Visual adaptation and the purpose of perception

Abstract

What is the purpose of perception? And how might the answer to this question help distinguish perception from other mental processes? Block’s landmark book, The Border Between Seeing and Thinking, investigates the nature of perception, how perception differs from cognition and why the distinction matters. It is, as one would expect, wide ranging, deeply informed by relevant science and hugely stimulating. Here, we explore a central project of the book – Block’s attempts to identify the features of perception that distinguish it from higher level cognition – by focusing on his suggestion that such features closely relate to perception’s purpose. As well as offering a detailed critical discussion of these proposals, our more general aim is to advertise both the promise and pitfalls of asking: what is perception for? Many theories, hypotheses and explanations in vision science appeal to the evolved purpose(s) of perceptual systems. Consider three examples. First, Goodale and Milner motivate their highly influential two visual systems hypothesis by appealing to the evolutionary benefits of separate processing streams: ‘It seems plausible from a functional standpoint that separate processing modules would have evolved to mediate the different uses to which vision can be put’ (1992: 20). Specifically, they contend that we should distinguish two such systems: an evolutionarily more ancient system whose function is to guide visuomotor action, and a more recent system that constructs a stable, conscious visual representation of the world, for use in memory, planning and decision-making. Similarly, Xu introduces her related two visual systems account by noting two competing functions of visual processing: ‘Visual information processing contains two opposite needs. There is both a need to comprehend the richness of the visual world and a need to extract only pertinent visual information to guide thoughts and behavior at a given moment’ (2018: 312). Xu’s hypothesis is that one system is ‘invariant’ (constructing a detailed model of the world independent of the perceiver’s goals or intentions) and the other ‘adaptive’ (representing only salient and task-relevant information to inform decisions). Second, both proponents and critics of the cognitive penetration of vision appeal to functional considerations. Bhalla and Proffitt (1999) notoriously claim that wearing a heavy backpack makes slopes look steeper, arguing that such an effect dissuades people from climbing slopes beyond their physiological potential (though see Durgin et al. 2009, Firestone 2013). Similarly, Balcetis and Dunning claim that desirable objects (e.g. chocolate) are seen as closer than undesirable objects (e.g. faeces), proposing that ‘these biases arise in order to encourage perceivers to engage in behaviors leading to the acquisition of the object’ (2010: 151). In contrast, Gilchrist offers a functional argument against penetration, likening an encapsulated architecture to a ‘free press’. His point is that visual (like public) information is needed for myriad, perhaps unpredictable purposes, and that distorting it for one end may leave the well ‘poisoned, with serious damage to other functions’ (2020: 1002). For example, exaggerating slopes and heights may mislead perceivers who intend to use a hill as a landmark for later navigation or to escape a flood (Firestone and Scholl 2016). Finally, many theorists offer functional arguments for the veridicality of perception.1 For instance, Palmer writes: ‘Evolutionarily speaking, visual perception is useful only if it is reasonably accurate’ (1999: 6). Against this, Hoffman et al. (2015) make the striking claim that perception’s goal – guiding adaptive behaviour – supports an opposing view on which our percepts are wholly non-veridical, deliberately hiding objective reality and instead offering an easily engageable interface (for criticism, see Berke et al. 2021). Block joins this tradition in outlining his approach to determining the border between perception and cognition. He begins by refining our common-sense grasp of the distinction, seeking out ‘scientific indicators that make sense of the pretheoretic classifications’ (33). Such markers offer empirical purchase on the joint between perception and cognition, and position us to ascertain its fundamental nature. Block offers five such indicators: rivalry, pop-out, illusory contours, processing speed, but, first and foremost, adaptation – the phenomenon whereby perceiving a given stimulus feature (e.g. blue) temporarily biases perception away from that feature (e.g. towards yellow; though see below for refinement). He singles out adaptation as ‘the main scientific indicator ... of what is perceptual and what cognitive’ (33) and ‘the most useful of the methods’ (61) for distinguishing perception from cognition. For Block, although not constitutive of perception, adaptation represents ‘a basic feature of perception that is present in all known perceptual systems’ (102). In an especially tantalizing passage, he suggests that these claims are connected to the purpose of perception: to deliver ‘news’. What do the indicators have to do with what perception is at the most fundamental level? Are they mere symptoms of perception or are they more deeply connected with what perception is? In the case of adaptation, one connection is very likely evolutionary. The evolutionary purpose of perception is acquiring information about what is happening here and now. Call that “news”. Any feature of the visual system that produces a constant effect has to be filtered out in order to focus on news. This evolutionary explanation of adaptation is commonly cited. “Sensory adaptation allows us to tune out stimuli that do not provide us with new information needed to cope with the environment. This is the property of adaptation that is generally used to define adaptation in textbooks” (McBurney, 2010, 406) ... By contrast, for cognition, facts that are not news are important too. We want to continue to know that tigers are dangerous, for example. In short, the psychological indicators for distinguishing the perceptual from the cognitive ... are closely related to the different functions of perception and cognition. (119–20) Here, we explore these ideas in more detail. We first look critically at the claim that adaptation is a key indicator of perception. Then, we turn to the connection Block draws between adaptation and the putative evolutionary purpose of perception, namely news. We cast doubt on the idea that the function of news acquisition distinguishes perception from cognition, and so question its usefulness in determining the markers of perception. More positively, we end by considering what light might be shed on other putative features of perception by reflecting on its purpose(s). What is perceptual adaptation (Figure 1)? Sceptical of strict definitions, Block proposes that adaptation is a natural kind, offering a paradigm example: the motion after-effect. A classic motion after-effect occurs after staring at continuous motion in one direction (e.g. leftward) for an extended period. Immediately afterwards, stationary stimuli in the same location will appear to move in the opposite direction (e.g. rightward). This example is paradigmatic in being short-lived (lasting seconds to minutes), repulsive (biased ‘away’ from what was experienced earlier) and retinotopic (specific to the retinally defined location of the original stimulation; Knapen et al. 2009). Importantly, however, not all adaptation exhibits these features. As Block notes, the McCollough effect, in which one adapts jointly to colour and orientation, can last months (72), and sometimes adaptation can be attractive (73). Nor is all adaptation retinotopic: some adaptation effects are spatiotopic (specific to an environmentally defined location, rather than the retinally defined location of earlier stimulation; and and many higher level effects are and of visual In the relevant effects can be experienced by first staring at for and to after staring at in the in the appear in the the in the after move the in the the in the appear The McCollough after staring at and appear and appear Here, the after staring at the two in the of which is and one of which is not very the two in the look different from one – the one than the after staring at the and in the the in the look different – the more than the from and 2009, and et al. For adaptation to its putative in distinguishing perception from cognition, it be both and as an indicator of perception. In other of perception adaptation be at as of adaptation perception. Consider adaptation is not to be constitutive of perception rather a of its Block’s view can of adaptation would be a of perception if cognitive adaptation – for example, if considering or some property as the of an later to be temporarily than they would This to a serious which on adaptation – that a of perceptual adaptation as to in the motion after-effect. who to motion to stationary stimuli as – a might – is this they the stationary stimuli as or is it that their for a stimulus to be has Block offers two to this and the Block effect has to be retinotopic or many perceptual adaptation effects are retinotopic or spatiotopic to some effects can be We have three about this First, as (2015) and Block in and this approach be to putative perceptual adaptation effects (e.g. as or these can be and Second, we that is against mere and Firestone it is to see why effects be are in and who are to one of two can their or et al. of stimulus location spatiotopic it would that a a spatiotopic the motion after-effect. to their for motion they have staring not in other This would the basic of of adaptation, in Finally, as or is of perception, it is how central or useful an indicator adaptation effects may in that an effect is retinotopic will to it as perceptual at it adaptation or For instance, are et al. 2010, et al. 2013). might be to be perceptual on the that they are retinotopic – they are not adaptation In other not adaptation, is the perception from cognition. Block’s to the of out in to Consider an to two a more one on the and a one on the at two (Figure on such a Block as to an of In the it as if are more on the than on the have these in many and have to the that have not with a that with more on the and to of There is to effects to This is not the of a It is a effect that can for the of This point might on in the of of a should be by What should we make of this appeal to our of adaptation is we have so a on the of a what would it to a and in It is not to us that such a not or be Nor is it that such a would not be natural to 2010, 2016). a the of we appeal to to distinguish perceptual adaptation from it is what adaptation is it seems that our grasp on the in our the heavy This our earlier about for Block, adaptation is to be ‘the main scientific of perception. if in it out that other features – features which are independent markers of perception – are needed to that the effect is the of adaptation is cast we have explanations of putative perceptual adaptation in of of our fundamental is with the of adaptation as a of perception. As Block writes: adaptation is to be useful in distinguishing perception from cognition, would have to be adaptation for perception not Block that are of cognitive effects would a of an of cognition and so is by a cognitive A example is (Figure to and biases in to do of a in the of and to in to the last For instance, if to a of in the a might if in a but, the will a to or or these effects are such that the one at a given biases the of in the not on the of or as cognitive In by and are to at The the from a In are very to the same by the central very likely to to their For example, after are to on the to or they are to only a (e.g. with or more likely than other In and the that are to is by a which the of visual and provide a model of in of and – a model for two and For our purposes, is key it is the of In that in the of the It this by towards the of the most recent the of the same on the and that this can the of in an which produces a the to of The for in the more recent on as a the to in The of in not to the both are however, is that as by and is by which are as cognitive as we see the of leading to both and is at in perception. is not a adaptive that this example of cognitive adaptation would to Block’s to other For instance, Block example of after first about very he suggests that this be to the to the the in The is not towards the of the of rather in direction away from the given from the Block putative example of cognitive adaptation, et in which as on such that a to than they Block that explanations of these (e.g. in of to ‘the of as well as its it seems natural to model these as a the explanation will have the of the that occurs in with well the more of and the of classic There is to that cognition its adaptive effects in may well be other of cognitive adaptation, to as in their relevant example might be adaptation and whereby an to or such as a in this phenomenon the of after the a might temporarily after some that would have are experienced as than they Similarly, whose for the might in that most their or for cognitive adaptation is the same for an extended makes that or and may make other more to rather than et al. though such are these are not they are at plausible of adaptation, they are not plausible of perception, and features such as that adaptation of as perception’s scientific his discussion of cognitive adaptation, Block the us doubt to that is cognitive We can distinguish cognitive adaptation from perceptual adaptation by of the features that have of perceptual features and other to of or adaptation, the being a function of stimulus and a of and in of visual and Block’s view is that are markers of perception which to out perception, Block by adaptation and on the of other by on other features to a for perception, Block would be the idea that adaptation is ‘the most useful of the methods’ we have for distinguishing perception from cognition. in appealing to highly features of perceptual adaptation, or to other Block to more about how such features should be and and a a general joint between perception and cognition. we have adaptation is a indicator of perception. is it a adaptation in all known perceptual systems’ As by Block, one of us that are perceptual may not For example, it seems that we can see objects as to our as being to our make a central appear to our with objects as and many There may be features of the different perceptual which do not (e.g. the of the visual these Block suggests that may In these a the perceiver’s point of view to the or such that objects that would have appear of and objects that would have of appear the are after the visual world in the other such that a who is to point at a will point of it is that a repulsive adaptation with to the perception of and First, one might that the effect is a of visuomotor as seems to be the case with et al. Second, it is how to the original about many on one do not the to see many on one they a the visual Block provide that adaptation occurs for et al. 2009). he that this adaptation for in et al. some in that though only for of objects a or and only for not for to stimuli other stimuli appear to stimuli not make other stimuli appear more Block is to that we know the of adaptation in is Block’s appeal to the purpose of perception may As we in the Block offers a evolutionary for adaptation to be a general perceptual that this argument begins with the idea that ‘the evolutionary purpose of perception is acquiring information about what is happening here and – what Block – and suggests that adaptation is a for out constant effects order to focus on In contrast, cognition, facts that are not news are important In this in distinguishing perception and cognition is related to the different functions of perception and This is an – and – of There is, of a what can be about and cognition by it as the of natural (for different see and this has not with the of and informed visual perception which Block in to such as or cognition, hypotheses about vision can be a wide of Block’s a to by reflecting on the evolved purpose(s) of perceptual systems. the promise in such to on our about perception, we see pitfalls with its present The central is that perception and cognition have evolutionary purposes, – in both – the acquisition of ‘news’. In that appeal to news acquisition to the or of adaptation as a of perception. are the for evolved systems. have a It functions to it us from and in some it our sense of it can and it may other have a In are used for and in other we might and too. As a of ... is used in such a of and from and of such as and We might many perception and allows us to and and so from his that of such different in different is that all use the same in different for different We to a A of this of can be in et of adaptation et al. offer a highly evolutionary explanation for the adaptation effects they This explanation is on the between which and so are relevant to and which are not out of This is a highly and and we should not it to may be it would be to the of is the adaptation This all suggests that we should be to highly and in given the of different perception may news acquisition be an general purpose of answer this we need more as to what is by ‘news’. news to important information about the here and it will be to that perception about news. The however, is that it is why the purpose of acquiring news in this sense should a visual system that exhibits visual is a repulsive effect a feature Such a will only news acquisition if we as to in the environment. Many of the by do with a focus on (e.g. and facts that are not are important for perceptual on a a will to an on her its location at the of her her if the location is It would be a to her out of the it is In a may to the position of some feature of the so it can it is to Similarly, might a on a so can see the of the and are in it in a visual a or only on Here, the is not are it to – and it is to continue perceiving may arise we an or – for example, determining a is or some would make a In such we may us for many and key visual we want such features to staring at Block may well with this and that his sense of only to this us to the question of why news acquisition in this sense should In other Block a as about what is happening here and and on the idea that this purpose perception as by or more as to a feature the of the claim that news is a general purpose of perception. who make an appeal to news in the sense of is In his related between vision and a he claims that have evolved to (2020: 1002). this Gilchrist that the of visual to ‘a to a visual this seems in two First, with the last of vision science has to be at the idea that perception’s goal is to us The and highly influential on and are to that is highly and – and it would be for perception to be so at out its Such more with a system which has as one as it the and benefits of these different to this idea Second, is of adaptation, the visual system of the which Gilchrist and such a in their influential of (Figure in visual perception. all effects in vision are In by and a to the of a The that recent stimulus can than perception, such that on the are towards the seen on the on and with only A function of vision is important in the and adaptation in to ... is a for visual to is a to the the world is and continuous and to the recent a of the The visual system may the need to to with the to on the of the environment. It may be to that the present visual is to the one seen Here, and that is one important goal of they point out that with this goal in with is to In with this, and in Specifically, perception of stimuli towards have in in et al. with our suggestion that perception has purposes, and that and adaptation effects competing goals of the visual In they make the important point that perception is in may not relevant in the rather or it if perception on it would not that adaptation is the only relevant information in more may to a more and view of the and that and at different and the visual it is not that both and will in all for all and For instance, et al. only for and in perception, and for and effects for have the to the of a and et al. a effect has for and may be it is to see why the for adaptation in perception would not in many of cognition. Consider why would it not be important to be to in A in and might be to or such as perhaps adaptation to position or of leading to and if perception is news this not provide a for the that adaptation is of perception. should Block to these would be to as feature of this would the of the claim that was in its As et al. in are not to visual perception a wide of cognitive memory, in and This at our cognitive system is and to the recent of and would be to on adaptation as the of perception and that is a visual There is for a model in which adaptation is a and is a et al. et al. et al. 2021). It would be an if all it is more plausible that and such at both perceptual and is for retinotopic on the model by et al. and et al. the of is higher its is perception. is, such in in perceptual The is a in perception. one is that adaptation is a of to visual on this it will not be a of perception Block other features of perception with of perception is the perceiving has to It do for perception to in as with the more of This is an it is how a of perception is with the has that to be as on a continuous or which the case is to some between and et al. et al. which is to on a model is that to make a in (Figure For instance, to a from a of who on their first will do than on their Such suggests that perception contains more information than a in in the case of be a to or it is why one a which if is a of a being one not need a to given a after an on a who the three well are to on of perception. these critical we are that of purpose in perception will provide In we that central feature of perception is very closely connected to purpose. A of Block’s book is that is as with cognition, which is It is to motivate this feature by appeal to perception’s function as theorists have a functional case for the of Xu’s that vision in comprehend the richness of the visual or not all perception is and it might well be that this purpose is by The critical idea here is that are especially well to information about features of the – a is a for a different functional argument for see Xu’s might be with an idea of to make a different Gilchrist that cognitive penetration is visual information has to many purposes, and that these are by a the same idea that vision many – which may not be in – suggests that would be in a of information and only more and to Such an idea has with Block’s on are and which we explore we end with an to Block and other should we the purpose – or – of perception? can we the of in so (e.g. and the of distinguishing present functions from the for which the evolved to And how do these relate to the indicators and constitutive features of This was in by to

Topics & Concepts