all, the two aspects of attention have deep functional interconnections, and rarely come apart. As such, the case of attention makes it uncomfortable for someone to countenance affordances for bodily action whilst rejecting affordances for actions that are at least partly mental.

The result that there are affordances to attend is valuable as it shows that there is at least one kind of mental affordance. What it does not do, however, is give us reason to posit affordances for a wider set of mental acts relevantly similar to attending. Attending is a sui generis act so there are no mental actions of the same genus that we can infer are also afforded. Our next target should be a mental act that is plausibly representative of a wider class of mental actions.

Case II: Affording the Imagining of a Bodily Action

Consider the following scenario. You are traversing a series of stepping stones across a river. The first dozen stones are fairly easy to deal with, and you perform the appropriate hops and steps without having to reflect on your actions. You then get to a trickier stone. You pause, mentally rehearse the required leaping manoeuvre, perform the rehearsed leap then continue on your way. I suggest that each of the actions involved in this scenario is best described in terms of affordance perception, and that when you reach the tricky stone the act afforded is the mental act of rehearsing your leap in imagination. The phenomenology of this scenario is plausibly captured in terms of our sensitivity to affordances.

Regarding perceptibility, as you traverse the easy stones, there is no need to infer what kind of step can be performed on the next stone. Rather, you perceive the specific stepping action afforded by the stone. I suggest that exactly the same applies to your experience of the tricky stone – that you perceive the stone as affording a certain mental act viz. the act of mental rehearsal. And this appearance seems to be cognitively impenetrable. If you’re convinced by a friend that the best strategy is to not think and to just keep going, then pausing and rehearsing your leap will be at odds with your beliefs. But since the appearance that the tricky stone affords this act is perceptual, it is relatively insensitive to your background beliefs.

Regarding potentiation, as you traverse the easy stones your experience is not of initiating the appropriate stepping action but is rather one of permitting yourself to perform steps potentiated by the perceived stone. Again, I suggest the same applies to your experience when you act on the tricky stone – you do not deliberately initiate the act of mentally rehearsing your jump but rather permit a potentiated mental action to unfold. Moving beyond phenomenological considerations, there is also indirect empirical evidence that imaginings of bodily acts can be potentiated. A wealth of data suggests that the neural underpinnings of an imagined bodily act overlap extensively with the neural underpinnings of the actual performance of that act (Jeannerod 1995; Kessler & Thomson 2010). This drives a convincing account of imagined bodily acts as ‘off-line’ performances of bodily acts. If we grant, as I think we must, that the processes responsible for bodily acts are automatically potentiated by stimuli, then we should also accept that the processes responsible for imagined acts can also be automatically potentiated by stimuli. After all, if they are to a significant extent the very same process, then if one can be potentiated it is plausible that the other can potentiated too. Even those who claim that only motoric processes can be potentiated have nothing to object to here.

We are thus led to the defeasible conclusion that there are affordances to mentally rehearse bodily actions. As with the case of attention, if one accepts affordances for bodily actions it is hard to resist the conclusion that there are at least some affordances for mental action. Unlike the case of attention, we have opened up a large class of mental actions for which there are plausibly affordance. If we can perceive affordances to mentally rehearse a certain kind of leap, we can presumably perceive affordances to mentally rehearse a host of other bodily actions. The space of perceptible affordances for bodily actions might even be duplicated in a space of perceptible affordances for the imaginative performance of those same bodily actions.

By considering mental acts with an intimate connection to bodily action, we put ourselves in a good position to argue that those who countenance affordances for bodily action ought also to countenance affordances for these mental actions. A limitation of this strategy, however, is that one might object that only those mental acts with a close connection to bodily action can be afforded. It might be objected that opportunities to perform abstract mental acts detached from the bodily are too complex to be perceived, or that the wholly non-motoric nature of such acts precludes them from being potentiated. This leads us to our third and final candidate mental affordance.

Case III: Affording Counting

Counting is a mental act. Sometimes we count in a way that involves the bodily act of pointing to items and numbering them out loud. Sometimes we count in a way that involves doing those bodily acts off-line i.e. by pointing and numbering in our heads. It is implausible, however, that the act of counting is exhausted by such overt or covert bodily action. We can count things without performing either of these acts, and we have a brain area – the intraparietal sulcus – that is directly associated with arithmetic without being directly implicated in those bodily acts (Dehaene et al. 2004). My target here is what you might call unassisted counting: a way of counting that depends on neither covert nor overt bodily action. Our environment can present opportunities for counting. Consider a jar full of marbles, a pile of pennies, or the leaves on a clover. The question is whether we perceive the marble jar as affording counting, and whether perceiving the marbles potentiates the act of counting. I trust that readers will be able to imagine the kind of phenomenological case I would make for this conclusion. For this case though, I will instead focus on some pathological data that suggests we are relevantly sensitive to opportunities to count.

The manifestations of utilization behaviour we have discussed so far are all bodily acts: eating an apple, brushing with a toothbrush and writing with a pen. Interestingly though, the patient discussed by Brazzelli & Spinnler also showed a ‘compulsion to count’ (1998, p. 350).[9] This indicates that the act of counting is potentiated by our perception of opportunities to count. Where neurotypical subjects naturally suppress this signal to act the patient is unable to do so, hence her atypical behaviour. It is worth noting that the patient’s symptoms are not naturally explained in terms of atypical behavioural urges: the characteristic feature of the disorder is that the patient’s behaviour is environment led, meaning that she acts on perceived opportunities for actions regardless of whether she has a desire to perform those actions. Consequently, the fact that she performs the act of counting on certain stimuli indicates that she perceives those stimuli as constituting an opportunity to count. This case again puts pressure on those who accept affordances for bodily action to countenance affordances for certain mental actions. If the compulsive bodily actions of Brazzelli & Spinnler’s patient are to be understood in terms of affordance perception then, other things being equal, the same treatment ought to be given to her compulsive mental actions.

We are thus led to the defeasible conclusion that there are affordances to count. Where does this leave us? Counting is part of a wider class of arithmetical actions, so to the extent that one finds it plausible that there are affordances to count one should also grant the possibility of affordances for other arithmetical actions. A pile of sweets, for example, might present an opportunity for division. And stimuli in the language of mathematics can present opportunities for far more sophisticated arithmetical actions. Moving beyond arithmetical actions, one might take the existence of counting-affordances as evidence that the scope of mental affordances is unlimited. We’ve moved beyond affordances for off-line bodily activity to mental acts more detached from bodily action, and if we can perceive opportunities to perform one act of this kind, why not all such acts?

Those convinced that we cannot perceptually represent complex properties like opportunities for abstract mental action will be resistant to this conclusion. Those convinced that potentiation is limited to motoric processes will be similarly resistant. The burden of proof will be on them, however, to demonstrate how cases like the countable marble jar are relevantly different to accepted cases of affordance-possession. Of course, the dispute here may well fragment, with objections to the perception of opportunities for abstract action differing in force from objections to the potentiation of mental action by stimuli. We should thus be open to the possibility that some opportunities for mental action will only qualify as mental affordances if we soften our conditions of affordance-possession so as not to require both perceptibility and potentiation.


5. Conclusion: The Mental Affordance Research Program

My preliminary case in favour of the Mental Affordance Hypothesis is by no means conclusive. It should, however, be enough to motivate the pursuit of a mental affordance research program. It will be worthwhile to explore whether there are cases in which the perception of a task-irrelevant opportunity to perform a mental act interferes with our behaviour in the same way as Ellis & Tucker’s grippable teapot. It will also be worth exploring whether pathological conditions associated with affordance perception can lend further support to the conclusion that we are sensitive to opportunities for mental action. Besides these direct investigations into mental affordances, we should also explore the theoretical applications of mental affordances. Recall the explanatory considerations that motivate the claims about perceptibility and potentiation for bodily affordances. Can mental affordances be put to the same explanatory use? Perhaps we need to posit mental affordances to properly capture our perceptual and agentive phenomenology. Perhaps sensitivity to mental affordances has an adaptive value of minimising cognitive load and enhancing response time. And perhaps we must cite a subject’s sensitivity to mental affordances to explain her transition from unskilled to skilled mental performance. Though we are not in a position to predict the answers to these questions with any certainty we are, I hope, in a position to recognise their value.


Besnard, J., Allain, P., Aubin, G., Chauvire, V., Etcharry-Bouyx, F., & Le Gall, D. (2011). A Contribution to the Study of Environmental Dependency Phenomena: The Social Hypothesis. Neuropsychologia (49), 3279-3294.

Brazzelli, M., & Spinnler, H. (1998). An Example of Lack of Frontal Inhibition: the ‘Utilization Behaviour’. European Journal of Neurology (5), 347-353.

Bub, D. (2016). tdlc.ucsd.edu/events/boot_camp_2015/DBubTDLCtalk2015.pdf. Retrieved from tdlc.ucsd.edu/events/boot_camp_2015/DBubTDLCtalk2015.pdf

Christensen, W., & Bicknell, K. (forthcoming). Affordances and the Anticipatory Control of Action.

Christensen, W.; Sutton, J.; McIlwain, D.J.F. (2008) Cognition in Skilled Action. Mind & Language (31), 37-66.

Cisek, P. (2007). Cortical Mechanisms of Action Selection: The Affordance Competition Hypothesis. Philosophical Transactions of the Royal Society (362), 1585-1599.

Cisek, P., & Kalaska, J. F. (2010). Neural Mechanisms for Interacting with a World Full of Action Choices. Annual Review of Neuroscience (33), 269-98.

Dehaene, S., Molko, N., Cohen, L., & Wilson, A. J. (2004). Arithmetic and the Brain. Current Opinion in Neurobiology (14), 218-224.

Dotov, D.G., Nie, L. & Wit, d. M.M. (2012). Understanding affordances. Avant, 3(2), 28-39.

Dreyfus, Hubert L. (2002). Intelligence without representation. Phenomenology and the Cognitive Sciences 1(4): 367-383.

Gaver, W. W. (1991). Technology Affordances. Proceedings of the SIGHI conference on Human factors in computing systems.

Gibson, J.J. (1966). The senses considered as perceptual systems. Boston: Houghton-Mifflin.

Gibson, J. (1979). The Ecological Approach to Visual Perception. New York: Psychology Press

Hartson, R. (2003). Cognitive, physical, sensory, and functional affordances in interaction design. Behaviour and Information Technology, 22(5), 315-338.

Heft, H. (2001). Ecological Psychology in Context: James Gibson, Roger Barker and the Legacy of William James’s Radical Empiricism. New York: Psychology Press.

Horton, T., Chakraborty, A., & St. Amant, R. (2012). Affordances for robots: a brief survey. Avant, 3(2), 70-86.

Ingold,  T. (2011). The Perception of the Environment:  Essays on livelihood, dwelling and skill. London: Routledge.

Jeannerod, M. (1995). Mental Imagery in the Motor Context. Neuropsychologia, 33(11), 1419-1432.

Jenkins, H. (2008). Gibson’s “Affordances”: Evolution of a Pivotal Concept. Journal of Scientific Psychology, 34-45.

Kessler, K., & Thomson, L. A. (2010). The Embodied Nature of Spatial Perspective Taking: Embodied Transformation versus Sensorimotor Interference. Cognition, 114(1), 72-88.

Krueger, J. (2011). Doing Things With Music. Phenomenology and Cognitive Science, (10), 1-22.

Koffka, K. (1935). Principles of Gestalt Psychology, Oxford: Routledge.

McClelland, T. (2015) Affording Introspection: An Alternative Model of Inner Awareness, Philosophical Studies, 172(9), 2469-2492.kreme

Michaels, C. (1988). S-R Compatibility Between Response Position and Destination of Apparent Motion: Evidence of the Detection of Affordances. Journal of Experimental Psychology: Human Perception and Performance, 14(2), 231-240.

Nanay, B. (2010). Action Oriented Perception. European Journal of Philosophy, 20(3), 430-446.

Nanay, B. (2013). Between Perception and Action. Oxford: OUP.

Norman, D. A. (1999). Affordance, conventions and designs. Interactions, 38-43.

Prinz, J. (2012) The Conscious Mind, Oxford: OUP.

Prosser, S. (2011). Affordances and Phenomenal Character in Spatial Perception. Philosophical Review, 120(4), 475-513.

Proust, J. (2016). The Evolution of Primate Communication and Metacommunication. Mind and Language, 31(2), 177-203.

Raubal, M., & Moratz, R. (2008). A Functional Model for Affordance Based Agents. In E. Rome, Affordance-based robotic control (pp. 91-105). Berlin: Springer.

Rietveld, P. (2012). Bodily Intentionality and Social Affordances in Context. In F. Paglieri, Consciousness in Interaction. Benjamins.

Rietveld, E., & Kiverstein, J. (2014). A Rich Landscape of Affordances. Ecological Psychology, 26(4), 325-352.

Scarantino, A. (2003). Affordances Explained. Philosophy of Science, 70(5), 949-961.

Siegel, S. (2014). Affordances and the contents of perception. In B. Brogaard, Does perception have content? Oxford: Oxford University Press.

Soteriou, M. & O’Brien, L. (eds.) (2009). Mental Actions. Oxford: OUP.

Sloman, A. (2008) The Well Designed Young Mathematician. Artificial Intelligence, 172, 2015–2034.

Tucker, M., & Ellis, R. (1998). On the Relations Between Seen Objects and Components of Potential Action. Journal of Experimental Psychology, 24(3), 830-846.

Turvey, M. T. (1992). Affordances and Prospective Control: An Outline of the Ontology, Ecological Psychology, 4(3), 173-187.

Zhang, J., & Patel, V. L. (2006). Distributed cognition, representation and affordance. Pragmatics & Cognition, 14(2), 333-341.



[1] Since there is no real literature in favour of mental affordances, it should be no surprise that there is very little literature against them either. The nearest is perhaps Nanay who holds that we perceive opportunities for bodily action but denies that this can be extended to mental action (2013, p. 18). Although Nanay shows that his own arguments for the perception of bodily action properties do not extend to mental action, he does not present reasons to doubt that arguments for the perception of mental action properties might be developed.

[2] Some adopt a more liberal conception of affordances on which the thing afforded need not be an action. Gibson (1979 p. 39) talks about a fire affording warmth, for example, even though warmth is not an action. However, Michaels (2003) argues convincingly against such a liberal conception.

[3] Here I stop short of the thesis that skilled actions are wholly automatic and ‘mindless’ (for discussion, see Christensen et al. 2016).

[4] Confusingly, this literature sometimes uses the term ‘affordance’ to refer to the motor state triggered by the perceived entity rather than to the entity’s property of triggering that state. I will maintain the latter use.

[5] The precise character of such interference has been mapped more recently by Bub (2015).

[6] Just as a teapot does not present an opportunity to grip as such, but rather presents an opportunity to grip that very teapot, so too a documentary does not present an opportunity to reflect as such, but rather an opportunity to reflect on that very documentary. It would be a mistake, however, to assume that all mental affordances are such that the mental act is performed on the affording object. A space might afford contemplation without affording contemplation of that very space and perhaps without affording contemplation of anything in particular. This is explored more closely in McClelland (2015).

[7] Although I am attempting to offer a conception of affordances that encompasses major strands in the literature, I do not claim to offer a conception free from theoretical biases. I am offering a conception that allows us to offer a plausible and theoretically interesting formulation of the Mental Affordance Hypothesis, meaning that some strands of the affordance literature have been disregarded.

[8] One might go further and suggest that we represent the aptness of the affordance: that the ball is represented as good-to-kick. My proposal stops short of this. Aptness is key to the aetiology of affordance sensitivity, and this is enough to assuage worries of proliferation without making a further claim about normative content.

[9] A complication here is that the cases of counting observed by Brazzelli & Spinnler are, of course, cases of overt counting. One might claim that it is this bodily act that is afforded rather than the mental act of counting. However, the burden of proof would be on the objector to say why this is so. Ordinary subjects perform these bodily acts to assist a mental act of determining how many of something there are, and there is no obvious reason to doubt that the patient is doing the same. Put another way, the patient is most likely compelled to make bodily gestures that aid counting precisely because she is compelled to perform the mental act of counting.

Invited Comments from Dark Jones (Evansville)

Cognitive Affordances – Reply to McClelland’s “The Mental Affordance Hypothesis”

Derek Jones
University of Evansville


Tom McClelland presents a strong case for extending the concept of an affordance to cover mental action.  His account also offers both a plausible way of extending ecological accounts of perception and action beyond their usual domains of application and an interesting way of distinguishing novice and skilled behavior.  The paper’s core argument is straightforward: Since some objects afford mental action to agents in the same way that others afford bodily action, anyone who accepts a theory of bodily affordances ought to extend that theory to cover mental affordances.  Of course, if too many objects are taken to afford mental action this strategy might motivate a reductio of bodily affordance theory, but McClelland is careful to hedge his bets on just how many kinds of mental affordances there truly are.

The varied uses of “affordance” across disciplines complicate McClelland’s task, for he must first stipulate a general definition of affordance and then show that this definition can usefully cover mental and bodily actions alike.  He offers a set of necessary conditions – opportunity, perceptibility, and potentiation – that are satisfied in the case of bodily affordances, and which are either mostly or wholly satisfied in the case of mental affordances.  McClelland advances the stronger claim that mental affordances satisfy all three conditions, but reserves the right to retreat to a more moderate condition.  I am uncertain that this move is available, however; if bodily affordances must satisfy McClelland’s three conditions, then mental affordances ought either to satisfy them as well or be classified as something other than affordances.


Affordances and their Relata

The proposed move from bodily affordances to mental affordances raises new questions.  For example, it unclear how close the relationship between the affording object and the afforded capacity must be.  Citing Nanay (2010), McClelland notes that “for a tree to be climbable for me is for it to stand in a certain relation to my ability to climb: it must be something toward which I can successfully deploy that ability.”  Note that in this example the tree itself is the object of the deployed capacity – I perceive that tree as affording the climbing of that very tree (it certainly doesn’t afford the climbing of just any tree!).   This is common to most of the canonical examples of bodily affordances – handles afford their being gripped, a hammer affords its being swung, and so on – and it comports with the design theoretic usage of the term.  In such cases the complementary dispositions of agent and object are very tightly related.

The link is far looser in many cases of mental action, however.  While some objects prompt mental actions directed at those same objects – a song might afford focal attention upon its melody, a jar of marbles might afford counting of its contents– others prompt action directed elsewhere.  Suppose that the old tree in my backyard reminds me of my grandmother, and that when I perceive that tree I am prompted to reminisce about time spent with her.  Does the tree afford reminiscing about my grandmother in the same sense that it affords climbing?  Perhaps this is a distinction without a difference, but if the aim of a theory of affordances is to explain the fast, here-and-now dynamic interplay between agents and environments, it seems that affordances that prime behavior directed at their potentiating causes might be more fundamental than those that do not.

This problematizes several plausible cases of mental affordance.  To use McClelland’s example, when I encounter a tricky step along a stone path, I pause and mentally rehearse the movements necessary to continue on.  This stone affords “the mental act of rehearsing [my] leap in imagination”.  But in this case the object of the afforded mental behavior is a mental representation of a movement, not the rock.

The rock case has another unusual component – I don’t believe that the potentiating cause of the rehearsal upon encountering an oddly-placed rock is the rock, but rather the awareness that the rock doesn’t afford stepping.  Does the awareness of a lack of affordance itself afford visualization?  If this is correct, the case is a step further removed from canonical cases of bodily affordance: not only is the object of the potentiated behavior distinct from its cause, but both the cause and the object of the behavior are mental.

Why should this matter?  I worry that if we continue down this path we will find that much of our mental lives will be shot through with affordances.  Consider a memory that prompts the mental rehearsal of a song – an earworm.  Is the rehearsal afforded by the memory?  A similar consideration holds for trains of thought, where one thought prompts the next, and so on.  It seems plausible to say that in at least some cases each thought is a deliberate mental act, primed by features of the last.  Do we think of each thought as affording the thinking of its successor?  If we require of a theory of affordances that afforded behaviors target their potentiating causes, we need not classify trains of thought in this way.  If not, affordances seem to proliferate.  Perhaps this is not a problem, but it bears consideration – too much proliferation and the concept loses its explanatory value.



McClelland’s discussion of potentiation is particularly useful, and his empirical evidence for the claim that perceived affordances potentiate the motor behaviors they afford is plausible.  This focus on priming motor responses highlights the ecological component of perception and action–the environment, rather than serving as a mere problem space for the agent to navigate, co-contributes to the production of behavior by triggering appropriate behaviors.  Action is as much a matter of suppressing triggered behaviors as it is a matter of initiating them. The agent’s job is to “surf” its environment, making corrections when necessary but also knowing when to allow the environment to drive movement.  The question of whether such behaviors ought to be viewed properly as active lingers, but it falls outside the scope of this paper and there are various non-causalist theories of action available.

Whether potentiation ought to be included as a necessary condition for a complete theory of affordances, however, will depend again on how narrowly we wish to define the concept of an affordance.  McClelland’s examples of utilization behavior deal primarily with overlearned, canonical object-oriented behaviors (grasping a handle, eating an apple, using a toothbrush to brush teeth).  But will a stimulus potentiate other possible stimulus-appropriate behavior?  If it does not, do we wish to restrict affordances only to actions that we are used to performing?

For example, I am an experienced door-opener, and it seems plausible to say that my perceiving a door handle potentiates opening behavior.  But what of other possible actions?  I can climb trees but rarely do.  Suppose I perceive a nearby tree that I can climb – it is “something toward which I can successfully deploy that ability.”  I am also aware, looking at the tree, that I could climb it if I were so inclined.  Minimally, I perceive the opportunity insofar as I would not be surprised if you told me I could climb it.  Nonetheless, I feel neither a pull to climb the tree nor any need to suppress climbing behavior.  Does the tree afford climbing for me? If potentiation is necessary for affordance then either the tree does not afford climbing for me or I am simply unaware both of the potentiation of the behavior and my suppression.



Heading off cognitive overload worries, McClelland concedes that we cannot perceive all of the opportunities for action presented by our environments at any given time.  As affordance perception must be sensitive to relevance, he suggests that we employ an “aptness filter” that restricts the set of perceived affordances. Presumably the permissiveness of this filter will be a function of the organism’s global state.  This may go some way towards addressing my earlier worry –the climbable tree does not afford climbing for me in normal cases because I am not in a state such that my aptness filter would admit it as a relevant option.  But this raises new questions.

First, how is aptness calculated?  Consider McClelland’s example of kicking one football out of many.  Observing a sea of kickable footballs, we would be overwhelmed if each football potentiated a kicking response, and thus a relevance filter is needed.  With the filter in place, we see one appropriately-positioned ball out of the group as particularly kickable and filter out the other affordances.  But this leaves the question of how the agent’s perceptual system determines relevance.  Clearly if the filter operates by comparing the actions afforded by various options and selecting the best candidate from among them, it would do nothing to diminish cognitive load.  The frame problem looms.

Note also that we cannot solve the relevance problem merely by appealing to attention, since attendability is one of the mental affordances McClelland discusses.  In a scene, certain elements will present themselves as to-be-attended and others will not.  How is the set of relevant attendable items to be discerned?  We can presumably distinguish deliberate, focal attention from a broader sort of awareness.  But it would be useful to discuss the nature of this more general kind of attention, given the myriad theories of attention available.

Regardless of the underlying mechanism, in many cases relevance clearly is dictated by the intentional states of the agent.  I might not ordinarily perceive the climbable tree as climbable, but if I were looking for things to climb I would.  This seems especially true of certain mental actions: The tree in my backyard does not normally invite reflection on my grandmother, but it does when I believe that today is her birthday.  The jar of marbles affords counting when I am interested in winning a contest at the fair; it does not when I know that I have been tasked with loading fifty jars of marbles into a truck.  If the relevance of perceived affordances depends on our intentional states, however, this calls into question the importance of doxastic impenetrability discussed throughout the paper.  Whereas my beliefs do nothing to diminish the effect of the Muller-Lyer illusion, different affordances may present themselves depending on what I believe or intend, and changes in beliefs may change the set of relevant perceived affordances.

Perhaps we wish to say that the affordances themselves are doxastically impenetrable, but our intentional states merely affect selection.  But this creates tension with the perceptibility condition, which sets perceptibility as a necessary condition on affordance-hood.  My beliefs dictate which affordances I perceive, which, on this account as I understand it, dictates which affordances exist.

An alternate approach is available, though: perhaps cognitive agents do not filter affordances at all!  Perhaps something like a Dennettian “pandemonium” model is more appropriate – conflicting affordances mutually suppress one another, with the strongest potentiators drowning out the weaker ones.  The dominant affordance is presented to consciousness.  This would seem to be consistent with Gibson’s view that all perception is affordance-laden.  It does commit us to the view that all affordances are perceived, but perhaps the cognitive load is not so great if a single affordance dominates the others at a sub-personal level of perceptual processing.



In sum, I believe McClelland has motivated an interesting and fruitful research project.  Most of my critical remarks have taken the form of questions that could be addressed in future projects, and I look forward to reading his future work on the topic.

Invited Comments from Julian Kiverstein (Amsterdam)

Commentary on Tom McClelland’s “The Mental Affordance Hypothesis” 

Julian Kiverstein
University of Amsterdam


Perceptual experiences present animals with an environment that offers affordances or possibilities for action. What types of possible actions can an animal detect in perception? Canonical examples are stairs that are climbable, chairs that can be sat on, gaps that can be passed through         , flying balls that can be caught by an agent running at a certain speed, roads that can be crossed, fish that can be caught by diving gannets, and so on. Does the environment that is present to us in perception also include possibilities for mental actions like judging, deciding, remembering, imagining, and attending? This is the interesting question McClelland takes up in his paper.

Like McClelland I will start my commentary with the matter of definitions. Definitions are important in this case for getting the phenomenon of interest in view, namely affordances. McClelland begins his analysis of the concept of affordances with “the notion of opportunities for action” that are dispositional properties of objects or situations. “Affordances” as Gibson introduces the concept in his (1979) are to be understood in the wider setting of an organism-environment pair. Gibson writes (though perhaps somewhat confusingly) that the term “refers to both the environment and the animal in a way that no existing term does” (1979: 127). This important point is reflected in McClelland’s talk of the tree’s climbability as having as its complement climbing abilities. Thus what McClelland dubs the Opportunity Condition would need to be spelled out further by including reference to abilities, or what Turvey (1992) refers to as “effectivities”. In other words part of what it is for X to constitute an opportunity for S to φ is for S to have the ability to φ.

The Opportunity Condition however seems to assume the correctness of Turvey’s(1992) treatment of affordances, according to which the term “affordances” should be read as referring to dispositional properties of physical objects. However, his influence notwithstanding, Turvey’s definition faces a number of problems (see Stoffregen 2003). Turvey writes for instance that “Dispositions never fail to be actualised when conjoined with suitable circumstances.” (Turvey 1992: 178). By “actualise” Turvey seems to mean that the physical object realises its potential to offer a possible action to some animal. We’ve seen above this requires the presence of an animal with the necessary abilities. Now here’s the trouble. As an animal moves into a particular location in its environment the number of possible actions that are then actualised is enormous. Which of these many possibilities makes it into the contents of perception? Presumably something in the circumstances of the animal is supposed to answer this question, but what this is isn’t immediately obvious.

“Actualisation” could also mean that under the right circumstances the animal cannot fail to act (or more plausibly prepare itself for acting) upon an opportunity for action. The presence of an animal with the complementary effectivities would then be sufficient for the actualisation of the affordance were this amounts to the animal’s exploiting, or preparing itself to act on an affordance. Something like this reading of “actualisation” seems to be at work in McClelland’s Potentiation Condition. According to this condition what it is for an apple to afford eating for example, is for the motor processes involved in eating to be “automatically potentiated” upon perceiving the apple. Now it seems to me we can raise a similar worry. For the apple affords numerous actions in addition to eating such as throwing, kicking, washing, and so on. Are all of these possible actions automatically potentiated when you perceive the apple? If so what makes it the case that you perceive the apple as affording eating rather than some or all the other possibilities it also offers?

This problem can be further brought out by thinking about the differences in how people with Utilisation Behaviour (UB) perceive affordances. Lhermitte (1983) writes that people with UB “without any internal motivation” will inappropriately grasp and use objects when they see them, even when told not to do so (p.253). A person might for instance start to undress upon seeing a bed in someone else’s house. The actions the person with UB performs are inappropriate given their “concrete personal and social context” (Rietveld 2012: 113). McClelland uses UB to argue that the actions that are automatically potentiated but typically inhibited in us are disinhibited or released in the patient with UB. This assumes however that there is something common to UB and ordinary cases of affordance perception – namely automatically potentiated actions. But then we run into Turvey’s problem that in any given situation a person must be automatically triggered to potentially perform an open ended number of actions. How do perceptual systems then make the cut selecting from among all of the many potential actions that are inhibited those that make into the contents of perception as possibilities the agent perceives? Moreover, is it really plausible to suppose that all of the many possible actions a given situation makes available are prepared automatically by the motor system?

I suggest that what is needed to make sense of UB and of affordances more generally in a way of carving of affordances that are relevant to a person from those that are not. The person with UB acts on affordances without “internal motivation” (Lhermitte 1983) in ways that are inappropriate both to them, and to their wider social context. This is a qualitatively different mode of engaging with the environment insofar as the individual’s sense of what matters to them and what is socially appropriate plays no part. I suggest (but cannot fully argue for this here) that dispositional analyses of affordances will have troubles accounting for this distinction. They will face difficulties in accounting for what makes an affordance relevant to an individual.

A rival approach to that of Turvey in the literature on affordances takes the animal-environment system to be the unit of analysis for understanding affordances (see Stoffregen 2003; Chemero 2003, 2009; Rietveld and Kiverstein 2014). Instead of treating affordances as being complemented by abilities, the affordances of the environment are defined in part in relation to animals and their abilities. Turning to McClelland’s Perceptibility Condition, what makes an affordance to φ perceptible for S are S’s abilities. A rock-face is perceived as climbable by a rock-climber S only if S is able and has the skill to make the climb. For others that lack the skill, the rock-face would not be presented in perception as climbable. It is the perceiver’s abilities – their sense of what matters for acting appropriately in a given situation – that determines which of the many affordances a situation offers make it into their perceptual experience. In UB, the person’s motor skills work fine but they suffer from a disturbance in their ability for responding to affordances selectively based on what matters to them in a particular situation (Rietveld 2012).

Let me finish up by returning to the question at the heart of McClelland’s paper – whether an agent can perceive the environment as providing opportunities for mental action. If affordances are relative to skill, this question can be resolved by reference to abilities. The cognitive anthropologist Ed Hutchins (2008) discusses the example of looking at the sky and seeing a constellation. Does a constellation in the sky afford attending to? To see a constellation requires knowing what to attend to, and what to see when attending. It requires attentional skills. On a skilled-based account of affordances such as I defend, it does indeed make sense to say that the constellation can afford attending to, but only for an astronomer that has the complex skills for looking. I would therefore conclude that opportunities for mental action can be perceived by individuals that have the necessary abilities for detecting such possibilities.



Chemero, A. 2009. Radical Embodied Cognitive Science. (Cambridge, MA: MIT Press)

Chemero, A. 2003. An outline of a theory of affordances. Ecological Psychology 15: 181-95.

Gibson, J.J. 1979. An Ecological Approach to Visual Perception. (Boston, MA: Houghton-Mifflin)

Hutchins, E. 2008. The role of cultural practices in the emergence of modern human intelligence. Philosophical Transactions of the Royal Society B, 363: 2011-19.

Lhermitte, F. 1983. “Utilisation behaviour” and its relation to lesions of the frontal lobes. Brain 106: 237-255

Rietveld, E. & Kiverstein, J. 2014. The rich landscape of affordances. Ecological Psychology 26 (4): 325-52.

Rietveld, E. 2012. Context switching and responsiveness to real relevance. In J. Kiverstein & M. Wheeler (Ed’s) Heidegger and Cognitive Science. (Basingstoke, Hampshire: Palgrave MacMillan)

Stoffregen, T.A. 2003. Affordances as properties of the animal-environment system. Ecological Psychology 15 (2): 115-34.

Turvey, M.T. 1992. Affordances and prospective control: an outline of the ontology. Ecological Psychology 4: 173-87.

Invited Comments from Carlos Muñoz-Suárez (Barcelona, Icesi)

Are Mental Affordances in the Landscape?*


Carlos Muñoz-Suárez
University of Barcelona, Spain
Universidad Icesi, Colombia  

“(…) a thing is full of threats and promises.
The dispositions or capacities of a thing (…)
are not less important to us than its overt behavior,
but they strike us by comparison as rather ethereal.
And so we are moved to inquired whether
we can bring them down to earth”

Fact, Fiction, and Forecast (1954: 40)


The idea that there are mental affordances is intriguing, probably, because the notion of ‘affordance’ is intriguing itself. Be that as it may be, an initial degree of perplexity doesn’t prevent us from presuming that the notion of ‘mental affordance’ could guide us within unexplored explanatory fields.

McClelland claims that “although there are murmurings about affordances for mental action, nobody has unpacked exactly what it means for there to be mental affordances or made an explicit case for their existence” (2. Italics mine). McClelland aims to fill in that explanatory blank by providing an initial articulation of what he calls the Mental Affordance Hypothesis (hereafter, MAH) – the thesis that agents are (not only) sensitive (to affordances for bodily actions but also) to affordances for mental actions (1).[1]

McClelland aims to support the metaphysical claim[2] that there are mental affordances by introducing plausibility consideration about MAH. I would spell out McClelland’s explanatory strategy as follows: “if we have evidence that agents are sensitive to affordances for mental action, then we have good reasons to believe that the metaphysical claim that there are mental affordances is true.” I don’t think that this strategy works. Let me explain.

McClelland claims that the metaphysical claim “is ultimately answerable to the empirical data” (11). There are good reasons to put that into question. In looking for empirical data that support MAH we are already presupposing – rather than assessing – the metaphysical claim, i.e., we are looking for something that we assume that exists. However, if the notion of ‘mental affordance’ is not overtly intelligible, the metaphysical claim either. In such a case, one should point out that the semantic vagueness of the notion clouds the empirical assessment of MAH, i.e., one should point out that we are looking for something that we barely know.

My complaint rests on the assumption that empirical data underdetermine the use of ontological categories. Rather, our ontological categories and commitments are the basis for making sense of empirical data. Therefore, addressing the question about the existence of mental affordances seems to be primarily a labor of armchair elucidation than one of collecting empirical data. Notwithstanding McClelland’s apparent empiricism, I think that his paper mirrors this strategy: he firstly introduced a notion of ‘mental affordance’ on the basis of a priori considerations and, then, proceeded to interpret particular empirical cases on the basis of that notion.

So, as a matter of prudence, before making empirical generalizations on the basis of a sketchy notion of ‘mental affordance’, we should try to unpack the notion to see if the metaphysical claim standing behind MAH is (more than prima facie) intelligible.

Like any piece of work that provides an initial articulation of a complex concept, I see McClelland’s paper as requiring some conceptual elaboration. In this commentary I’ll focus in elucidating the notion of ‘mental affordance’ that’s required for the intelligibility of the metaphysical claim. Thence, we’ll be in a position to see if the mental-affordance talk is merely purported to plunge a “new fashion” term into the waves of academic seas or if, otherwise, that talk is the touchstone of a genuinely innovative way of thinking about human psychology.

1. Towards a minimal notion of ‘affordance’. The kernel of Gibson’s ecological theory is the thesis that animals do not interact with action-neutral environmental items. Gibson’s theory was purported to explain behavioral activity in vivo, i.e., as a dynamic ecological flow which is relative to the abilities of the animal and the sources of its niche.[3] According to this view, agency is ecological agency; it is not the product of sequential processes that filter mental representations of (action-neutral) distal, causal sources. The essence of the agent-environment interaction is the complementarity between agentive and environmental potentialities (Gibson 1979).

Gibson’s theory was purported to show that the Classical Sandwich View doesn’t explain ecological agency, i.e., behavioral activity construed as a dynamic flow that shapes the environment in which animals live. According to the Classical Sandwich View (Hurley 2001), animals form sensory representations of action-neutral distal stimuli; such sensory representations are peripheral inputs to internal hierarchies of processing whose outcomes are motor instructions for bodily actions. In contrast with this picture, Gibson claims that ecological agency can be explained without appealing to inner representational processors that regulate the traffic between purely perceptual representations of distal stimuli and purely motor protocols that change external physical items. Thus, explaining ecological agency requires to abandon the view that the ontology of the environment is independent of agents’ abilities. As Dennett claims:

“[e]very organism, whether a bacterium or a member of Homo sapiens, has a set of things in the world that matter to it and which it (therefore) needs to discriminate and anticipate as best it can. Call this the ontology of the organism, or the organism’s “Umwelt” (von Uexküll 1957) (…) An animal’s “Umwelt” consists in the first place of affordances (Gibson 1979), things to eat or mate with, openings to walk through or look out of, holes to hide in, things to stand on, and so forth” (Dennett 2015: 4-5).

Radical Gibsonians think being committed Gibson’s theory requires to endorse Gibson’s anti-representationalism. Thus, they monger the corollary that endorsing anti-representationalism is fundamental for explaining ecological agency. Moreover, Radical Gibsonians will accuse representationalists of not having understood the central theoretical construct that’s endemic to Gibson’s theoretical fields, namely the notion of ‘affordance’.

McClelland is not a Radical Gibsonian. In particular, in contrast with Radical Gibsonians, who claim that affordances are opportunities for action that are exploited without any kind of representational interface or internal processing, McClelland appeals to what can be construed as a minimal notion of ‘affordance’. This minimal notion is compatible with the idea that affordance-perception can be understood in terms of internal representations of opportunities for action – a concession that, as McClelland recognizes, “Gibson would doubtless have resisted” (McClelland 2017: 3).

As far as I see, this concession leads McClelland to overlap conditions for affordancepossession and conditions for the ascription of dispositional properties. I’ll try to carve the ontological categories that I think should be made explicit in order to understand McClelland’s notion of ‘affordance’.

2. McClelland’s conditions for affordance-possession. According to McClelland there are three necessary conditions that are jointly sufficient for affordance-possession:

Opportunity Condition: x affords φ-ing for S only if x constitutes an opportunity for S to φ (3).

Perceptibility Condition: x affords φ-ing for S only if x’s property of constituting an opportunity for S to φ is perceptible by S (4).

Potentiation Condition: x affords φ-ing for S only if S’s perception of x potentiates the process responsible for S φ-ing (7).

3. Ontological neurosis towards the claim that affordances are opportunities for action. McClelland claims that affordances are opportunities for action. According to McClelland, opportunities for action are situations “in which it is possible for a subject to deploy some ability” (3, italics mine) that she possesses. Taking the phrase literally, McClelland seems to be talking about situations which make it possible for an agent to deploy some ability. I’ll call these situations enabling situations. In this way, an agent exploits an opportunity for action when the deployment of one of her abilities is enabled by a corresponding enabling situation of which the agent is a constituent.

Enabling situations have different aspects. As opportunities, they constitutively have a singular for-someone aspect, i.e., they involve particular agents that are able to exploit opportunities (i.e., to deploy an ability in virtue of an enabling situation). Moreover, as ecologically embedded, enabling situations involve ecologically located relations between particular agents and environmental items.

As McClelland highlights, enabling situations involve relations of “dispositional complementarity” (3). Complementarity hereto should be understood as a symmetrical relation between dispositions, such that a disposition, DA, of a particular agent, A, is complementary of a disposition, Dx, of an environmental item, x, only if Dx is complementary of DA, and vice versa. This relation of complementarity is dispositional because DA  and Dx  can be manifested relative to each other.

Dispositional complementarity is ecologically embedded, i.e., it should not be understood as a mere logical relation between dispositions – it is rather actual in the ecological sense. So, DA and Dx are actually complementary only if there’s an enabling situation in which A and x are ecologically related. Strictly speaking, such an ecological relation is not between the manifestations of DA and Dx but between their bearers, namely A and x; otherwise, the relation would not involve dispositional properties but, say, occurrent properties.

As far as I can see, this conceptual clarifications fit well with Gibson’s thesis that affordances are response-independent in the sense that they don’t depend on the actual responses and needs of agents (Gibson 1979, 139 ff.). Thence, opportunities for action, construed as enabling situations, involve ecological relations between non-manifested complementary dispositions. (This doesn’t entail that manifesting complementary dispositions breaks down the dispositional relation. Here it’s worth noticing that, although affordances – as enabling situations – constitutively involve ecologically located relations in virtue of which dispositions can be manifested, such ecological relations would be updated if the corresponding dispositions were manifested (see Turvey 1992 and Dotov et al. 2012, 31).)

Ecological relations are ecologically located (or concrete in the ecological space) in the sense that they are grounded in the constituents of the physical space between A and x. According to Gibson, ecological physics, in contrasts with “physical physics” (Gibson 1979, 140), can specify the physical grounds of ecological relations. In particular, those are the material grounds of enabling situations and, a fortiori, what enables dispositional manifestations. For instance, in the case of the disposition of an environmental item of being graspable, the item has “opposite surfaces separated by a distance less than the span of the hand” (Gibson 1979, 133).[4]

Furthermore, these conceptual clarifications help to make explicit some details about the ontology of affordances that’s seems to frame McClelland’s view. Let’s me state the argument as follows:

(1) Affordances are opportunities for action (as McClelland claims)
(2) Opportunities for action are enabling situations (as McClelland seems to suggest)
(C1) Affordances are enabling situations (as McClelland should overtly assert) (1, 2).

Moreover, if

(3) enabling situations constitutively involve ecologically located relations between dispositional bearers (according to Gibson).
(C2) Affordances constitutively involve ecological relations (C1, 3).

This argument seems to indicate that affordances are relational in nature (see Chemero 2003).

Thence, a first preliminary conclusion: if one accepts McClelland’s claim that affordances are (opportunities for action which are) situations “in which it is possible” for an agent to deploy some ability, then it seems implausible to accept that affordances are dispositional properties and, a fortiori, that conditions for affordance-possession are conditions for the possession of dispositional properties.

I think that these conceptual clarifications regarding the ontology of affordances are in order insofar as MAH rests on a metaphysical claim regarding the existence of a type of affordances, namely mental affordances.

4. Are affordances perceptible? According to the first preliminary conclusion, affordances are the conditions of possibility of perception. Thence, for instance, an agent, A, can perceive x as affording φ-ing only if there is an affordance that enables A’s φ-ing and x’s being φ-ed. I think that this interpretation fits well with Gibson’s original picture, according to which:

“[a]n important fact about the affordances of the environment is that they are in a sense objective, real, and physical, unlike values and meanings, which are often supposed to be subjective, phenomenal, and mental. But, actually, an affordance is neither an objective property nor a subjective property; or it is both if you like. An affordance cuts across the dichotomy of subjective-objective and helps us to understand its inadequacy. It is equally a fact of the environment and a fact of behavior. It is both physical and psychical, yet neither. An affordance points both ways, to the environment and to the observer” (Gibson 1979: 129).

Roughly speaking, affordances are postulated to explain the origins of the gap between objectivity and subjectivity in the space of ecological activity – in other terms, ‘affordance’ seems to be a topic-neutral term. In this sense, affordances (as well as the activity trails that they enable) shape the separation between what is the mind and the perceptible world for a particular agent.[5] The separation is shaped by means of a long-term dynamical, mutual stabilization of the agent’s niche, on the one hand, and of her abilities, on the other (see Cussins 1992, 659 ff.).

Hereto, we obtain a second preliminary conclusion. If affordances are conditions of possibility of perception, then affordances are not perceptible (in the sense of being admissible objects of perceptual states). Rather, environmental items involved in enabling situations are the kind of things that can be perceived as affording certain actions, but the affordances themselves – in virtue of which such environmental items afford are not.

If we rather apply here an inversion of reasoning, we could end up saying that items have affordances because we perceive them as affording. Therefore, we should conclude that affordances don’t explain the origins of the mind-world separation by means of niche construction processes, but that the separation is given and that agents live in a shared objective world that they perceive. There’s no agreement on this controversial points about the ontology of affordances (see Dotov et al. 2012) but, as far as I can see, the core motivation for the ecological theory is rejecting the idea that animals live in an already-made world that they represent.

5. Are McClelland’s conditions about affordance-possession? If my clarifications are right, the analysandum of McClelland’s necessary conditions (i.e., ‘x affords φ-ing for S’) don’t refer to affordances. Rather, they refer to dispositions of environmental items, i.e., to a relata of the ecological relations that ground dispositional complementarity (which is constitutive of opportunities for action and, a fortiori, of affordances).

I think that we should echo McClelland’s emphasis on the relevance of characterizing affordances as opportunities for action, but we shouldn’t treat his necessary conditions as conditions for affordance-possession. If I’m right, being an enabling situation is not a condition for something to possess an affordance, it is rather an individuation condition of affordances.

The notion of ‘affordance-possession’ involves the notion of ‘affordance-bearer’, therefore (given my interpretation of McClelland’s characterization of affordances, i.e., affordances as opportunities, where opportunities are enabling situations) we should agree that the bearers of affordances are ecologically located relations between particular agents and environmental items that possess complementary dispositions. Thus, a third preliminary conclusion: the analysandum of McClelland’s necessary conditions (i.e., x affords φ-ing for S) doesn’t refer to affordance-possession.

6. Two notions of ‘opportunity’. Probably, McClelland grants that his conditions are conditions of affordance-possession because of the weight that he gives to phenomenological considerations.

McClelland claims that, e.g., “for a tree to be climbable for me is for it to stand in a certain relation to my ability to climb: it must be something toward which I can successfully deploy that ability” (3. Italics mine). As far as I see, these sentences (separated by the colon) capture different things.

The second sentence indicates that, in order to be manifested, a disposition of a particular agent requires certain triggering conditions as well as certain physical grounds that maintain a causal chain between the agent and the environment. For instance, the ability of A to attend an attend-able item, x, can be manifested, on the one hand, only if x triggers the manifestation of A’s disposition to attend x-wise items and, on the other hand, only if there’s a physically grounded, ecological relation between A and x that grounds a causal chain between A and x. So, as far as I see, ‘something towards which I can successfully deploy that ability’ refers, on the one hand, to the triggering environmental item and, on the other hand, to the physical grounds that maintain a causal chain between I and the triggering item.

On the other hand, regarding the first sentence, ‘for me’ seems to be used in the epistemic sense, i.e., as ‘for x (to appear) to be F to me…’. The phrase can be generalized to the case of attention as follows: ‘for x (to appear) to be attend-able to me is for x to stand in a certain relation to my ability to attend’. In other words, the sentence seems to indicate that ‘from my point of view, x epistemically looks to be attend-able in case that it is related to my ability to attend’. And this sentence is an instance of a more general condition of the form: “x appears to be such-and-such to a subject S insofar as x is related so-and-so to S”. So stated, the condition is not about being attend-able but about appearing to be attend-able.

A remedy to the inconsistency between these phrases might be to replace ‘for me’ with ‘by me’, thereby switching from an indication to subjective representation to an indication of behavioral capacities. This is a significant point because, as far as I can see, McClelland inadvertently oscillates between objective conditions of affordance-individuation and subjective conditions for the conscious detection of environmental dispositions. This oscillation can be noticed by attending to another notion of ‘opportunity’ that features in McClelland’s paper.

McClelland claims that “the entity x that constitutes an opportunity for action needn’t be an object. A scenario on the street [an environmental arrangement of objects] might present an opportunity for me to cross the road, but here it is the overall situation that presents the opportunity to φ rather than some particular object” (3. Italics mine). In this passage McClelland is not talking of opportunities for action as enabling situations “in which it is possible to me” to cross the road. McClelland is rather saying that certain arrangements of objects can also be relata in the ecological relations that form enabling situations. Such arrangements of objects are perceptible, by contrast to enabling situations. So, strictly speaking, in this case there is a dispositional complementarity (constitutive of an enabling situation) between my disposition to cross the road and the road’s disposition of being cross-able in the situation C (the scenario in the street).

Dispositional complementarity ontologically depends on enabling situations and, merely, contingently depends on (situations construed as) environmental arrangements of objects. Yet in other words, enabling situations may, or may not, involve situations-as-environmental arrangements. The scenario on the street is a mundane perceptible situation whereas enabling situations are conditions of possibility for perception.

I agree with McClelland that arrangements of environmental items as well as particular environmental items can be afforders. Such arrangements of items are situations by which an agent can exercise an ability that shouldn’t be confused with the situations in which it is possible that an agent exercise an ability.

I think that failing to distinguish enabling situations from perceptible situations lead McClelland to think that affordances are perceptible opportunities and, more importantly, that subjective representations of perceptible opportunities entail the perception of affordances. This remarks lead me to other preliminary conclusion: affordances (construed as objective opportunities, i.e., enabling situations) are not perceptible situations.

7. On the elasticity of ‘affordance’. I think that McClelland’s minimal notion of ‘affordance’ (as something of which agents can form internal representations) is too liberal. My point is that it seems pretty difficult (if not implausible) to detach the notion of ‘affordance’ from Gibson’s claim that they are explanatorily and ontologically prior to phenomenology and mental representation.

I’m not a Radical Gibsonian because I could concede that the notion of ‘affordance’ (understood as referring to a kind of enabling situation) can be the touchstone for developing an enactive account of representational content. However, as far as I can see, phenomenological evidence is not the right kind of evidence for ascribing affordances, although phenomenological evidence can be well placed for ascribing dispositional properties, that is to say, for representing environmental items as affording. Having said this, I think that the analysandum of McClelland’s necessary conditions refers to disposition-possession, rather than to affordance-possesion.

8. From affordances to dispositional properties. The verb ‘affords’ and ‘affordance’ don’t have the same meaning. As Gibson highlights, ‘affordance’, in contrast with ‘affords’, “refers to both the environment and the animal in a way that no existing term does” (1979, 127). If one assumes that affordances are dispositional properties, the analysandum of McClelland’s conditions should be read elliptically.

The analysandum can be restated in order to overtly refer to dispositional properties (assuming that ‘affords’ – in the present indicative – cannot play that semantic function). Following Scarantino (2003, 950), the right kind of dispositional properties can be referred by means of the linguistic construct [verb phrase]-able. So, I suggest to use this analysandum: ‘x is φ-able for S’.

9. Restating the conditions. In line with what I’ve said, McClelland’s conditions can be restated as follows:

Opportunity Condition*: x is φ-able by S only if x is a constituent of an enabling situation in virtue of which S can deploy her ability to φ.

Perceptibility Condition*: x is φ-able by S only if x’s property of constituting an opportunity for S to φ is perceptible by S.

Potentiation Condition*: x is φ-able by S only if S’s perception of x potentiates the process responsible for S φ-in.

According to the Opportunity Condition*, x possesses a dispositional property that can be activated by S only if both, x and S, are parts of an enabling situation grounded in an ecologically located relation. Such ecological relation is an opportunity for S insofar as the physical world has contingently set it up, not because the relation is available to S’s perceptual capacities. So spelled out, going hand in hand with Gibson, affordances don’t depend on being perceived.

According to what I’ve said in § 4., ‘opportunity’ in the Perceptibility Condition* should not be understood as referring to affordances (as enabling situations). However,  as far as I can see, ‘opportunity’ in the analysans should be understood as referring to a situation in which “it is possible for S to  φ”, i.e., to an enabling situation. Thereby, I don’t see a redeemable version of the Perceptibility Condition.

10. On the Potentiation Condition*. The Potentiation Condition* says that x has a dispositional property, Dx, that can be activated by S only if S’s perception of x potentiates the manifestation of a corresponding disposition, DS.

I think that here McClelland elaborates a very interesting point that helps us to understand the actualization of affordances, understood as enabling situations. Let me elaborate on this point. In more detail, I think that actualizing enabling situations involves the activation of potentiation loops.

Consider McClelland’s case in which you are crossing a series of stepping stones across a river (14) – this case shapes an enabling situation. Let’s say that an arrangement of stones has the dispositional property of being leap-able by you and that you have the disposition to mentally rehearse a leap in imagination. In this case, there’s a dispositional complementarity that shapes an enabling situation, i.e., you and the arrangements of stones are the relata of an ecologically located relation grounded on the physical properties of the environment.

Here is worth noticing (as McClelland does) that:

“there is […] indirect empirical evidence that imaginings of bodily acts can be potentiated. A wealth of data suggests that the neural underpinnings of an imagined bodily act overlap extensively with the neural underpinnings of the actual performance of that act […] This drives a convincing account of imagined bodily acts as ‘off-line’ performances of bodily acts. If we grant, as I think we must, that the processes responsible for bodily acts are automatically potentiated by stimuli, then we should also accept that the processes responsible for imagined acts can also be automatically potentiated by stimuli. After all, if they are to a significant extent the very same process, then if one can be potentiated it is plausible that the other can potentiated too” (14-15).

When the stones (i.e., the bearers of the dispositional property of being leap-able) are perceive by you, your perceptual state automatically triggers your dispositional property of mentally rehearsing a leap, i.e., the stones indirectly cause the manifestation of your mental disposition to rehearse a leap. That dispositional manifestation involves the formation of motor protocols; some of which are manifested in externally observable bodily movements. Those bodily movements interact with the physical properties of the stones that ground their disposition of being leap-able, thereby triggering their dispositional manifestations. Then such manifestations trigger again your disposition to mentally rehearse the leap. In this way a potentiation loop is activated by a cycle of dispositional manifestations.

The stones’ dispositional manifestations update your previous dispositional manifestations, thence triggering an updated mental rehearse of the leap, which involves an updated set of motor protocols; some of which will be externalized. This potentiation loop will be maintained until you leave the stones behind.

According to this picture, affordances (understood as enabling situations) are actualized and updated through the activation of potentiation loops. Furthermore, potentiation loops help to understand nonlinearly the skilled engagement of agents in their ecological niches. This interpretation is incompatible with McClelland’s view because he takes affordances to be the properties of the stones that trigger the agent’s motor response (6, f.n. 4.). I think that this is a terminological issue that doesn’t jeopardize the Potentiation Condition* – we could simply accept that affordances are enabling situations and the Potentiation Condition* remains equally strong.

I would like to introduce a couple of considerations here. First, by analogy to his analysis of gripping a teapot, one could suspect that McClelland would be willing to accept that the motor protocols triggered by the stones constitute a perceptual representation of the stones leap-ability, such that the motor protocols would be accurate were the stones be lead-able and inaccurate if they were not (9). I think that this view about the assessability of perceptual states that are constitutive of potentiation loops is misleading.

As far as I see, when an agent is engaged in a potentiation loop, she perceives the dispositional manifestations of the dispositions of environmental items that she has triggered, not the dispositional properties themselves. So, if we want to infer some norms of correctness from the Potentiation Condition*, it would be better to say that a motor protocol is accurate when the dispositional manifestations that it triggers in environmental items potentiates the agent’s ability to mentally rehearse the leap, and inaccurate when they don’t. This doesn’t imply that motor protocols are accurate if they maintain potentiation loops simpliciter; this  would lead the agent, e.g., to a behavioral disorder like the “utilization behavior” discussed by McClelland (6 ff.). This norm of correctness indicates the accuracy of motor protocols depends on how much they contribute to the skilled engagement of the agent in the environment, i.e., with how much they help the agent to develop her skills.

11. Final remark. It’s not clear to me whether McClelland is talking of affordances. As I see it, the undeniable value of McClelland’s paper rests on describing how certain mental abilities (like attending, imagining, and counting) are potentiated by perceptible dispositional properties of the environment.

I don’t think that McClelland’s paper requires a notion of ‘affordance’. The argument could work if, instead of affordances, McClelland articulates his discussion around a phenomenology of environmental dispositional properties, and how their manifestations potentiate mental skills. In this sense, neither I think that McClelland’s paper contributes to confirm the metaphysical claim that there are mental affordances, nor do I think that it supports MAH. Rather, I think that his central hypothesis should be the thesis that agents are (not only) sensitive (to environmental dispositions for bodily actions but also) to environmental dispositions for mental actions.[6]



Chemero, A. (2003), “An Outline of a Theory of Affordances”, Ecological Psychology, 15(2): 181-195.

Cussins, A. (1992), “Content, Embodiment and Objectivity: The Theory of Cognitive Trails”, Mind, New Series, 101(404), 651-688.

Cussins, A. (2003), “Content, Conceptual Content, and Non-Conceptual Content”, in Gunther, Y. (ed.), Essays on Nonconceptual Content, MIT Press,133-163.

Dotov, D., Nie, L. and Wit, d. M. (2012), “Understanding affordances”, Avant, 3(2), 28-39.

Dennett, D. C. (2015), “Why and How Does Consciousness Seem the Way it Seems?”, in Metzinger, T. & Windt, J. (Eds). Open MIND: 10. Frankfurt am Main: MIND Group.

Gibson, J. (1979), The Ecological Approach to Visual Perception. New York: Psychology Press.

Hurley, S. (2001), “Perception And Action: Alternative Views”, Synthese, 129(1), 3–40.

McClelland, T. (2015), “Affording Introspection: An Alternative Model of Inner Awareness”, Philosophical Studies, 172(9), 2469-2492.

McClelland, T. (2017), “AI and affordances for mental action”, Proceedings of the AISB Annual Convention 2017, 372-379.

Scarantino, A. (2003), “Affordances Explained”, Philosophy of Science 70 (5): 949-961.

Turvey, M. (1992), “Affordances and prospective control: An outline of the ontology”, Ecological Psychology, 4(3): 173-187

von Uexküll, J. (1957), “A stroll through the worlds of animals and men: A picture book of invisible worlds”, in Schiller, C. (Ed.) Instinctive behavior: The development of a modern concept. New York, NY: International Universities Press.


I thank to the organizers of MindsOnline2017 for inviting me to comment on “The Mental Affordance Hypothesis” by John McClelland. References to the paper will be solely indicated by the page numbers.

[1] For instance, in his “AI and Affordances for Mental Action” (2017), McClelland argues that the notion of ‘mental affordance’ leads to fruitful explanations of deliberating about a choice within the fields of Artificial Intelligence. And in (2015) he introduces the “Affordance Model of inner awareness”, according to which “all conscious states of normal adult humans are characterised by an affordance of introspectability” (p. 2469).

[2] This claim must be distinguished from the explanatory claim according to which the notion of ‘mental affordance’ helps us to understand phenomena that, in themselves, are not affordances.

[3] Gibson claims that a “species of animal is said to utilize or occupy a certain niche in the environment. This is not quite the same as the habitat of the species; a niche refers more to how an animals lives than to where it lives. I suggest that a niche is a set of affordances” (Gibson 1979, 128).

[4] Metaphorically speaking, ecological relations are like “doors” that the physical world “opens” in order to propitiate, on the one hand, the structuring of environments into niches and, on the other hand, the structuring of activity trails (dynamic manifestations of dispositions of the agent) into behavioral skills.

[5]  Strictly speaking, enabling situations cannot be reduced to affordances because the environment of activity also involves resistances (see Cussins 2003, 156 ff.).

[6] I thank Flavia Felletti for her valuable comments.

10 thoughts on “The Mental Affordance Hypothesis”

  1. Many thanks to the organisers of the Minds Online conference for allowing me to present my paper, and to the three commentators for their thoughtful commentaries. I’m going to post separate responses to each commentary in turn. I’m pleased to see the Mental Affordance Hypothesis such lively discussion and am glad of the opportunity to clarify and develop my ideas in a little more detail.


    As I note early in the paper, Kiverstein is one of the few authors who have proposed something in the ballpark of the Mental Affordance Hypothesis. As such, I’m extremely grateful for his thoughts about my argument. Kiverstein’s commentary revolves around how best to define affordances. With Kiverstein, I think this is not an issue of mere terminological preference. Rather, settling on a definition of affordances involves confronting some substantive theoretical issues. One important division in the literature is between those (such as Turvey) who define affordances relative to dispositions to act, and those (such as Kiverstein) who define them relative to abilities to engage in skilled activity. Kiverstein suggests that I adopt the former view and that I ought instead to adopt the latter. In response, I argue that my own position is actually somewhere in between the two positions described. Moreover, I suggest that the considerations that might push one towards Kiverstein’s view can be dealt with in a different way.

    Take the case of the teapot affording gripping for me. Turvey would characterise the teapots affordance of grippability in terms of something like the following complementary pair of dispositions: I have the disposition to grip the teapot in appropriate circumstances and the teapot has the disposition to be gripped by me in those circumstance. My own appeal to complimentary dispositions diverges from Turvey’s in at least two respects.

    First, I use the term to pin down what it takes for something to constitute an opportunity for action. For me, opportunities for action can only be described as affordances if they satisfy further, more demanding, conditions.

    Second, the dispositions to which I appeal in my characterisation of opportunities for action are not quite the same as those to which Turvey appeals in his characterisation of affordances. Where Turvey says that a teapot affords gripping insofar as I am disposed to actually grip it, I say that I a teapot affords gripping insofar as I have the ability to grip it. My later appeal to potentiation seems to confuse matters here. Although I agree with Turvey that affording gripping has something to do with triggering gripping behaviour, it should be noted: a) that potentiation is a matter of readying a certain action rather than of actually making it happen and b) that potentiation has nothing to do with my dispositional analysis of opportunities for action, but is instead a further condition on affordance-possession.

    If I’m defining opportunities to grip in terms of a relation between my dispositional property of having the ability to grip things and the teapot’s dispositional property of being grippable by me, does that mean that Kiverstein and I are in agreement? Not quite. Kiverstein’s understanding of abilities is more demanding than mine. For me, an ability is just a power to do something. For Kiverstein, having an ability means knowing how to participate in some norm-governed socio-cultural practice, such as the practice of oil painting. My understanding of abilities is thus far more liberal (and perhaps leads to a more liberal picture of affordances).

    The real question is thus whether my liberal view is sustainable, or whether I should follow Kiverstein in building in this ‘practice’ dimension. Kiverstein’s main motivation for appealing to practice is that it promises to overcome worries about the proliferation of affordances. Opportunities for action are ubiquitous – just think of all the actions (bodily or mental) that I can perform toward the teapot! It’s wildly implausible that I am perceptually sensitive to all these opportunities, and even more implausible that all of these possible actions are potentiated.

    I take these worries very seriously, and think the threat they present is especially acute with respect to mental affordances. In the paper I suggest that perception and potentiation are selective: we are perceptually sensitive only to a useful subset of opportunities for action, and only a useful subset of possible actions are potentiated by perceived objects. We’ve thus acquired a sensitivity to the teapot’s grippability that we don’t have to, say, its throwability. But, Kiverstein asks, what explains this selectivity? This is ultimately an empirical question. I think skill acquisition is likely part of the story, so to that extent I’m in line with Kiverstein. Where our views diverge, however, is with respect to task-relevance.

    My proposal is that we’re sensitive to the teapot’s grippability because it is generally useful to be so. Kiverstein’s view, by contrast, is that we are only sensitive’s to the teapot’s grippability if is useful for the skilled task in which we are engaged. (The way I present things in the paper actually sounds a lot more like Kiverstein’s view but this is not the position I wish to endorse). In the striking teapot study I cite, we find evidence of gripping being potentiated by the image of the teapot even though gripping is irrelevant to the task at hand. I think this tips the balance in favour of my view. This is not to say that the task in which we are skillfully engaged is irrelevant: it may modulate our perceptual sensitivity and propagation of potentiated motor signals. Against Kiverstein, I would suggest that actions are potentiated in neurotypical subjects in just the same way as they are in UB patients. The difference is that UB patients have a deficit in the context-sensitive downstream processes that would suppress those motor signals in healthy subjects.

    What does all of the above mean for the case for mental affordances? Kiverstein is open to mental affordances insofar as there are norm governed mental skills. I am sticking with the idea that run-of-the-mill mental abilities are sufficient. But Kiverstein’s case of the skilled observer of constellations is extremely useful. The most informative cases of affordance perception are those in which a new perceptual sensitivity comes hand-in-hand with the acquisition of a new skill. I thought such cases would be hard to find for the case of attention, but the example Kiverstein cites may well provide exactly what I need.


    Many thanks to Derek Jones for his illuminating remarks. Jones’s remarks are divided into three sections so I will follow that structure:

    Affordances and Their Relata

    Jones notes a possible disanalogy between bodily affordances and putative mental affordances. In the bodily case, the afforded action is directed toward an object. But mental actions often aren’t directed at the environmental stimuli that trigger them. One option here is for me to concede that this limits the scope of mental affordances and that only those mental acts directed toward mental objects (such as attending and counting) are candidates for mental affordance. My preferred option, however, is to deny that affordances need to be linked to objects. Sometimes it is better to talk in terms of situations affording actions rather than objects. Siegel introduces the case of a situation affording dancing. Here there is no object that gets danced in the way that a teapot gets gripped. Nevertheless, the action of dancing constrained by the situation in a way that maintains a tight link between environment and performance. I propose that many mental affordances will be more like the dancing case than the teapot case.

    This provides a possible way out of Jones’s worries about the stepping-stone case. I think that the imagined act is one of imagining leaping onto that particular rock, so is a mental act directed at an object in the environment. But if that doesn’t wash, then I am happy to say that it is the situation on the river that affords the imaginative act, and not strictly the rock.
    Jones also worries that my criteria might entail that mental objects can afford mental action, and that this will lead to a proliferation problem. The first thing to note is that the perceptibility condition seems to preclude this: affordances are perceptible opportunities for action and the opportunities for action presented by mental objects are not perceptible. I would be comfortable, however, with softening the perceptual condition so that it accommodates the quasi-perceptual awareness we have of our mental states. This would put mental objects affording mental action back on the table which, to my mind, is a good thing.

    Jones raises a concern that any time one mental state prompts another we’ll have a case of mental affordance. However, if we reflect on my conditions this worry can be ameliorated. Take some mental state that triggers a mental action. My (revised) perceptibility condition requires us to be aware of the mental state as presenting an opportunity for mental action. The fact it triggers that mental action is not enough. Furthermore, the potentiation condition requires this triggering to have very particular causal profile: one in which a mental action is automatically readied but not necessarily performed.


    Jones presents a dilemma regarding the potentiation condition. When in the company of trees, I am not generally aware of climbing being potentiated. I must say either that the trees do not afford climbing, or that they do afford climbing and I am unaware of the potentiation. On the first horn, if the tree turns out not to potentiate climbing, and therefore not to afford climbing, that’s fine by me. I doubt those in the ecological school would be happy with such a concession, but this is just one of many respects in which my position isn’t strictly Gibsonian. On the second horn, if it turns out that I am unaware of the potentiation and the suppression, that’s fine too. In fact, I think it’s the more plausible possibility. Jones is happy to accept my interpretation of the toothbrush case for Utilization Behaviour, and thus my claim that perceived toothbrushes potentiate brushing actions in neurotypical subjects. I assume that, like me, when Jones sees a toothbrush in a pharmacist he does not feel the pull toward brushing behaviour and does not experience any suppression of this motor signal. He thus seems to be on board with unexperienced potentiation and suppression. But if these things can occur outside awareness, why do I appeal to phenomenology so much in my case for the potentiation of mental action? I claim that the occurrence of certain experiences make it likely that potentiation has occurred. I deny, however, the converse claim that potentiation having occurred makes it likely that a subject has had those experiences. Potentiation and suppression are, for the most part, unconscious affairs.


    I appeal to an ‘apt-ness filter’ to explain why not all opportunities for action figure in perception and potentiate action. Jones rightly asks how this apt-ness filter is supposed to work. On reflection, I think that my appeal to aptness was not apt. I think we routinely perceive affordances that are irrelevant to our desires and our tasks. Furthermore, the teapot study suggests that actions can be potentiated even when they are totally irrelevant. What I should have said is that we sensitive to those opportunities for action that it is generally helpful for us to be sensitive to. Thus being sensitive to grippability is psychologically useful whereas being sensitive to opportunities to hit-like-a-bongo is, I would hazard, not quite so adaptive. This leaves me with the proliferation worry that Jones so vividly highlights.

    My preferred solution to this problem is actually the Dennettian ‘pandemonium’ model that Jones considers. Regarding perception, I avoid worries about perceptual experience becoming implausibly rich by saying that only a sub-set of perceived opportunities for action make it into perceptual experience. And regarding potentiation, Cisek & Kalaska’s affordance competition hypothesis suggests that our environment triggers any number of motor signals which then compete for propagation. Only the winners of this neurological Darwinian competition make it into consciousness and have a strong influence on our actions.

    This model also allows me to deal with Jones’ worries about cognitive penetrability. On the affordance competition hypothesis the propagation of potentiated motor signals is modulated by our desires and the task we are performing. This explains how affordance perception and potentiation can be sensitive to context but also shows how it can often occur in a way that doesn’t fit with our task, our beliefs or our desires. Motor-perceptual signals modulated in this way aren’t fully cognitively impenetrable but have enough independence from the cognitive to place them firmly in the space of perception rather than cognition.


    Muñoz-Suárez’s commentary applies pressure to my conception of affordances in a number of ways. He suggests that the paper commits me to various claims that are either inherently implausible or else at odds with the core tenets of ecological psychology. I will argue that many of these putative commitments are spurious and that I do not have to adopt the positions with which I am lumbered in the commentary. Furthermore, to the extent that I am committed to positions at odds with the core tenets of ecological psychology I do not consider this to be an objection to my proposal: I am one of many authors who deploys Gibson’s concept of affordances whilst resisting any commitment to a radical Gibsonian conception of psychology.

    Problems of Strategy

    Muñoz-Suárez’s first worry about my argument revolves around the following strategy that he attributes to me: “if we have evidence that agents are sensitive to affordances for mental action, then we have good reasons to believe that the metaphysical claim that there are mental affordances is true.” I’m not sure that this is my strategy. My first concern with this sketch is that I’m not sure that the Mental Affordance Hypothesis is a metaphysical claim. My discussion of the nature of opportunities for action may be construed as metaphysical but this is not motivated at all by empirical claims about agent’s sensitivity to such opportunities. I am confident positing opportunities for mental action because I know that there are mental abilities and that, as dispositional properties, abilities are the kinds of thing that must have complementary dispositions viz. opportunities for the exercise of those abilities. That’s a claim that I take not to need any real argument and which should be sharply distinguished from the question of whether and how we are sensitive to these opportunities for action.

    Of course, to claim that there are opportunities for mental action is not, on my account, to claim that there are mental affordances. Mental affordances require further conditions to be met. But whether these further conditions are ever satisfied – conditions regarding perceptibility and potentiation – is surely an empirical matter. Consequently, Muñoz-Suárez’s objection that I am confusing a metaphysical question for an empirical one does not stick. He claims ‘In looking for empirical data that support MAH we are already presupposing – rather than assessing – the metaphysical claim, i.e., we are looking for something that we assume that exists.’ I do not assume that there are mental affordances. I assume, quite defensibly, that there are opportunities for mental action. I then look at the data to determine whether our sensitivity to these opportunities means that my further conditions of affordance possession are satisfied. I do move from claims about the mind to claims about the world: I move from claims about subjects perceiving things to claims about things being perceptible, and from claims about subjects being motorically affected by things to claims about those things having the power to motorically effect us. I think that such moves are innocuous and follow straighforwardly from the psychological conclusions that I reach.

    Enabling Conditions

    Muñoz-Suárez’s next line of argument concerns enabling conditions. In the context of bodily affordances, an object has the property of being grippable in virtue of certain categorical features (i.e. non-dispositional properties) of its surface and certain categorical features of the span of my hand. Here there is an interesting relational property that holds between the spatial properties of the object and the spatial properties of my hand. Muñoz-Suárez claims that when I talk about opportunities for action I must be talking about enabling situations, and that since enabling situations are not dispositional my dispositional analysis of opportunities is in error.

    In response, I would maintain that both abilities and opportunities for action are dispositional properties that should not be conflated with the categorical properties that ground or otherwise enable them. Put another way, opportunities for action are possibilities for action and not the properties of the subject and her environment that make it possible. The teapot’s being grippable by me does not consist in the teapot having certain categorical surface properties that stand in a certain relation to certain categorical spatial properties of my hand span. After all, if I didn’t have the requisite motor skills this relation could hold without me having the ability to grip the teapot and so without the teapot being grippable by me. A teapot being grippable consists in a pair of complementary dispostional properties: the fact that this is underwritten by certain occurrent categorical properties should not confuse us. It may well have confused some in the ecological school, but that’s no reason for us to make the same mistake.

    Later in the commentary is seems that Muñoz-Suárez may be motivated by a concern that we do not perceive dispositional properties. If affordances are perceptible, they had then better be categorical. Here I’m prepared to double-down: dispositional properties are perceptible, and if it turns out that they’re not then affordances are imperceptible. I think the burden of proof is on a skeptic to say why dispositional properties aren’t the kinds of thing we perceive.

    Bringing Forth Worlds

    Muñoz-Suárez reflects upon the fact that the core motivation for the ecological theory is rejecting the idea that animals live in an already-made world that they then represent. The first thing I’d note here is that I do not share this motivation, so if my account entails that we live in an already made world then that wouldn’t be a problem. The second point I’d make is that my dispositional account of opportunities for action is actually compatible with the idea of a mutual dependence between the animal and its world: as the abilities of an animal change, so too do the opportunities presented by an animal’s environment, thus to the extent that the animals environment is constiuted by opportunities for action the animal and its environment are mutually dependent. As I say though, if claims about mutual dependence have to go by the wayside, I would not consider that an objection to the Mental Affordance Hypothesis.

    Mind Dependent vs. Mind-Independent Affordances

    Muñoz-Suárez also takes issue with my use of the term ‘for me’, claiming that it ‘seems to be used in the epistemic sense’. On this reading, something being climable for me involves me taking it to be climbable. This is not my intention. I state ‘…an entity constitutes an opportunity for action for a subject independently of how the subject takes it to be (Gibson 1979, p. 139). A tree might be climbable by me even if I fail to notice its climbability, and might be unclimbable by me even when I mistakenly take it to be climbable (see Gaver 1991).’ The ‘for me’ serves to emphasise that I am talking about an opportunity for me to climb rather about the tree being climbable in some broader sense.

    Affordances and Phenomenology

    Muñoz-Suárez states: ‘I think that McClelland’s minimal notion of ‘affordance’ (as something of which agents can form internal representations) is too liberal. My point is that it seems pretty difficult (if not implausible) to detach the notion of ‘affordance’ from Gibson’s claim that they are explanatorily and ontologically prior to phenomenology and mental representation.’ First, I don’t say that affordances are things of which agents can form internal representations. I state that my account is neutral between representational and non-representational views of perception. As for priority, my conditions are compatible with affordances being explanatorily and ontologically prior to phenomenology.

    Phenomenological considerations figure in the paper as evidence of our sensivity to opportunities for action, so are in no sense ‘prior’ to affordances. More to the point though, I would reject the claim that our understanding of affordances cannot be detached from Gibson’s claims about explanatory priority. I explicitly distance myself from Gibson’s theoretical ambitions and a great number of those contributing to the affordance literature have done the same. Viable notions of affordances can be offered that eschew Gibson’s commitments, including his commitments regarding explanatory priority.

    Muñoz-Suárez develops his objection by stating that ‘phenomenological evidence is not the right kind of evidence for ascribing affordances’. Again, I think this problem disappears if we are clear about how phenomenology figures in the dialectic. It would be problematic to infer the existence of opportunities for action from our phenomenology, not least because it would be at odds with realist conception of affordances that both Muñoz-Suárez and I agree must be adopted. I am instead using phenomenology as a clue to what is going on a subject’s mind: a window into the content of their perceptual states and the dynamics of their behaviour. This route is, as I emphasise, far from infallible but it can provide us with valuable insights.

    Why ‘Affordances’?

    Muñoz-Suárez concludes ‘It’s not clear to me whether McClelland is talking of affordances. As I see it, the undeniable value of McClelland’s paper rests on describing how certain mental abilities (like attending, imagining, and counting) are potentiated by perceptible dispositional properties of the environment.’ Although I hope to have defended my characterisation of affordances against his arguments, I should note that this verdict would not be so very unwelcome. What I’m interested in is the psychology of mental action. If I’ve made a good case for mental action working in the way I have described then I have achieved my goal. If it transpires that these psychological facts aren’t best described in terms of affordances then this is a result I could bear.

  5. Hi Tom,
    I really like the mental affordances idea, and it sounds prima facie convincing.
    I would like to throw something into the discussion, hoping it would be interesting (it’s in the ballpark of Jones’s worry about attention and relevance, but I think it is a different point).

    The issue is that the notions of relevance and attention are intertwined in (most of?) the psychological literature. You suggest (roughly) that a relevant stimulus affords attention. This sounds prima facie right. But it also seems right to say (very roughly) that attention enables relevance. More precisely, attention (ideally) amplifies task-relevant information and suppresses task-irrelevant information. It allows task-relevant information to impact our mental lives, and blocks such an impact from task-irrelevant information. Perhaps it is even correct to say that the distinction between task-relevant and task-irrelevant information, in cognitive science, makes sense only in light of attentional mechanisms.
    (I’m just waving my hands here. The literature is complicated, there are plenty of disagreements, and I haven’t thought the matter through)

    I’m not sure whether this is a worry. Perhaps your account and the one I presented are two sides of the same coin. Perhaps it makes sense to say: a relevant stimulus affords attention, and attention allows a relevant stimulus to be effective.

    (by the way, what would you say about a task-irrelevant stimulus: would you say that it affords suppression (or inhibition)? It “invites” us to suppress it?)

    1. Dear Assaf,

      Many thanks for your comment. I’ve been grappling with the attention literature lately and think I’d go for your ‘two sides of the same coin’ suggestion. The mechanisms of attention seem to be best described in terms of salience maps. I’d suggest that if something figures in the salience space then it is represented as attendable, and the stronger it’s status in the salience space the stronger the affordance.
      I think that task-irrelevant stimuli can afford attention. In my example of the distracting noise, the noise affords our focal attention despite the fact that the noise is irrelevant to our task. Although the mechanisms of attention help us pick out what’s most important for a task, they also function to highlight other ‘important’ things in our environment that might, for instance, lead us to switch task. Presumably that’s why the irrelevant noise pulls on our attention despite it’s irrelevance to our task.

      Thanks again!

  6. I very much enjoyed this paper. Before reading, if asked what an affordance was, I think I would have guessed that affordances involve the binding of motor representations into perceptual object files. When I see the cup as grippable, my motor cortex’s representation of gripping is combined with the visual cortex’s representation of perceptual properties. This interpretation seems to require that the afforded action be represented in an appropriate way to be bound together with perceptual representations.

    I found the proposal that there might be attention affordances especially striking because I’m not aware of anything that relates to attention (especially bottom-up forms of attention like those involved in your music example) the way that motor or perceptual representations relate to the properties they represent. Instead, I thought that bottom-up attention is more a product of the system as a whole than something that is discretely represented anywhere.

    Your argument for attention affordances relies on the thought that they occupy the general functional role of affordances rather than that they participate in the mechanism that typically implements affordances, and I am curious about the motivation for this. Do you think that there there is a single mechanism behind affordances in all their guises, or is affordance a cognitive kind that can be implemented in a variety of different mechanisms? If it turned out that traditional examples of affordances involve one sort of cognitive mechanism, and the effects that you discuss another, would that alter your conclusion?

  7. Thanks for the question Derek. I think of sensitivity to affordances as a cognitive kind that can be implemented in a variety of different mechanisms. I think the beauty of affordance perception is that it implicates the very mechanisms responsible for performing the afforded action. If, instead, affordances were represented by a kind of affordance-detecting module isolated from the mechanisms of mental and bodily behaviour, this value would be lost.

    Seeing sensitivity to affordances as implemented by different mechanisms also allows me to accommodate the differences between different kinds of sensitivity. Despite offering an argument premised on an analogy between bodily affordances and mental affordances, it looks pretty clear that the two kinds will differ. If the mechanism responsible for attention affordances is distinct from the mechanism responsible for gripping affordances, these differences needn’t present a problem.

  8. Hey Tom,

    Cool paper! I’ve been doing some research in the vicinity, so it’s exciting for me to see you doing good work here.

    Let me attempt a challenge, though. When one first encounters the notion of affordances, it’s natural to ask whether affordances are genuinely perceived. Perhaps there is strong motivation for an affirmative answer to that question if you endorse the full Gibsonian/ecological package, but, as you’ve noted, you don’t want to commit to those other Gibsonian ideas, like anti-representationalism, optic flow, etc. So, you will want strong, independent motivation for the existence of affordance perception. I, myself, have reservations about phenomenological motivations for at least two reasons. First, it’s not easy to establish that the experience of being solicited to act on an affordance is perceptual, and, second, even if you can establish that, it’s a further question what kind of action-guiding role, if any, those solicitation experiences play. It’s not clear, in particular, that they play the role they seem, introspectively, to play.

    However, I think the notion of affordance perception receives significant empirical support from the “vision for action” component of the “two visual system hypothesis”. The visual states that are thought to be responsible for online action-guidance appear to code visually represented object egocentrically and in terms of the movements required to intervene on it. That can look awfully affordance-like, since the claim is that the object is visually represented in relation to the agent’s possibilities for action. So, actually I’m optimistic that affordance perception can be defended empirically.

    But now the challenge is whether there is anything like that kind of empirical support for the mental affordance case. Other features of the two visual systems hypothesis might lead one to think the answer here is ‘no’. For, it is the other (ventral/vision-for-perception’) stream that is thought to inform cognition. So, the challenge is (roughly): the strongest evidence for affordance perception comes from the discovery of vision-for-action, but that evidence seems restricted to bodily action.

  9. Hi Aaron

    That’s a great challenge. I too have reservations about relying on introspection too much here. My own view on the content of perceptual (and this is developed independently of what I say about affordances) is that you need to have empirical and introspective evidence working in tandem to make a case for some property being perceptually experienced. The empirical evidence might suggest some property is perceived but this doesn’t mean it’s necessarily perceptually experienced. Introspection might tell us that something’s happening in experience but, as you say, might be a poor guide as to exactly what is being experienced and as to whether it’s experienced perceptually. Working empirically and introspectively, we might avoid each problem.

    But where does this leave things with affordances? It means I have to concede to you that introspective table-thumping is inadequate, and that I need introspective support for the mental affordance hypothesis. I hope to have given some anecdata that’s kind of proto-empirical but I basically think this stuff needs to be investigated properly before we can really say there are mental affordances.

    I feel the pull of your challenge regarding vision-for-action. I think the best empirical case for bodily affordances will almost certainly draw on these ideas and also concede that this kind of data won’t help make a case for mental affordances. But here’s why I’m not too worried: it’s an open possibility that the mechanisms responsible for mental affordance perception are quite different to those responsible for bodily affordance perception (this comes back to my response to Derek’s comment above). On the one hand, I need a close link between bodily and mental affordance perception to motivate the hypothesis, but on the other I’m free to say that there are certain respects in which mental affordance perception is quite unlike bodily affordance perception. I think that’s what I’d have to say in response to your challenge.


Comments are closed.