Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Perception without processing : J.J. Gibson's ecological approach Smart, Brent Maxwell 1988

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-UBC_1988_A8 S57.pdf [ 6.78MB ]
Metadata
JSON: 831-1.0097806.json
JSON-LD: 831-1.0097806-ld.json
RDF/XML (Pretty): 831-1.0097806-rdf.xml
RDF/JSON: 831-1.0097806-rdf.json
Turtle: 831-1.0097806-turtle.txt
N-Triples: 831-1.0097806-rdf-ntriples.txt
Original Record: 831-1.0097806-source.json
Full Text
831-1.0097806-fulltext.txt
Citation
831-1.0097806.ris

Full Text

PERCEPTION WITHOUT PROCESSING: J . J . GIBSON'S ECOLOGICAL APPROACH By BRENT MAXWELL SMART B.A., The Universi ty of Calgary, 1983 A THESIS SUBMITTED IN PARTIAL FULLFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF ARTS in THE FACULTY OF GRADUATE STUDIES DEPARTMENT OF PHILOSOPHY We accept th i s thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA August 1988 .® Brent Maxwell Smart, 1988 In p resen t ing this thesis in partial f u l f i lmen t o f t he r e q u i r e m e n t s f o r an advanced d e g r e e at t he Univers i ty o f Brit ish C o l u m b i a , I agree tha t t h e Library shall make it f ree ly available fo r re ference and s tudy . I f u r the r agree that pe rmiss ion f o r ex tens ive c o p y i n g o f th is thesis f o r scholar ly pu rposes may be g ran ted by the head o f m y d e p a r t m e n t o r by his o r her representat ives. It is u n d e r s t o o d that c o p y i n g o r p u b l i c a t i o n o f th is thesis fo r f inancial gain shall n o t be a l l o w e d w i t h o u t m y w r i t t e n pe rm iss ion . D e p a r t m e n t o f Philosophy The Univers i ty o f Brit ish C o l u m b i a Vancouver , Canada Date A u g u s t l f 1 9 8 8  DE-6 (2/88) Abstract The ecological movement in the psychology of perception, founded by James J . Gibson, hold that t rad i t i ona l approaches to perception are based upon cer ta in fundamental mistakes. The chief one, ecological theor is ts c la im, is that perceptual information pickup consists of the appl icat ion of cer ta in cogni t ive processes to sensory input which is not speci f ic to features of organisms' environment. Gibson's fundamental claim is that perception does not require the processing of some form of sensory input. In th i s sense, the ecological approach is said to be a theory of d i rec t perception. An important debate over the Gibsonian view concerns the question of whether or not perceptual information pickup without cognit ive processing is a coherent not ion. Among the more recent wr i te rs who claim that the ecological view w i l l not work as i t stands are Jerry A. Fodor and Zenon W. Pylyshyn. They claim, essent ia l l y , that Gibson's approach has no means fo r accounting for i n t e n t i o n a l i t y . Fodor and Pylyshyn are answered by four prominent Gibsonians who claim such c r i t i c i sms are u t t e r l y baseless. These ecological t heo r i s t s , Michael Turvey, Robert Shaw, Edward Reed, and Will iam Mace endeavour to show how the i r approach can indeed account fo r i n t e n t i o n a l i t y . This debate between Fodor and Pylyshyn on the one hand, and Turvey, Shaw, Reed, arid Mace on the other is a perfect example of the kinds of misunderstandings that have arisen between Gibsonians and proponents of t rad i t i ona l view. In t h i s thes is , I supply a detai led descr ipt ion of Gibson's model as i t re lates to the issue of how i n t e n t i o n a l i t y could survive perception without processing. Fodor and Pylyshyn's understanding and assessment of the Gibsonian posi t ion w i l l then be examined. Although these defenders of t r a d i t i o n a l views have, some important concerns,.they also seem not to have a proper grasp.of,:some Gibsonian concepts. In p a r t i c u l a r , Fodor arid Pylyshyn have an unsat isfactory grasp of the notion of an invar ian t . There are more serious misunderstandings evident in the response to Fodor arid Pylyshyn given by Turvey et a l . I point out that these ecological theor is ts have d i f f i c u l t i e s with philosophical terms and theories they employ in defense of Gibson. As a resu l t of evident confusions over notions of in tension, extension, and property, arid confusions over the nature of Fred I . Dretske's theory of natural laws and H i l l a r y Putnam's theory of natural kinds, Turvey et al do not manage to show how Gibson's approach could account for i n tens iona l i t y . I conclude by suggesting that the ecological approach nevertheless is compatible with the idea,of analyzing perceptual information pickup in terms of behaviour, or d isposi t ions to behave. On such an i n te r -p re ta t ion , the ecological approach is s imi lar in many important respects to the D.M. Armstrong's philosophical theory of perception. The compar-ison provides ecological theor is ts with a precedent as well as ph i lo -sophical model to consu l t . in order to better, understand the philosophical language and terminology.'-..On the other hand, the comparison with Arm-strong provides philosophers of perception wi th a means fo r approaching Gibson's view and the problems with which i t w i l l he confronted. iv Table of Contents ABSTRACT i i ACKNOWLEDGEMENT v Chapter I 1 ( i ) • 2 ( i i ) 6 ( i i i ) 11 ( i v ) 21 I I 26 ( i ) 26 ( i i ) • 28 ( i i i ) 35 ( i v ) 44 I I I 48 ( i ) 48 ( i i ) 64 IV 67 ( i ) 67 ( i i ) 70 ( i i i ) 71 ( i v ) 77 (v) 80 V 89 ( i ) 89 ( i i ) 96 VI 109 ( i ) 109 ( i i ) 114 ( i i i ) 125 ( i v ) 127 SOURCES CITED • 137 Acknowledgements I would l i k e to express my appreciation to Dr. Gary Wedeking and Dr. Steven Savi t t fo r t he i r considerable help in the development of my thesis top ic , and for the many hours they have spent ca re fu l l y reading d r a f t s . I would also l i k e to thank my parents, Max and Sylvia Smart, and my wife Louise, fo r a l l t he i r support, material and otherwise. F ina l l y , I would l i k e to express my grat i tude to G.M. Grieg, professor of philosophy (emeritus) at the Universi ty of Calgary fo r a l l that he taught me about philosophy. His unequalled dedication to his students w i l l be an example to me always. Chapter One James J . Gibson's ecological approach to perception^'s an approach s t i l l very much undergoing refinement and hard c r i t i c a l scrut iny . Gibson claims to have begun developing the theory shor t ly a f te r the Second World War. A number of well-known psychologists continue today to expound Gibson's basic view and to attempt to contr ibute more deta i l to i t . Gibson's approach is taken by i t s founder and his fol lowers to const i tu te a new form of d i rec t perceptual theory. Some even suggest that i t is a var ie ty of d i rec t real ism. A l l Gibsonians regard the outlook as. rad i ca l , as being an important divergence from v i r t u a l l y a l l previous perceptual theor ies. A few of them have even speculated that the ecological approach's supercession of p r io r t r a d i t i o n a l views would amount to a 2 Kuhnian-style revo lu t ion. Others who are rather less indiscrete in the i r predict ion (Gibson being one of these) are nevertheless at pains to emphasize the alleged radical nature of t h e i r ideas. Jerry A. Fodor and Z.W. Pylyshyn believe there is a way of reading the ecological theory which reconciles i t with the "Establishment" (Fodor and Pylyshyn's term for t r a d i t i o n a l t h e o r i s t s ) , but go on to note quite r i g h t l y that Gibson does 3 not want his views interpreted in such a conc i l ia tory fashion. This f i r s t chapter is one ha l f of an out l ine of the ecological approach. I w i l l endeavor to introduce Gibson's view in a very general way. The emphasis w i l l be on his mot ivat ion: what he believes are the mistakes of the Establishment and why. Detailed questions relevant to the phi los-ophical aspects of the Gibsonian Theory w i l l be dealt wi th in subsequent chapters. 2 (1) Gibson c i tes f i ve important points over which his view d i f f e r s from 4 that which he regards as the t r a d i t i o n a l approach to perception : 1. The notion of perception is conceived as the pickup of information. To perceive is to pick up information about oneself or one's surroundings. 2. The categories of perception are reconceived. Gibson suggests that organisms perceive t he i r surroundings predominantly in terms of u t i l i t y . Organisms perceive t h e i r environment in terms of what i t a f fords. 3. There is a new view of the basis of perception (Gibson often refers to th is as " information for" perception). This is Gibson's notion of an " i nva r i an t " , or invar iant st ructure of ambient energy. 4. Perceptual systems are taken to be hierarchies of organs rather than indiv idual organs or banks of receptors. These sets of organs funct ion together in order to f a c i l i t a t e the pickup of information. 5. Perceptual systems must concurrently reg is ter persistence and change in the f lux of ambient energy to which each system is sens i t ive . 4) and 5) are relevant to the main issues wi th which th is thesis w i l l be concerned although they are not cen t ra l . To the extent that they require e laborat ion, they w i l l be explained in Chapter two. Chapter two w i l l deal far more with terminology of the ecological approach. Much of i t w i l l be devoted to 3) since the concept of an invar iant is both a crucia l and d i f f i c u l t idea of Gibson's. As with 3 ) , 2) also pertains to a showpiece, so to speak, of the ecological approach. To explain what i t means to say that an organism perceives i t s environment in terms of u t i l i t y requires an explanation of the notion of an "affordance". That w i l l be one 3 of the la te r tasks of th is chapter. F i r s t i t is necessary to expand and comment on 1 ) , however. The assert ion from which Gibson's theory grows is that to perceive is to pickup informat ion. For a number of reasons, the focus here w i l l mainly be on visual perception. Not least of these reasons is that Gibson's last book, in which his new approach is most developed, is wr i t ten en t i r e l y about v i s ion . C la r i t y of exposit ion is also f a c i l i t a t e d by thus constraining the discussion. One should not, however, take the concentration on v is ion to be a t a c i t assumption that an explanation of visual perception i s , with only terminological adjustments, a sui table account of the operations of other modal i t ies. Gibson supposes there are always two d i f fe ren t sorts of information avai lable for pickup: information about the environment and information about oneself. These two var ie t ies are correlated with his terms "exteroception" (perception of the environment) and "proprioception" (perception of one's own body). As one examines the ecological approach, i t should become apparent that the two kinds of information pick up are very much interconnected. The information one can pick up about one's own surroundings is p a r t i a l l y contingent upon propriocept ion. Nevertheless, my discussion w i l l be concentrated on exteroception since th is is the kind of perception which is t y p i c a l l y . o f in teres t in the philosophy of perception. Thus, as i t is to be considered in subsequent chapters, perception w i l l mainly re fer to the pickup of information about the environment via visual systems. A very important point to make about Gibson's theory is that inform-at ion pickup is an epistemic not ion. Perceiving considered as information pickup is epistemic perception. Epistemic perception occurs when and only when the process of perceiving y ie lds knowing, be l iev ing , judging, or the l i k e ( that i s , some kind of epistemic s ta te ) . I t is to be dist inguished 4 from so-cal led "simple perceiv ing" ,^ which is not the acquis i t ion of some epistemic s ta te . The key dif ference between the two is that perceiv ing, in the epistemic sense, denotes an in tent ional re la t ion whereas simple perceiving does not. I t is important to remember that Gibson's approach is an account of epistemic perception, because a large c r i t i c i s m of the ecological approach w i l l turn out to be that i t cannot explain the in tent ional component in perceiving. This c r i t i c i s m could only be a problem for a theory concerned with epistemic perception. Although in ordinary language i t is often evident from the context whether an occurrence of the word "perceives" is being used in the simple or in the epistemic sense, the use w i l l be made e x p l i c i t throughout th is thes is . Any occurrences of "perceives" in the simple sense w i l l be c lear ly indicated by some construction such as "(simple) perceives", or by subst i tu t ing "senses" for "perceives". Otherwise, the word should be taken in the epistemic sense. F i n a l l y , the d i s t i n c t i o n between epistemic and simple perceiving is sometimes i d e n t i f i e d with the d i s t i n c t i o n between "perceiving things" and "perceiving tha t " . This is not qui te accurate because epistemic perception is any sort of perception which requires the perceiver to be in some 8 epistemic (usual ly be l i e f ) s ta te . "Perceiving as" , as well as "perceiving t h a t " , is a var ie ty of epistemic perception. These two notions d i f f e r both grammatically and substant ively. With respect to grammar, when "perceives that " occurs in a sentence, "perceives" acts as a bridge verb, taking a sentence as i t s grammatical object. "Perceives as" takes a noun or noun phrase as i t s grammatical object. Also, the t ru th of "John perceives that there is a hawk overhead" ( fo r example) requires the embedded sentence, "there is a hawk overhead", to be 5 t rue . However, "John perceives the thing overhead ( t h a t , i t , something...) as a hawk" can be true even i f that which John sees as a hawk is not a hawk at a l l . There is an exception to th is general observation, though, whenever the noun phrases occurring to the l e f t and to the r i gh t of "as" are e i ther the same, as in "perceives X as (qua) X", or e f fec t i ve l y the same, as in "perceives X as such". Both "perceiving as" and "perceiving that" form r e f e r e n t i a l l y opaque contexts since both involve epistemic states of the perceiver. With reference to the previous example, suppose that the very hawk John perceives is a red- ta i led hawk. From the t ru th of "John perceives that there is a hawk overhead", i t does not fol low that John perceives that there is a red-t a i l e d hawk overhead. S im i la r l y , from the fac t that John perceives some-thing as a hawk, i t does not fol low that John perceives i t as a red- ta i led hawk. The di f ference in the locut ions ' opacity is that whereas "perceives that " does not guarantee t ru th - preserving subst i tu t ion of codesignative terms anywhere to the r igh t of the verb, "perceives as" r e s t r i c t s subs t i t -ut ion only to the r i gh t of "as" . With the exception of the expressions "perceives X as such" and "perceives X as X", noun phrases occurring to the l e f t of "as". The important substantive di f ference concerns the aforementioned di f ference in the t ru th conditions of the expressions. That i s , John can perceive something as a hawk, but cannot perceive that there is a hawk when there are only ravens overhead. I f one bases a theory of perception on the notion of "perceiving t h a t " , then non-veridical perception must simply be regarded as fa i lu res to perceive. I f one's perceptual theory is based on the idea of "perceiving as" , then non-veridical can be counted as misper-9 cept ion, but perception nonetheless. 6 I t is not clear which notion of epistemic perception Gibson would adopt, par t l y because he is not completely decided on how to handle perceptual mistakes, and par t ly because he is not as careful about his terminology as would be phi losophical ly desirable. Fodor and Pylyshyn sometimes seem inc l ined to categorize Gibson as holding a theory which analyzes perception in terms of "perceiving t h a t " . 1 ^ However, Gibson's discussion can be read very comfortably in terms of "perceiving as". My discussion w i l l therefore make fa r greater use of the l a t t e r locut ion . ( i i ) I t must seem quite presumptuous of Gibsonians to claim that v i r t u a l l y the en t i re community, past and present, of researchers in perception have been labouring under certain mistaken assumptions. This is especial ly t rue since there is no systematic discussion in Gibson of various Establishment a l ternat ives and how they pa r t i cu la r l y exemplify the mistakes he suggests. As a r e s u l t , Gibson stands accused by some c r i t i c s , such as S. Ul lman, 1 * of arguing se lect ive ly against proposed a l te rna t ives . Although there is a cer ta in force in the complaint, the adequacy of Gibson's approach w i l l of course depend upon i t s own success in explaining the phenomena to which i t i s supposed to apply, rather than on the def ic iencies of r i va l s ( I am in no way suggesting that Ullman is unaware of t h i s ) . I t is therefore not merely presumptuous, but unnecessarily so, fo r Gibsonians to assert that nearly the ent i re h is tory of perceptual theory indulges in the same basic er rors . A l t e r n a t i v e l y , one can simply i den t i f y the general assumptions Gibson supposes philosophers and psychologists should abandon and take the Establishment theories to be whichever theories happen to subscribe to these. U l t imate ly , Gibson must show that his own view succeeds in avoiding the d i f f i c u l t i e s in accounting fo r perception that he envisages (wi thout , of 7 course, ra is ing fur ther ones). A f a i r l y general attempt of Gibson's to d i f f e r e n t i a t e between the t r a d i t i o n and his own view is as fo l lows: Up to the present t ime, theories of sense-perception have taken for granted that perception depends wholly on sensations that are spec i f ic to receptors. I have cal led these theories of sensation-based perception. The present theory asserts the p o s s i b i l i t y of perceptual experience without underlying sensory qua l i t i es that are spec i f ic to receptors, and I have cal led th is a theory of information-based perception.12 Now one reason that Gibson is accused of s e l e c t i v i t y is t h a t , in spi te of his use of "sensation" and "sensory q u a l i t y " , he means to take aim at modern-day "information-processing" views as well as such t r a d i t i o n a l 13 14 theories as those of Hermann von Helmholtz, or even John Locke. Gibson takes the information-processing theories to be l i t t l e more than dressed-up versions of Lockean or Helmholtzian views: Not even the current theory that the inputs of the sensory channels are subject to "cognit ive processing" w i l l do. The inputs are described in terms of information theory, but the processes are described in terms of old-fashioned mental acts: recogni t ion, i n te rp re ta t i on , inference, concepts, ideas, and storage and re t r i eva l of ideas. These are s t i l l the operations of the mind on the deliverances of the senses, and there are too many complexities entai led in th is theory.15 In a s imi lar ve in , Gibson complains about theor is ts who have taken computer analogies too much to heart: Adherents to the t r a d i t i o n a l theories of perception have recently been making the claim that what they assume is the processing of information in a modern sense of the term, not sensations, and that therefore they are not bound by the t r a d i t -ional theories of perception. But i t seems to me that a l l they are doing is climbing on the la tes t bandwagon, the computer band-wagon, without reappraising the t r a d i t i o n a l assumption that perceiving is the processing of i n p u t s . ^ To put matters in the most neutral language possible, Gibson's object ion is to the view that perceptual pickup of information is a matter of sensory/perceptual systems receiving meaningless input which is then 8 transformed in to information by cognit ive processes that in te rpre t i t . The conception of input Gibson purports to have in mind is "sensory or af ferent nerve impu lses . " ^ Elsewhere, however, Gibson's attack is directed against views on which the input would be cal led sense-data, sensations, visual or re t ina l images, or qua l i t i es or disturbances in sense-f ie lds. His chief concern is wi th the idea of input being "spec i f ic to receptors" rather than to features of one's surroundings (environment). The term "spec i f i ca t ion" is a technical one on the ecological approach, one that is important in understanding the view. In general, i f a sensory input is uniquely associated with some state of the nervous system, then the input is said to be spec i f ic to that state (Gibson has such a general notion of input that i t is d i f f i c u l t to def ine, but I take i t that i t is something of which perceiving organisms may be cogni t ive ly aware. More w i l l be said about the notion of sensory input s h o r t l y ) . In terms of the ecological notion of " s p e c i f i c " , the t r a d i t i o n a l view amounts to the claim that a sensory input specif ies a state of the nervous system or of the receptor. That i s , by obtaining input , a perceiving organism picks up information about some state of i t s e l f . One and the same kind of environmental phenomenon can bring about d i f f e ren t states of one's nervous system and one and the same state may be brought about by d i f fe ren t kinds of environmental phenomena. Therefore, a sensory input which is spec i f ic to a nervous system state w i l l only be spec i f ic to such a s ta te . I t w i l l not also be uniquely associated with some environmental feature. Any psychological theory of perception which has some form of sensory input as i t s basis w i l l count as an Establishment theory fo r Gibson. The ecological theory of spec i f i ca t i on , as i t has been explained by some of Gibson's supporters, w i l l receive more thorough, 9 c r i t i c a l treatment in la te r chapters. The problems Gibson envisages for receptor-speci f ic approaches to perception w i l l be described presently. Gibson claims that sensation-based theories (theories based on sensory input that i s ) require a cognit ive structure capable of get t ing information about one's environment out of the meaningless input. Operations postulated to explain how receptor-speci f ic input is transformed in to information about one's environment are the sorts of cogni t ive , mental, or in ternal processes Gibson claims are not necessary fo r perception. These processes f a l l in the general area of inference and memory (storage and r e c a l l ) . In l i s t i n g 18 operations he takes to be those invoked by sensation-based approaches, Gibson uses, in addit ion to memory, the categories of "menta l " , "semi log ica l " , and "decoding operat ions". Examples found under these headings range from Kantian-style appl icat ion of a p r i o r i categories of understanding, through deduction of features of the world v ia unconscious inference (a t t r i bu ted to Helmholtz), to the decoding of signals and use of perceptual cues. Language communication can be used as an analogy to i l l u s t r a t e the basic model of t r a d i t i o n a l approaches fo r Gibson. A language, in the general sense, consists of a set of symbols (physical signals) which has some in te rp re ta t i on . A receiver 's being communicated to is therefore a two-stage process. A tokening (occurrence or use) of some of the symbols must be detected by the receiver. This is the correlate of reception of sensory input . No information is imparted to the receiver by the mere reception of s igna ls , however. I t s t i l l must have in te rp re ta t i ve s k i l l s of certain kinds in order to determine what the received input is supposed to convey. In fe ren t ia l processes are necessary, fo r example, in order to recognize a token as being one of a given type. Also, depending on the complexity of the system (and language) some symbols may be ambiguous. Some funct ion of the receiver therefore must be to recognize the context in which a symbol occurs in order to determine which of several meanings the given token bears. F ina l l y , any such system has to have some memory system so that stored knowledge of the re la t ions between symbols and what they denote can be recal led at the appropriate time and applied to present input . Any perceptual theory conforming to the "sense-and-interpret" analogy const i tutes a version of the t r a d i t i o n a l approach in the sense Gibson means. While he and his fol lowers exh ib i t a broad suspicion of cognitive/mental processes in perceiv ing, however, Gibson is most luc id and persistent in his doubts about the role of memory. His main conceptual concern pertains to the notion that epistemic perception of one's environment is dependent upon previously acquired background knowledge, or concepts (innate or learned), and p r io r sensory input. Regarding sensation-based theor ies , Gibson notes that "A l l theor is ts seem to agree that past experience is brought to bear on the sensory inputs, which means that memories are somehow applied to 19 them. Cont ra r i l y , Gibson's own view is that perceiving does not require the appl icat ion of memory to input . Before ind icat ing the kinds of support Gibson c i tes fo r his approach, a comment should be made on the status of sensory input. Nowhere does Gibson deny that there are such things as visual sensations. His view is that these simply do not f igure in a cognit ive theory of perception. Sensations, ra ther , are by-products of the physiological equipment with which organisms perceive. In f a c t , fa r from being the basis of perception, sensations are thought by Gibson to be a hindrance to i t . He l i s t s the 20 obtrusion of sensations on perception as a source of de f ic ien t perception. For v i s i o n , the obtrusion would amount to a perceiver attending in an uncharacter is t ic way to a two-dimensional visual f i e l d rather than the three-dimensional world. Claiming they are i r re levant to an account of perceptual cognit ion might well be the task Gibson plans for any notion of sensory input , although the spec i f ic reason need not always be the same. In the case of sensations or s imi la r "menta l is t ic" notions (sense-data, images in a visual f i e l d , for example), or even with regard to re t ina l images, i t is f a i r l y plausible to suggest these are j u s t by-products rather than the basis or cause of perception. Perhaps with other, physiological construals of input , such as nerve impulses or exc i ta t ion of receptor-banks, i t could be claimed t h a t , although these f igure in the causation.of perception, they are part of a phys io log ica l , rather than a cognit ive or psychological account of perceiving. On Gibson's behalf one could say that a psychologist or philosopher concerned with perception is not expected to analyze perception in terms of neural impulses any more than he would be expected to analyze i t in terms of chemical reactions or the behavior of energy pa r t i c l es . Making th i s kind of point requires only the recognit ion of the existence of some hierarchy of levels of descr ipt ion and explanation in science. Gibson 21 and his fol lowers c lear ly do recognize some such hierarchy. One should therefore not be too quick to suppose Gibson denies the very existence of well-recognized psychological or physiological phenomena. ( i i i ) Gibson states that his o r ig ina l motivation fo r considering a d i f fe ren t approach to perception is the resu l t of studies involv ing the notion of depth-perception ( th i s is a term, i nc iden ta l l y , that.Gibson t r i e s to avoid) in the 1940's. The experiments were an attempt to apply psychological theory concerning depth-perception to problems in aviat ion and f l i g h t t r a i n i n g . An important i n i t i a l assumption apparently was that depth-percep-t ion was based on the detection of cues in a f l a t (two-dimensional) visual f i e l d . Gibson notes: The trouble was that none of the tests based on cues for depth predicted the success or f a i l u r e of a student p i l o t , and none of the proposals fo r improving deptti perception by t ra in ing made i t any easier to learn to f l y . . . ^2 To th is he adds: I now say that there is information in ambient l i g h t fo r the perception of the layout of surfaces but not that there are cues or clues fo r the perception of depth. The t r a d i t i o n a l l i s t of cues is worthless i f perception does not begin wi th a f l a t p ic ture . I t r i e d to reformulate the l i s t in 1950 as "gradients and steps of re t i na l s t imulat ion (Gibson, 1950b, pp. 137f f ) . The hypothesis, of gradients was a good beginning, but the reformulation f a i l e d . Since Gibson f i r s t attempted to develop a new approach to perception, he has suggested a number of problems for so-cal led Establishment views, including a var ie ty of experimental resul ts which he takes to count against d i f f e ren t versions. The main general object ions, though, continue to be founded on the same theme underlying the foregoing quotes that the very basis fo r perception as i t is t r a d i t i o n a l l y conceived is deeply flawed and needs to be replaced. This is often put, to repeat, by claiming that sensory input is spec i f i c to receptors (states of a sensory or nervous system), rather than being spec i f i c to features of the environment. Propert ies, objects, events, processes, states of a f f a i r s may a l l be captured by the use of " fea tures" : although Gibsonians take special in terest in the perception of events, and of cer ta in kinds of proper t ies, they are not as par t i cu la r about t h e i r ontological commitments as a philosopher might be. To say that input is not spec i f ic to environmental features, in Gibson's view, is to say that with respect to the environment the input is meaningless. Gibson objects to the notion that the pickup of information (perceptual) awareness of what one's physical surroundings contains, must somehow be mentally derived or constructed from meaningless sensory input. The processes one needs to postulate in order to explain how information is gleaned from receptor-speci f ic input , Gibson th inks , lead to too many theoret ica l complexities and/or perp lex i t ies . He also suggests that per-ception based on meaningless input is fundamentally flawed because i t requires possession of p r io r knowledge or concepts in order fo r a perceiving organism to process present input in to information. I f sensory input is spec i f ic to receptors, then a given input is supposed to be uniquely associated with some state of an organism's sensory or nervous system. I t uniquely corresponds to a kind of receptor-s tate. I t may or may not also be pecul iar ly correlated with some feature of an organism's environment. When a given kind of sensory input i§_ uniquely associated with some environmental phenomenon, i t is a mere accident that i t is so re la ted. I t is an accident in the sense that i t merely happens that only one sort of environmental phenomenon is uniquely correlated with a cer ta in kind of sensory input and therefore to a par t i cu la r state of the organism's nervous system. I f sensory input is spec i f ic to receptor-s tates, then changes of input w i l l correspond to changes in these states. As is the case with associations between given sensory input and environmental phenomena, there may or may not be correlat ions between changes of input and changes (whether of a par t i cu la r type or of d i f fe ren t ones) in features of one's environment. A l terat ions in the frequency-mixture or d i rec t ion of the source of i l l u m i n -at ion w i l l change the state of one's sensory system, as w i l l a l te r ing one's perspective (pos i t ion re la t i ve to one's physical surroundings). Some states and changes of state of an organism's nervous system are the resul t of physiological vagaries, connected e i ther i n d i r e c t l y or not at a l l to percep-t ion of the environment (after-images and hal lucinat ions fo r instance). At the same t ime, many of the a l te ra t ions in the sensory portions of an organism's nervous system are the resu l t of changes in i t s environment. Any perceptible event or process w i l l do as an example. 14 Since receptor-speci f ic input may or may not be uniquely associated with par t i cu la r kinds of environmental phenomena, i t does not re l i ab l y indicate consti tuents of an organism's physical surroundings. S im i la r l y , since change of input may or may not correspond to change in the environment, i t is no consistent ind icator for the occurrence of environmental transform-at ions. Gibson supposes that because sensory input is meaningless in th is sense organisms must be endowed with capab i l i t i es for i n t e r p r e t a t i o n , means for adding to sensory input , that leads to some d i f f i c u l t questions: A) Given a theory on which input is not spec i f ic to features of the environment, a par t i cu la r sensory input could hardly be expected to be especial ly informat ive. That i s , i t is more plausible to suppose on an Establishment theory that information about one's surroundings is the resu l t of processing a series of inputs rather than indiv idual ones. Unless one happens to be f i x a t i n g , a highly non-typical condit ion fo r an act ive organism, a series of visual inputs is going to consist of d i f f e r i n g i n d i v i d -uals. In sp i te of the variable input , perceiving organisms (human ones at least) are aware of t he i r environment as stable and unchanging. Gibson takes i t to be a mistake to t r y to explain th is phenomenon in terms of some-how applying mental processes, reca l l of past s imi lar occurrences, applying appropriate p r io r knowledge (concepts) and the l i k e to the input in order to derive awareness of unchanging features of one's physical surroundings. He wr i tes : The century-old problem of why the world does not seem to move when the eyes move and the analogous problem of why the room does not appear to go around when one looks around are unnecessary. They only ar ise from the assumption that visual s t imul i andpVisual sensations are the elements of visual perception. Although the solut ion to the puzzle given by Gibson involves a good deal more than is indicated in th is passage (a new notion of the contr ibu-t i on of v is ion to proprioception and of the re la t ion between proprioception and exterocept ion) , i t s resistance to resolut ion in the past is c lear ly seen as resu l t ing from the assumption that perception is based on sensory inputs (visual sensation, in th is case). B) A puzzle closely related to how s t a b i l i t y , or nonchange, is perceived in the face of var iable inputs , is that of how an object is perceived as pers is t ing in the face of variable input. Any given kind of environmental feature may be associated with an i nde f i n i t e var ie ty of sensory input. Some part of an organism's sensory apparatus, then, must recognize any of an i n d e f i n i t e l y large se t , or any of i n d e f i n i t e l y many sequences of input as re lated to some par t i cu la r kind of environmental phenomenon. To the extent that i t remains mysterious as to how a sensory system could perform th is task, Gibson would regard the question of how large and var iable sets of inputs manage to y i e l d awareness of the same, pers is t ing object as a problem the Establishment has yet to resolve. This is a close kin to a problem 25 that has troubled cer ta in philosophers : how can one recognize an object as one and the same by obtaining d i f f e ren t sense-impressions, sense-data, or the l i k e , belonging to i t ? C) While these f i r s t two issues concern perception of pers is t ing features of the environment, Gibson thinks the Establishment approach leads to complications in the perception of events (changing features) as w e l l . The reg is ter ing of a sequence of sensory input occurs over t ime. The input lasts only as long as the environment is const i tuted so as to cause the state of an organism's nervous system which is speci f ied by the input. According to Gibson's version of the t r a d i t i o n a l approach, as one perceives some progressive change in the condit ion of one's environment, there is a succession of sensory inputs as new states of the environment bring about d i f f e r e n t , successive receptor-states. An organism must therefore have some complex s torage- reca l l - in tegra t ion system in order to re ta in and combine immediately past inputs with successive to bring about awareness of an event. Without some memory process, there would be naught but the discrete reg is-te r ing of indiv idual inputs. Gibson apparently believes any such storage-reca l l system would be far too cumbersome and that there is no coherent explanation of how indiv idual sensory input could become "fused in to a scene" (or some cor re la te ) . On the contrary, he claims: The simple fac t is that perceiving is not focused down to the present item in a temporal ser ies. Animals and men perceive motions, events, episodes, and whole sequences. The doctr ine of sensation-based perception requires the assumption that a success-ion of items can be grasped only i f the e a r l i e r ones are held over so as to be combined with la te r ones in a single composite. From th is comes the theory of t races, requir ing that every percept lay down a t race , that they accumulate, and that every trace be theo re t i ca l l y able to re instate i t s proper, percept. This can be pushed to absurdi ty. I t is bet ter to assume that a succession of items can be grasped without having to convert a l l of them into a simultaneous composite.'<Q The a l te rnat ive Gibson alludes to in the last sentence, as w i l l be discussed next chapter, is to suppose there are certa in kinds of abstract properties cal led invar iants which are detectable by organisms' perceptual systems and which specify events and processes in the environment. D) The re jec t ion of the idea that percept ion"is "focused down to the present item in a temporal ser ies" is re f lec ted as well in what Gibson thinks about perception of that which is about to occur. He adopts the idea that information pickup "slops over", so to speak, to include imminent events in addit ion to those which have already occurred. The conclusion Gibson draws is the resu l t of experiments conducted by himself along with 28 29 two colleagues and la te r by W. Sch i f f . An observer is placed close to a translucent screen upon which a small s i lhouet te is magnified rap id ly . Observers, including a var ie ty of animals used by Schi f f in the la te r study, apparently experience th is as rapid approach of a r i g i d object (as opposed, fo r instance, to enlargement or expansion of an e l a s t i c , non-approaching one). That is to say, the observer's behavior, when there is rapid magnif icat ion to the l i m i t (where the subject 's borders of the s i lhouet te extend beyond the boundaries of f i e l d of view) indicates expectation of c o l l i s i o n . The test-subjects 30 b l inked, averted t h e i r eyes, cringed, and so f o r t h . Gibsonians would no doubt want to make the general point that the consistency and speed of subject 's response, even among f a i r l y p r im i t i ve creatures (crabs) tested by Sch i f f , would make an explanation in terms of processing s l ices of the expanding pattern and i n fe r r i ng " c o l l i s i o n " (an alleged t r a d i t i o n a l explan-31 at ion) qui te implausible. The conclusion Gibson draws which counts expl i c i t l y only against explanations of the behavior of the observers in terms of expectat ion, or inference, based on previous experience, is the fo l lowin The experiments of Sch i f f , Caviness, and Gibson (1962) and Schi f f (1965) on opt ical magnif ication of a s i lhouet te in the f i e l d of view demonstrate that " looming", the visual information fo r imminent c o l l i s i o n , is often detected by young animals who have never had painful encounters with an approaching object. They shrink away or b l ink t h e i r eyes, or otherwise make protect ive responses without having any reason to "expect" c o l l i s i o n by reason of past experience. In th is case the visual nervous system is presumably attuned to the information at b i r th .32 Now Gibson is no more inc l ined to accept a theory of perception which would explain the " c o l l i s i o n " experiments in terms of expectations or in fe r ences from some innate phenomenon, such as innate knowledge or concepts, than he is inc l ined to accept an empi r ic is t view. I t is not clear from the previous quote whether Gibson would also consider these experiments as counting against the so-cal led n a t i v i s t version of the Establishment approach, though. However, since he would c lear ly not consider the resul ts obtained by himself , Sch i f f , and Caviness as supporting some "innate ideas" theory, there must be some d i s t i n c t i o n between such notions and that of "the attunement of perceptual systems at b i r t h " . Gibson never makes such a d i s t i n c t i o n e x p l i c i t . Perhaps the d i s t i nc t i on Gibson has in mind would be that "attunement" of a perceptual system consists in an organism's being physio logical ly structured so as to pick up certa in kinds of information. To Gibson, th is would be d i s t i n c t from a theory which endows an organism, not wi th a par t i cu la r physiological makeup, but with the possession, p r io r to b i r t h , of cer ta in information. E) F i n a l l y , to take input spec i f ic to states of receptors as the basis 33 of perception is as Shaw, Turvey and Mace have put i t , to subscribe to "the doctrine of in t ractab le non-spec i f i c i t y " . Since the postulated input is not uniquely associated with environmental features, i t could only y i e l d information about the environment (be taken as an indicator of a part-i cu la r environmental type, say) i f i t were known what kind of th ing under various circumstances produced the receptor state(s) speci f ied by some input. The conversion of sensory input in to information about one's envir-onment therefore requires p r io r possession of knowledge and/or concepts. Either the p r io r knowledge required by an organism to pick up presently avai lable information is i t s e l f acquired or i t is innate. Taking the former empi r ic is t option produces a regress: an organism cannot pick up information about i t s environment without already having some other information about i t s environment that i t cannot pick up without s t i l l other in format ion, and so on. Taking the l a t t e r , n a t i v i s t , horn only postpones the problem. By postulat ing certa in innate informat ion, concepts, pr inc ip les of reason, or the l i k e , one can explain the o r ig in of the inform-at ion an indiv idual organism possesses. However, i t s t i l l remains to be explained how the species to which an organism belongs has come to possess innate knowledge (and presumably how i t is passed on through successive generations). The n a t i v i s t is thus seen as encountering a problem at the phylogenic level s imi lar to the one posed fo r the empi r ic is t at the onto-genic l eve l . A present species member is said to have cer ta in information i t acquired genet ica l ly . I t s parents had the information to pass on because they acquired genet i ca l l y , and so on without end. The n a t i v i s t view thus makes the o r ig in of our information as mysterious as the empi r i c is t . In general terms, Gibson explains his version of the object ion by saying: The error l i e s , i t seems to me, in assuming that e i ther innate ideas or acquired ideas must be applied to bare sensory inputs for perceiving to occur. The fa l lacy is to assume that because inputs convey no knowledge they can somehow be made to y i e l d knowledge by "processing" them. Knowledge of the world must come from somewhere; the debate is over whether i t comes from stored knowledge, from innate knowledge, or from reason. But a l l three doctrines beg the question. Knowledge of the world cannot be explained by supposing that knowledge of the world already ex is ts . A l l forms of 34 cognit ive processing imply cognit ion so as to explain cogni t ion. The d i f f i c u l t y Gibson sees, once again, is c lear ly directed at the assumption that perceiving is based on some form of meaningless input. He objects to any notion that p r io r knowledge, concepts, or some correlate is required to get fu r ther information from sensory input. Thus Gibson wri tes that " i f you agree to abandon the dogma that 'percepts without concepts are b l i n d 1 , as Kant put i t , a deep theoret ica l mess, a genuine quagmire, w i l l 35 dry up." The problem posed fo r the Establishment here is assigned a great deal of importance by a number of Gibsonians. There are a l te rnat ive ways of s ta t ing exactly what Gibson intends to r e j e c t , although they amount to much the same th ing . One may say that he rejects e i ther the necessity of p r io r information or the necessity of cognitive/mental processing in perception. The re jec t ion of processing leaves no means for concepts or the l i k e to become e f fec t i ve in perception. The re jec t ion of p r io r knowledge places serious res t r i c t i ons on that fo r which cognit ive processes could be used. I t is not clear which point should be given precedence but they converge on the same fundamental c r i t i c i s m of the Establishment: the assumption that perceptual pickup of information is based upon what Gibson would regard as meaningless input , input that does not specify features of an organism's environment, is a mistake. Any theory which accepts the assumption to which Gibson objects , whatever the theoret ica l language in which the view is couched, is an instance of what Gibson and his fol lowers ca l l the t r a d i t i o n a l approach to perception (the Establishment view). Gibson's a l te rnat ive is to develop a new basis fo r perception on which the input is uniquely associated with par t i cu la r features of an organism's environment, and on which the detection of such input by some perceptual system, by i t s e l f , const i tutes the pick up of information. Michael T. Turvey, Robert E. Shaw, Edward S. Reed, and Will iam M. Mace put the new assert ion th is way: The fundamental hypothesis of the ecological approach to v i s i o n , elaborated at great length by Gibson (1966, 1979) is that opt ical s t ructure specif ies i t s environmental source and tha t , therefore , mobile organisms with active visual systems that can pick up th is information w i l l see t h e i r environments and sui tably adjust t h e i r a c t i v i t y , i f and when they detect that information (and only then). Gibson himself expresses his fundamental view by saying that inform-37 at ion is "simply ava i lab le " , or by claiming that organisms perceive 38 "meanings" or "values". The sense in which the Gibsonian theory is intended to be a theory of d i rec t perception is that the perceptual pick-up of information does not require the mediation of cognit ive processes, especial ly as would be used to apply p r i o r knowledge/concepts to sensory input : when input is properly construed, i t s detection by perceptual systems is the pickup of informat ion. ( i v ) Any d i rec t theory of perception must be so with respect to some set of phenomena. Gibson's view is no exception. Throughout th is discussion, i t has been understood that the ecological approach is a theory on which perceptual pickup of information about the environment is d i rec t . I have a lso, though not e x p l i c i t l y , roughly equated environments wi th organisms' physical surroundings. "Environment" is a technical notion of Gibson's, however, and even though i t cannot be defined prec ise ly , i t is important to bring out points about the general const i tu t ion of an environment. Philosophical debate concerning the immediacy of human perceptual connections to things are most frequently concerned with these in re la t ion to physical or material objects or some s imi la r not ion, these being taken 39 as sui table samples of the so-cal led external world. The external world could be thought of as containing, from some given t h e o r i s t ' s point of view, a l l that is mater ia l : physical objects, events, processes, or states of a f f a i r s of any size and s i tuated anywhere w i th in the physical universe. Among the population of the ex terna l , physical world is usually counted perceivers' bodies, although perceivers themselves are to be regarded as d i s t i n c t and not part of i t . Gibson would not make such a d i s t i n c t i o n . A perceiving organism is part of i t s environment, on the one hand. On the other, when Gibson ta lks about propr iocept ion, he is not re fe r r ing to aware-ness of an ephemeral s e l f , a Cartesian th inking thing which is d i s t i n c t from the physical organism that moves through and in teracts wi th i t s envi r -onment. Sel f -percept ion, in Gibsonian terms, involves awareness of the states of one's own body and i t s re la t ions to surrounding features of the environment. Although environments include perceivers while the external world t y p i c a l l y does not, an environment is largely a subset of the external world. I t is a place that supports l i f e , holds features necessary fo r org-anisms of some kind to l i v e . Put roughly, an environment consists of "animal-relevant" features of the external world. "External" or "physical" world more closely resembles what Gibson means by "the world of physics". Gibson dist inguishes the world of physics from an environment by w r i t i n g : The world of physics encompasses everything from atoms through t e r r e s t r i a l objects to galaxies. These things ex is t at d i f fe ren t levels of size that go to almost unimaginable extremes. The physical world of atoms and t h e i r ul t imate par t ic les is measured at the level of mi l l ion ths of a mi l l imeter and less. The astronomical world of stars and galaxies is measured at the level of l i g h t years and more. Neither of these extremes is an environment. The size-level at which the environment exists is the intermediate one that is measured in mi l l imeters and meters. The ordinary, fami l i a r things of earth are of th is size - - actual ly a narrow band re la t i ve to the fa r ex t remes.^ With respect to l i f e on th is planet, an environment consists of f a i r l y local physical phenomena. The sun, other planets and more distant astronom-ica l objects would a l l be excluded. The most important point here, though is the notion of scale. I t is very important to Gibson that perceiving should be l inked to what an organism must in teract with in the physical regions in which i t has grown up and in which i t s species has evolved. In general, animals in teract with moderately s ized, tangible physical phenomena and therefore these are the kinds of th ings , Gibson supposes, about which organisms ought to be concerned with and suited fo r picking up information. What w i l l be said in subsequent chapters w i l l be said; bearing in mind that environmental phenomena, that which we perceive™, consists of the "narrow band" of physical things with which organisms would t y p i c a l l y i n te rac t . The basic components of an environment, to be a l i t t l e more rigorous 41 about Gibson's concept, are substances, surfaces, and a medium. Gibson postulates two environments for ear th . These are the t e r r e s t r i a l and the aquatic. For t e r r e s t r i a l animals (and i t is a species of these which is obviously of the greatest i n t e r e s t ) , the medium is a i r . The medium, according to Gibson is "transparent" to perception, t e r r e s t r i a l organisms perceive things in and through i t . I t permits locomotion. The notions of "substance" and "surface" are a pai r because the surfaces of an environ-ment are the surfaces of i t s substances. Gibson defines "substance" very generally as "sol ids and l iqu ids that vary in composition and in resistance 42 to change." In shor t , they are the tangible s t u f f of which the environ-ment is composed: f l e s h , wood, g ran i te , and ( fo r the t e r r e s t r i a l environ-ment) water would a l l count as typ ica l examples. Gibson characterizes "surface" very general ly , saying that i t refers to the boundary between a substance and a medium. Every substance has some surface. These two notions are v i t a l because the layout of surfaces and the nature of the substances to which they belong (sometimes j o i n t l y referred to as the envi r -onment layout) are the terms in which Gibson organizes his account of visual perception. Only structured l i g h t energy, as w i l l be explained next chapter can be a source fo r the visual pickup of information. The substances and the surfaces make up the source of that which serves to provide l i g h t with the requis i te s t ruc ture . The concept of an environment is s t i l l obviously f a i r l y general and i t seems destined to remain so. There are plenty of examples, fo r instance, of "poly-environmental animals", such as amphibians and certa in aquatic mammals that could cause trouble for any attempt to c lear ly individuate the Earth's two environments. Also, some of Gibson's proponents use "environment" to mean " h a b i t a t " , 4 ^ which would c lear ly resu l t in hundreds of environments on Earth. The main point to bear in mind, though, is that Gibson wants to confine his analysis of perception to the consideration of how i t allows the pickup of information about the places where an organism l ives and the phenomena i t must in terac t with there. This genuine in teres t in the ecological or evolut ionary considerations is p a r t i c u l a r l y clear in the new categories of perception Gibson suggests. This refers to the second of the f i ve differences he claims between his approach and that of the Establishment. The essence of the claim is that organisms perceive t h e i r environments pr imar i ly in funct ional terms, or in terms of u t i l i t y . The word " func t iona l " here simply refers to that which the environment can do to and can do fo r organisms. In Gibsonian terms, organisms perceive environmental properties cal led "affordances". Affordances can be explained as fo l lows. Every animal has in terests and/or needs,, and a b i l i t i e s and vu lnerab i l -i t i e s which determine how these can be f u l l f i l l e d . These a b i l i t i e s and l im i ta t ions determine how an organism of some type could make use of i t s environment. The structure of the environment determines what there is to be made use of by an animal, what actual opportunit ies there are for i t to use i t s a b i l i t i e s to sa t is fy i t s needs (or for i t s vu lne rab i l i t i es to lead i t to g r i e f ) . This is to consider the environment in terms of that which i t affords organisms. "Affordance" is "a word Gibson coins to act as the substantive for the 44 verb " to a f f o r d " . An affordance is a- d isposi t ional property of the envir -onment. The notion is most easi ly i l l u s t r a t e d by example (although i t is d i f f i c u l t to convey properly the general i ty of the notion th is way): ( r e l a t i v e to humans) a chef 's kni fe affords cut t ing and also being cut. A ratt lesnake affords being b i t t en and poisoned, and also ( i t is alleged) eat ing. Relat ively l e v e l , so l i d ground affords standing and walking. Nelson Goodman has w r i t t e n , with respect to d isposi t ional proper t ies, that an object is f u l l of " threats and promises" and th is can be used to give a good, quick i l l u s t r a t i o n of affordances. An affordance is a threat or promise the substances and layout of surfaces of the environment hold fo r an organism ( fo l lowing Goodman on another po in t , affordances are often indicated by Gibsonians by the a r t i f i c i a l addit ion of the suf f ixes " - i b l e " or " - a b l e " ) . Gibson sometimes puts th i s by suggesting that affordances are values or 46 meanings. One of the most i n t r i gu ing and unique aspects of the ecological approach to perception is Gibson's assertion that perception of affordances is the basic form of perception. He takes the Establishment to be committed to the primacy of "form percept ion", perception in terms of manifest, sensible q u a l i t i e s . Cont ra r i l y , Gibson would deny that we recognize things pr imar i ly as red, bulgy, heavy or the l i k e , claiming instead that things are pr imar i ly recognized as ed ib le , h ide- in-ab le , cu t -w i th-ab le , f a l l - o f f -able, and so on. Thus Gibson holds that the perceptual pickup of information only requires the detection of invar iants by perceptual systems. He also main-tains that the values or meanings of things ( re la t i ve to a perceiving organism) are the terms in which the environment is pr imar i ly perceived. The resu l t is a theory according to which i t is claimed that things are d i r e c t l y perceived as having some meaning. Gibson and his fol lowers are sometimes inc l ined to put t he i r point in even more s t a r t l i n g terms: organisms " d i r e c t l y perceive" meanings. Chapter Two The previous chapter began the task of expl icat ing the Gibsonian theory of perception, explaining the f i r s t two of the f i ve essential dif ferences Gibson sees between his own view and t r a d i t i o n a l approaches. The other three respects in which Gibson is supposed to d i f f e r from the Establishment is by: 3. suggesting a new basis for perception: invar iant st ructure or propert ies 4. taking perceptual systems to be overlapping hierarchies of organs 5. saying the funct ion of perceptual systems involves the concurrent reg is ter ing of persistence and change This chapter w i l l continue the task which chapter one begins. The necessary de ta i l w i l l be added to explain (v isual ) perceptual pickup of information through the notions of v i sua l , or opt ic arrays and invar iants . The foregoing points 3, 4, and 5 w i l l thereby be explained. This chapter w i l l conclude wi th a discussion of Gibson's use of the term " in format ion" . (1) In The Senses Considered as Perceptual Systems, Gibson i den t i f i es f i ve perceptual systems*: the basic or ient ing system, the auditory system, the hapt ic , the tas te-smel l , and the visual systems. Each of these is sensi t ive to some form of energy (on a very l i be ra l use of "energy"). The basic or ient ing system is responsive to grav i ty and acceleration and the audi tory, system is sensi t ive to v ibrat ions in the medium, to give two examples. The only system to be elaborated here is the visual system, which is sensi t ive to (stimulated by) l i g h t energy. Gibson says that t r ad i t i ona l views hold that perceiving requires only the st imulat ion of receptors, or banks of receptors. He ca l l s these passive senses. These "passive" senses are act ive only insofar as the a c t i v i t y concerns the f i r i n g of neurons in the part of the brain to which the receptors are connected. Perceptual systems, on the other hand, are act ive sets of organs designed to "o r ien t , explore, invest igate, adjust , • • 2 optimize, resonate, ex t rac t , and come to an equ i l ib r ium." I t is l e f t la rge ly up to the reader to imagine precisely what is entai led by these a c t i v i t i e s which is especial ly unfortunate in the case of "ex t rac t ing" or " resonat ing". The detection of invar ian ts , which serves as Gibson's replacement fo r the reg is ter ing of input , is often described as extract ion of or resonating to an invar iant by a perceptual system. I t could only have aided Gibson's case had he taken care to explain these a c t i v i t i e s of perceptual systems in some d e t a i l . The cons t i tu t ion of the visual system is given in terms of i t s const i tuent organs and the i r adjustments: F i r s t , the lens, p u p i l , chamber, and re t ina comprise an organ. Second, the eye with i t s muscles in the o r b i t comprise an organ that is both s tab i l i zed and mobile. Th i rd , the two eyes in the head comprise a binocular organ. Fourth, the eyes in a mobile head that can turn comprise an organ for the pickup of ambient informat ion. F i f t h , the eyes in a head on a body const i tu te a superordinate organ fo r information pickup over paths of locomo-t i o n . The adjustments of accomodation, in tens i t y modulation, and dark adaptation go with the f i r s t l e v e l . The movements of compensation, f i x a t i o n , and scanning go with the second l e v e l . The movements of vergence and the pickup of d i spar i t y go wi th the t h i r d l e v e l . The movements of the head, and of the body as a whole go wi th the four th and f i f t h l e v e l s . 3 Al l of these adjustments, a c t i v i t i e s , and movements are undoubtedly important to the proper funct ioning of a visual system but only those re -lated to the las t two levels w i l l f i nd the i r way into the discussion of i nvar ian ts . The adjustments of the eye-head system const i tutes the pickup of information by looking around. The adjustments of the superordinate organ of eyes in a mobile body const i tu te the pickup of information through changing perspectives that an organism takes on i t s environment. The crucia l di f ference between a bank of receptors and a perceptual system is that the former is incapable of the kinds of adjustments Gibson envisages fo r perceptual systems. This di f ference captures the sense in which receptors are said to be "passive". Receptors merely reg is ter changing patterns of s t imulat ion and therefore do not comprise the " input-output loops" of information pickup and adjustment that are supposed to be comprised by Gibsonian perceptual systems. There is no doubt, of course, that receptors and the regions of the brain to which they are connected are parts of perceptual systems. I t is clear Gibson does not deny th is from 4 column three of his table mapping perceptual systems in Senses Considered. His point is j us t that such sets of receptors, because they merely reg is ter s t imu la t ion , are not by themselves suited fo r the detection of invar iants . They can therefore only be parts of the systems required fo r the perceptual pickup of information. ( i i ) Once again, the form of energy to which the visual system is sensi t ive is l i g h t . However, th is is only to say that the receptors contained in the organs of the visual system are stimulated by l i g h t energy. Not a l l that stimulates the visual system y ie lds perception since, in Gibson's terms, only s t ruc tured, ambient l i g h t is a source of information. Ambient l i g h t surrounding a perceiving organism is described by Gibson as a "sea" or a " f l ux " of energy. In ecological terms th is is known as the visual or opt ic array. "Ambient array" is the general case for re fe r r ing to any modali ty. Since ambient l i g h t occurs in the absence of perceivers, thou' Gibson defines "opt ic array" more accurately and generally as the ambient l i g h t surrounding a "point of observation". The spec i f ic use of "ambient l i g h t " is to d is t inguish between energy which is re f lec ted by environmental surfaces and therefore is t y p i c a l l y s t ructured, and radiant l i g h t , which is l i g h t from a source and is never structured in Gibson's sense. "Point of observation" i t s e l f requires a l i t t l e explanation. Gibson says i t should be regarded as a place or pos i t i on , not as a point in some 5 geometric sense. I t is qui te s imi lar to the concept of a s ta t ion-po in t in perspective geometry except that a s ta t ion-po in t is f i xed . Gibson intends his notion of a point of observation to be a moving point . Since any given point can move in re la t ion to any other and since they need not be occupied by organisms, i t is quite d i f f i c u l t to imagine the terms in which these points of observation are supposed to be i d e n t i f i e d or d i f f e r e n t i a t e d : perhaps i t is easier to envisage Gibson's conception in terms of actual and possible perceivers, rather than in terms of po in ts , places, or posit ions with no apparent co-ordinates. The two operative words in the conception of the opt ic array as the f lux of l i g h t energy surrounding a point of observation are "s t ructure" and "surround". An opt ic array consists of i l l um ina t i on , l i g h t that has been re f lec ted by surfaces of the environmental layout. I t is structured so long as i t is not homogenous. An unstructured array could, fo r example, be produced by t o t a l darkness, b l ind ing ly intense levels of i l l um ina t i on , or by some d i f fus ing substance, such as a th ick fog. Leaving aside such excep-t i o n s , an opt ic array has two kinds of s t ruc ture : " va r ian t " , which Gibson sometimes refers to as "perspective s t ruc tu re " , and "underlying invar iant s t ruc tu re" . These two kinds and t h e i r i n t e r r e l a t i o n w i l l be made clear once the funct ioning of the opt ic array has been explained. For structured ambient l i g h t to surround a point of observation, or to be "ambient at a p o i n t " , as Gibson sometimes puts i t , is explained when he wr i tes : To be ambient, an array must surround the point completely. I t must be environing. The f i e l d must be closed in the geometrical sense in which the surface of the sphere returns upon i t s e l f . More prec ise ly , the f i e l d is unbounded.^ A par t i cu la r visual f i e l d is confined to that which is in s ight for a f i xed pair of eyes in an unmoving organism. This makes a visual f i e l d bounded. An array, c o n t r a r i l y , is unbounded in the sense that i t is not confined to that which is in s ight from a given, f ixed pos i t i on , but consists of a l l the i l luminated surfaces which face a point of observation and are not obstructed by other surfaces from that point . The st ructure of the visual array consists of "visual so l id angles", or "angles of i n te rcep t " , of surfaces.^ Gibson mentions Euclid and Ptolemy in th is connection fo r pos tu la t ing , respect ive ly , visual cones and visual pyramids. These notions are v i r t u a l l y the same as the Gibsonian one, except that Gibson shows recognit ion of the obvious fact that environmental surfaces are not neatly c lass i f i ab le as e i ther e l l ipses or rectangles. Thus, considered as bases fo r three-dimensional f i gu res , these surfaces do not invar iab ly form e i ther cones or pyramids. A visual so l i d angle is a f igure with some surface as i t s base and an eye (bet ter s t i l l : a point of observation) as i t s apex. Every surface which is unobstructed (or unoccluded) from a point of observation subtends some visual so l id angle. Among other th ings , th is implies that there is no " f igure-ground" d i s t i n c t i o n in the ecological approach. Also, the angles are "nested" w i th in others, the smallest being surfaces of very small objects or textures of larger surfaces. A very rough d i s t i n c t i o n is made by Gibson between these r e l a t i v e l y small elements of the opt ic array, which he ca l ls " facets" and the larger surfaces that he ca l ls " faces". An opt ic or visual array, then, is a set of visual so l id angles, typ ica l comprised of a nested set of facets and faces of the environmental layout. Gibson's explanation may be found in the fo l lowing quote: There are several advantages in conceiving the opt ic array in th is way, as a nested hierarchy of so l id angles a l l having a common apex instead of as a set of rays in tersect ing at a po int . Every so l id angle, no matter how smal l , has form in the sense that i t s crossection has form, and a so l i d angle is qui te unl ike a ray in th is respect. Each so l id angle is unique, whereas a ray is not unique and can only be i d e n t i f i e d a r b i t r a r i l y , by a pair of co-ordinates. Sol id angles can f i l l up a sphere in the way that sectors can f i l l up a c i r c l e , but i t must be remembered that there are angles w i th in angles, so that t h e i r sum does not add_ up to a sphere.8 There is one fur ther piece of terminology to add here. Sol id angles are separated from one another by edges and comers. The s igni f icance of th is point has to do with the importance in Gibson's theory with the beha-v io r of edges re la t i ve to a moving point of observation as part of the foundation of var iant s t ruc tu re , thereby playing a key ro le in the detec-t i on of invar iants . This w i l l be explained present ly. The study of visual so l id angles is cal led the study of natural per-spective by Gibson and he says i t is "a continuation of ancient and medieval op t i cs " . This kind of opt ics is mainly supposed to be the examin-at ion of tr igonometr ic re lat ions elements of the environment and visual angles. He dist inguishes i t from the study of " a r t i f i c i a l perspective". 9 This l a t t e r study is what Gibson refers to as "the a r t of picture-making" because i t pertains mainly to problems of representing three dimensions in two. Gibson is very careful to d i f f e r e n t i a t e between the two d isc ip l ines but he nevertheless concedes t h a t , as they are t r a d i t i o n a l l y conceived, they share the same l i m i t a t i o n of being concerned with "frozen" or arrested structure rather than the changing structure avai lable at a moving point of observation. The ecological approach therefore extends the study of natural perspective so that i t considers changing s t ruc ture . Gibson also modifies the t r a d i t i o n a l d isc ip l ine in two other respects. F i r s t , objects are replaced by i l luminated surfaces as the basic items of study. Second, Gibson gives consideration to shading of the layout of envi-ronmental surfaces. This is a point apparently not accounted fo r by the opt ica l theor is ts Gibson takes himself to be fo l lowing. Shading and changes therein are the resu l t of the fac t that the prevai l ing source of i l l umina-t i on which is re f lec ted by surfaces t y p i c a l l y comes from some d i rec t ion and the spec i f i c d i rec t ion var ies. Af ter elaborat ing his idea of the opt ic (orv isual ) ar ray, Gibson goes on to note that the term "s t ructure" is vague as i t applies to his theory of information pickup because there are two d i f fe ren t sorts of st ructure in an array. The d i s t i n c t i o n is o r i g i n a l l y cast in terms of perspective structure that is disturbed with every movement of the point of observa-t i on versus what is known in the ecological theory as "underlying i n v a r i -ant s t r u c t u r e " . ^ Later in Gibson's discussion, however, i t comes to l i g h t that the proper contrast is a more general one between invar iant and var iant structure.* ' ' ' Perspective structure is actual ly merely one of four sources of var iant s t ruc ture . The other sources are 2) movements in an organism's eye-head system, 3) changes in i l l u m i n a t i o n , and 4) perceptible events that occur in an organism's local environment. These w i l l be explained shor t l y . With regard to 2 ) , movements of the eye-head system are dist inguished by Gibson from the other bodily movements that would const i tu te a d is turb-ance of the point o f observation. Looking around is the chief funct ion of the eye-head system. Scanning with the eyes might be included as well although Gibson does not e x p l i c i t l y say so when he discusses scanning. With 33 regard to 3 ) , both d i rec t ion and the frequency of sources of i l luminat ion are considerations. In Gibson's exposit ion of invar ian ts , there is a d iv is ion made among the four sources of variance, the f i r s t three being discussed p r io r to and independently of the fou r th . The main rat ionale behind the order could be that sources 1 ) , 2 ) , and 3) are a l l in some sense "external" to an organism's environment. A source of radiant l i g h t , i l luminat ion is never part of the environment (al though, Gibson would note, the object providing the source could be). Also, from the point of view of a par t i cu la r organism, i t is d i s t i n c t from i t s environment (although from a purely object ive stand-point animals are of course part of the fu rn i tu re of the environment): a l l of the organism's perceptions of i t s own body are proprioceptions and these are c lear ly dist inguished on.the ecological approach from the pickup of information about i t s environment. From a perceiving organism's own unique perspective, then, i t s own body is as "ex te rna l " , as much a non-environmental phenomenon, as a source of i l l um ina t ion . Adjustments of an organism's body and changes of i l l um ina t i on , there fore , are in some sense not changes in the environment. The fourth source of variance, environmental events, quite obviously const i tutes change in the environment. A l l of the sources of variance are such in the sense that they produce disturbances of opt ica l s t ructure (changes in the opt ic array — Gibson actual ly prefers to use "disturbance" and "non-disturbance" in re la t ion to the opt ic ar ray, confining "constancy" and "change" to features of the envi-12 ronment). Variances due to eye-head adjustment or change in e i ther point of observation or i l luminat ion can be, and by Gibson are, discussed with respect to an unchanging environment. The discussion concerns the percep-t ion of constancy or persistence through disturbance of opt ica l s t ruc ture . The addit ion of environmental events adds a new tw is t to the theory because i t pertains to the perception of environmental change through opt ical disturbance. Now a theory in which the mere reg is ter ing of variances in the opt ic array brought about the pickup of information would s t i l l be a kind of Establishment theory in some respects. Perception would s t i l l be based on some notion of variable input ; i t could s t i l l require some means of constructing information from variable ser ies. On the ecological approach, however, disturbance of opt ica l s t ructure is essential to the pickup of information but is not the basis of pickup. Information is picked up only when underlying invar iants are detected by an organism's visual system. The re la t ion between invar iant and var iant structure is complementary though. While the former is the basis of perception, the l a t t e r allows i t s detection because, as Gibson would put i t , the "f low" of the array, series of opt ica l disturbances produced by changing perspective and so on, "separates o f f " the underlying invar iant st ructure from the changing, ambient f l u x . The perceptual pickup of information thus involves the concurrent reg is ter ing of persistence and change. More prec ise ly , i t involves the concurrent reg is ter ing of disturbance and non-disturbance of opt ica l s t ruc-tu re . This is one of the previously l i s t e d f i ve respects in which Gibson's theory is alleged to be revolut ionary. From the ecological point of view, the t r a d i t i o n a l approach to our perception of constancy (of a stable envi-ronment) would suggest that i t is perceived in spi te of the reg is ter ing of var iant s t ruc ture . The Gibsonian approach is t h a t , although var iant s t ruc-ture is not i t s e l f a source of information about the environment, i t plays a crucia l (causal) ro le in enabling perceptual system to extract the abstract ra t ios and r e l a t i o n s , which are invar iant properties and which remain in tac t through movements of the perceiving organism, the source of i l l u m i n -a t i o n , or even environmental objects themselves. When a visual system detects an invar iant specifying some pers is t ing feature of the environment, an organism perceives s t a b i l i t y through, not in spi te o f , disturbance of the opt ic array. This d i f fe ren t way of th ink ing about the ro le of variance in perception expla ins, fo r example, why Gibson is opposed to studying percep-13 t i on through experiments that unduly constrain a perceiver 's movements. ( i i i ) In the most general terms, invar iant properties are ra t ios and r e l a -t ions involv ing substances and surfaces of the environment that remain con-stant through certa in kinds of opt ical disturbance. Gibson postulates four types of invar ian ts , each associated with one of the four sources of variance in the visual array. These are e n t i t l e d : " Invar iants underlying change of point of observat ion", " Invar iants of opt ica l s t ructure under changing i l l u m i n a t i o n " , Invariants across sampling of the ambient opt ic ar ray" , "and "Local invar iants of the ambient array under local disturbances of i t s s t ruc tu re" . A. Invar iants Under Changing Perspective From a temporari ly f ixed point of observation, cer ta in surfaces facing i t w i l l be obstructed by others. As the point of observation changes, that i s , the perceiving organism takes d i f fe ren t perspectives on i t s envi -ronment, previously covered surfaces become uncovered and vice versa. The covering of one so l id angle in the visual array by another is referred to 14 by Gibson as occlusion. In ecological terms, a moving point of observa-t ion produces delet ion of opt ical texture (what visual so l id angles consist of) along cer ta in edges and accretion of texture along others. Where there is delet ion of tex tu re , a surface is being occluded. Where there is accre-t i on of tex tu re , a surface is becoming unoccluded. One must of course remember that these disturbances occur in comprehensive pat terns, not as isolated instances. Among the kinds of things one may suppose would be speci f ied in the delet ion and accretion of texture along edges would be spat ia l re la t ions . Patterns and rates of occlusion re la t i ve to par t i cu la r changes in the point of observation specify things in the environment as being behind or in f ron t of others, as being certa in distances apart , and as being at cer ta in d is-tances from the perceiver. S i t t i n g down produces a compression of some of the so l id angles in the visual array. Perhaps the degree of compression re la t i ve to the amount of displacement of point of observation specif ies the angle of the surface re la t i ve to an organism's l ine of s ight . Examples might also be constructed which do not rea l l y involve an actual movement of the point of observation but which may s t i l l be thought of as the detection of invar iants through changing perspective. Perceivers are sometimes in a posi t ion to see d i f fe ren t environmental phenomena which are of the same kind (exact ly s imi la r in some respect) but which are pos i t -ioned rather d i f f e r e n t l y from them. In his book, Gibson uses the example of a series of telephone poles, a l l of the same height, extending in to the distance away from a perceiver. In open country, the poles are seen as the same height , even though, because the perceiver is taking a d i f f e ren t per-spective on each, each pole subtends a progressively smaller visual angle. The supposed explanation is that the ra t i o of the port ion of opt ica l texture above the horizon to the port ion below is constant. Perceiving the poles as the same size is the resu l t of i t s detection of th is invar ian t . In general, more or less complex r a t i o s , ra tes , and re lat ions emerge through the patterns of disturbance, the "opt ica l f low" , as Gibson sometimes has i t , produced by a perceiving organism moving through i t s environment. These are invar iants under changing perspective. Such constancies to be extracted from the opt ica l flow perhaps would specify s izes, spat ia l posi-t i o n s , r i g i d i t y . I f Gibson's approach to perception were cor rect , they would cer ta in ly specify objects as being s ized, shaped and s i tuated so as to a f ford she l te r , grasping, c l imbing, walk ing, and so on. That i s , the invar iants would specify affordances of the environmental layout. B. Invariants of Optical Structure Under Changing I l luminat ion The three sources of variance f a l l i n g under "changing i l l umina t ion" that Gibson mentions- are changes in d i rec t ion of the prevai l ing source, changes in i n t e n s i t y , and changes in colour or f requency-mixture.^ The d i s t r i b u t i o n of shaded and l ighted surfaces in the environment, fo r example, is a funct ion of the d i rec t ion of the prevai l ing source of i l l u m i n a t i o n . ^ As the d i rec t ion changes (the sun moving across the sky, a perceiver moving around a cave with a hand-held lan te rn , are a couple of examples), there are patterns of opt ical disturbance in the shading. Arrangement and patterns of disturbance would be thought to br ing out invariants that help to specify size and distances of th ings , and convexit ies and concavities in surfaces. Change in spectral composition or in tens i ty of prevai l ing i l luminat ion means a di f ference in the nature of l i g h t absorbed or re f lec ted by a surface. The resu l t is d i f f e r i n g absolute colour of par t i cu la r surfaces and of de ta i l in the opt ic array. With regard to i n t e n s i t y , the extremes simply make the visual pickup of information impossible. I f the i l luminat ion is too st rong, b l ind ing occurs. I f i t is too weak, i t is said that conditions are too dim in which to see. I mention th i s to underscore the fact t h a t , on Gibson's view, visual perception can f a i l to occur even when the receptors in the visual system are receiving some measure of s t imula t ion. Each of these cases can provide an example of how there can be visual sensation without visual perception. Among the features of the environment that are detected as constant through both changing in tens i ty and spectral composition are texture and pigmentation. Perception of a surface as being of a constant colour would involve the detection of charac ter is t ic ra t ios of ref lectance fo r various pigmentations. Pickup of information concerning environmental substances may be related to charac ter is t ic scat ter-pat terns that remain constant under 18 changes in the source of i l l um ina t ion . C. Invariants Across the Sampling of the Ambient Array This kind of invar iant is related to adjustments of the eye-head sys-tem. In par t i cu la r i t is re lated to the spec i f ic task of get t ing informa-t i on by looking around. As I have already noted, movements of the eye-head system are dist inguished in Gibson's c lass i f i ca to ry system from other bodi ly movements (eg . , s i t t i n g down, crouching, walk ing). Adjustments of the eye-head system are therefore, not counted as a change in the point of observation. An explanation of invar iants across sampling of the opt ic array requires some account of the notion of the process, and of "sample". A sample of the opt ic array is the port ion of i t which is in from a tempor-a r i l y f ixed viewing pos i t ion . Any motion of the eye-head system produces a new sample. That which is in s ight from a f ixed eye-head posi t ion const i tutes the contents of an organism's f i e l d of view. "Field of view", i n c i d e n t a l l y , i s a notion Gibson c lear ly wishes to 19 dist inguish from "visual f i e l d " A f i e l d of view and a visual f i e l d are both bounded in some sense, but Gibson takes the l a t t e r to re fer to some form of sensory mosaic accessible only through in t rospect ion, a "patchwork of visual sensations". A f i e l d of view is a bounded port ion of the opt ic array and therefore consists of i l luminated surfaces. To look around is to sample the opt ic array via movements of the eye-head system, to obtain successive, overlapping, and often reversible series of samples. There is somewhat of a complication in th is explanation in that the eye-head system must move whenever the point of observation does. However, t h i s can be remedied simply enough by specifying that the eye-head system i t s e l f must change state fo r invar iants to be detected through sampling, rather than a change in perspective. There w i l l no doubt be some s i m i l a r i t i e s in the contr ibut ion made to visual information pickup by looking around and by changing perspective. Use of the eye-head system w i l l produce series of samples of opt ica l s t ruc-ture through which there w i l l occur systematic.and progressive delet ion and accretion of opt ical texture along edges of so l id angles. As is the case with occlusion (and i t s reverse) through changing perspective, the covering and uncovering of texture along edges through looking around should be important to perception of spat ia l posit ions of th ings. Gibson also regards the r e v e r s i b i l i t y often associated with occlusion and wi th things going out of and coming in to the f i e l d of view as important to the perception of persistence, and of the coexistence and connectedness of that which is 20 temporari ly in sight with that which is temporari ly out of s ight . D. Local Invariants of the Ambient Array under Local Disturbances  of I t s Structure The invar iant- types considered so far have been explained in re la t ion to a "frozen" environmental layout (though they should not be thought of only as related to an unchanging environment). A tenet in the ecological approach, however, is that events and the perception of events must be part of an account of perception. Environmental events are themselves a source of opt ica l disturbance, through which invar iant properties are separated o f f and may be detected, as well as being perceptual objects. Unlike the other sorts of invar ian ts , the source of disturbance of opt ica l s t ructure (some environmental event) can be that which is speci f ied by the invar iant revealed by the associated opt ica l disturbance. Elaboration of invar iants under local disturbance of environmental st ructure requires some descr ipt ion of events. Gibson's discussion pa r t i c -u la r l y concerns what he ca l ls " t e r r e s t r i a l events". The only example given of a non- te r res t r ia l one is that of the sun's progress across the sky. The point of the d i s t i n c t i o n Gibson makes seems only to confine the discourse, in accordance with his rough notion of the " t e r r e s t r i a l environment", to events at or near the surface of the ear th. 21 Three general event-types are c i t ed . "Change of layout" includes disrupt ions and deformations of surfaces, c o l l i s i o n s , turns and displace-ments of objects. "Change in colour and texture due to change in compo-s i t i o n " pertains to chemical metamorphosis, such as the bluing of heated s t e e l . Heating, ox idat ion, reduct ion, the production of chemical substances by l i v i n g things (one of Gibson's examples is of the fading colouration of a plant surface through a decrease in ch lorophy l l . F ina l l y , there is "waxing and waning of a surface due to change in the state of matter" , such as the r o t t i n g of vegetable matter, evaporation or freezing of l i q u i d s , or the d isso lu t ion of crystals in l i q u i d . Changes of these kinds const i tute coming in to or going out of existence of surfaces. In general, the l i s t Gibson provides takes in to account a l te ra t ion of a given surface, a l te ra t ion between surfaces, and t h e i r creation or destruct ion. The supposed revolut ionary approach to events in perception taken by Gibson and his fol lowers is that events are perceived j us t the same as per-22 s i s t i ng features of the environment. That i s , on the Gibsonian view there are pe rs i s t i ng , abstract ra t ios and re lat ions underlying disturbances of opt ical s t ructure which specify events of certain kinds, as well as specify-ing objects and the i r unal ter ing propert ies. Also, a d i s t i n c t i o n needs to be maintained between events as sources of opt ical disturbance and events considered as perceptible features of the environment. In shor t , the ecol-ogical approach t reats environmental events both as objects of perception and as sources of opt ica l disturbance. Gibson could perhaps have made th i s point c learer . The reason fo r such a d i s t i n c t i o n is that not every event that resul ts in some charac ter is t ic opt ical disturbance is necessarily perceived as such. Optical disturbances produced by events should.be able to separate o f f invar iants that specify things other than simply the source of the d is turb-ance. Af ter a l l , changes in perspective and i l luminat ion reveal invar iants that specify features of the environment, rather than the sources of those disturbances (point of observation or l igh t -source) . One of the favour i te examples for ecological theor is ts of disturbance of opt ical s t ructure by local events that allows the detection (or extrac-t i on of an invar iant is the previously mentioned case of the perception of imminent c o l l i s i o n : The magnif icat ion of the visual so l id angle of an object normally accelerates as i t approaches the l i m i t of a hemispheric angle, as the object comes up to the eye. The accelerated port ion of th i s sequence was cal led "looming" by Sch i f f , Caviness, and Gibson (1962). I t speci f ies impending c o l l i s i o n , and the rate of^ magnif icat ion is proport ional to the imminence of the c o l l i s i o n . An explosive rate of surface magnif icat ion is part of the invar iant st ructure that speci f ies a cer ta in event-type: c o l l i s i o n with an object or surface. This pa r t i cu la r example is so frequently c i ted by Gibsonians because i t is taken as a source for one of the f i r s t concrete examples of 24 an invar ian t . David Lee has apparently worked out a mathematical variable to explain human reaction to impending co l l i s ions in ecological terms. Another somewhat s imi lar example ( th is one is mainly of my own design: i t is intended to i l l u s t r a t e the notion of invariants in environmental chang but the deta i ls may or may not be borne out in experiment) would be of a 25 bal l thrown toward a perceiver. The disturbance of the opt ic array would consist of the so l id angle subtended by the bal l occluding progressively greater amounts of surrounding tex ture . The perceiver 's visual system, on Gibson's view, would reg is ter the opt ica l disturbance created by the b a l l ' s changing posi t ion re la t i ve to the surrounding environment. This event, fo r one t h i n g , is qui te d i f fe ren t from a perceiver moving toward a stat ionary object , however quick ly . Part of the reason could be that a p r o j e c t i l e has a charac ter is t ic arc that is not reproduced when one merely approaches a s t i l l object . Part of the reason would also no' doubt be l inked to proprioception (v ia the visual system and otherwise). An organ ism's a b i l i t y to perceive change and constancy in the environment is con-nected very c lear ly to i t s awareness of i t s own temporari ly changing or f ixed posi t ion in i t . Also, the re la t ion between the so l id angle subtended by the bal l and those subtended by surrounding surfaces is qui te d i f fe ren t when the perceiver is in motion re la t i ve to a stat ionary b a l l , as opposed to the reverse. Surrounding texture is increasingly occluded by the b a l l ' s so l id angle in each case, but when the bal l is s t i l l and the perceiver is moving toward i t , the "background" texture w i l l be magnified at the same ra te . When the bal l is in f l i g h t (assuming a stat ionary perce iver) , the background texture w i l l not be magnified at a l l . The clear resu l t is a s t r i k i n g dif ference in the texture that is occluded in each instance. The d i f fe ren t patterns of opt ica l disturbance which resu l t from the two d i f -ferent occurrences would therefore make d i f fe ren t invar iant ra t ios or rates avai lable fo r detection by one's visual system. A perceiver 's a b i l i t y to time his catch, or to avoid being h i t by the b a l l , is no doubt contingent upon his a b i l i t y to perceive the speed of the bal l accurately. At least a pa r t i a l factor involves the detection of a rate o f magni f icat ion. Another factor could have to do wi th the b a l l ' s arc. A fas ter ba l l has a f l a t t e r arc. A ba l l t r a v e l l i n g r e l a t i v e l y quickly would therefore produce somewhat less decretion of surrounding opt ical texture along i t s lower edge and somewhat less accretion of texture along i t s upper edge than would a so l id angle associated with a s lower- t rave l l ing ba l l that covered the same distance with greater arc. Perceiving a b a l l ' s speed and path of f l i g h t , in these terms, would thus be a matter of the detection of a cer ta in invar ian t , consist ing of f a i r l y subtle mathematical re la t ions amongst a se t of visual so l id angles which are not in motion, but are undergoing s t ruc tura l disturbances. I t is important to remember that even though the two examples given concern perception of changing spat ia l r e l a t i o n s , visual perception of events concerns fa r more. There are chemical events as well as mechanical ones. Chemical events bear some s i m i l a r i t i e s to changing i l luminat ion because a chemical ( including biochemical) event produces change in sub-stance, usually resu l t ing in some di f ference in pigmentation or surface tex ture . The invar iants underlying such change w i l l , as with changing, i l l u m i n a t i o n , pertain to how l i g h t from a source is re f l ec ted , absorbed, and scattered by surfaces. While the thought Gibson has put in to the categorizat ion of environ-mental events is impressive, i t involves a curious oversight. Not every-thing on Gibson's l i s t of t e r r e s t r i a l events would produce a disturbance of opt ica l s t ructure because noticeable change occurs in some of the cases over very long periods of t ime. Some environmental events, l i ke the r ipening of a peach, w i l l not be registered as changing opt ica l s t ruc-t u r e , unless one is extremely v i g i l a n t . Others, l i ke the motion of t ree-branches in the wind w i l l count as sources of disturbances of opt ica l s t ruc-tu re . Gradual events may be considered as things perceived on the ecolo-g ical approach: i t is conceivable that peaches are perceived as r ipening, s ta lac t i t es as lengthening, or whatever. These sorts of gradual changes in the environment, however, cannot be considered as sources of opt ica l d i s t -urbances because no change is registered by the perceptual system. One is aware of a peach's r ipening because of a state the f r u i t is i n , not because of a discernible change i t is undergoing. ( i v ) Some general points about invar iants should be made in summary. These are, to emphasize, abstract ra t ios and re lat ions (mathematical, but not necessari ly geometric) that hold among components of the ambient array, and are constant through certa in kinds of disturbance of ambient (op t ica l ) s t r u -cture. These are said to be detected or extracted by perceptual systems from the array. I t is also said by Gibson that perceptual systems (or "nervous systems") resonate to invar iant propert ies. Neither "detect" nor "ex t rac t " , nor "resonate" is explained. Invariants should by no means be thought of as some novel conception of sense-data, as one might be inc l ined to do. An invar iant is not , and in many cases, Gibson bel ieves, could not be an object of awareness. In f a c t , he claims they w i l l fo r the most part f a i l to be "open to analyt ic in t rospect ion" . This must mean more than that one is not standardly aware of invar iants . Rather, i t must be that one could not attend to 45 one's own perceptual experience in such a way as to discover the invar-iants upon which pickup of certain information is based. Invar iant propert ies are qui te unl ike sense-data or s imi lar notions in th is respect. Representational theor is ts as early as Locke are prepared to grant that perceivers are not t y p i c a l l y perceptually aware of the sensory qua l i t i es 27 upon which t h e i r awareness of the world is based. I t is customary to maintain some condit ional view instead: a perceiver could be aware of bare sense-data i f he were to attend to his perceptual experience in the r igh t way. For Gibson to claim that invar iants w i l l often not be in t rospect ib le is to deny they can be objects of awareness even in th is weak, condit ional sense. An immediate consequence of th is point is that (assuming of course that invar iants are indeed discoverable), to the extent that they are not revealed by analy t ica l in t rospect ion, they must be so through independent, empirical t es t i ng . This explains Gibson's (otherwise puzzling) remark regarding invar iants under changing i l luminat ion that "they are not yet known but they almost cer ta in ly involve ra t ios of in tens i ty and color 28 among parts of the array". The study of invar iants consists of d is -covering which par t i cu la r ones there are ( that the study is r e l a t i v e l y new perhaps stands as a temporary explanation fo r the decided lack of concrete examples). This makes the f ind ing of invar ian ts , what they speci fy , and hence the question of what information can be picked up via t h e i r detection of a l l empirical issues. This is an important point for understanding how Gibson dist inguishes perceptible and non-perceptible features of the environment, which w i l l be a central issue in the next chapter. F i n a l l y , there is the matter of the re la t ion between the notions of " invar iant propert ies" and " in format ion" , or of the nature of in format ion, on Gibson's conception. In important respects, i t is an issue that cannot be resolved un t i l l a te r in th is discussion since the major controversy between Gibsonians and Establishment theor is ts centers on the notion of information and whether i t manages to do a l l that is required of i t . The only cer ta in ty is that Gibson wants to d is t inguish c lear ly between his conception and one that has come to philosophy and psychology via communication-theory: The information fo r perception is not t ransmit ted, does not consist of signals and does not enta i l a sender or a receiver. The environment does not communicate with the observers who inhabit i t . 2 5 Elsewhere, Gibson notes that informat ion, when defined (as in communications-engineering) in terms of the reduction of uncer ta inty , allows d iscr iminat ion , not "perception of" (presumably meaning "meaningful perception" ) . ^ The basis of the controversy consists of whether or not invar iant propert ies and information should be i d e n t i f i e d . Some of Gibson's comments and the locutions he uses suggests that the two should be i d e n t i f i e d . He t a l k s , fo r example, about " information fo r perception" and information 31 specifying features of the environment". The most natural reading of these phrases resul ts from assuming that " informat ion" here refers to invar ian ts : propert ies of s t ructured, ambient energy that uniquely (and lawfu l l y ) correspond to environmental phenomena. Also, information about the environment is said to be "simply ava i lab le" , i n the ambient array. On the other hand, invar iants are supposed to be generally non-in t rospec t ib le i they could not be objects of awareness. Surely i t would be incredib le to hold such a view of the information organisms pick up about t h e i r environments. The pickup of information is anyway defined by Gibson sometimes as awareness of features of the environment and th is raises a precise logical problem for the i d e n t i f i c a t i o n of information and invar ian ts . Add i t iona l l y , examples of information pickup would be such as "awareness of a surface as walk-on-able", "awareness of something as an enclosure". The information picked up about the environment could be reducible to or analyzed in to invar iant properties of the opt ic array. I f there is such an analysis in Gibson's works, i t is not the least b i t e x p l i c i t . Now Turvey, Shaw, Reed, and Mace have wr i t ten that "opt ica l s t ructure speci f ies i t s environmental source and that therefore mobile organisms with act ive visual systems that can pickup th is information w i l l see the i r environments and su i tab ly adjust t h e i r a c t i v i t y , i f and when they detect 32 that information (and only then) " . By subs t i tu t ing " invar iant " fo r " informat ion" in th is quote (s ince" invar iant" is f a i r l y c lear ly what the authors have in mind), one can derive a workable proposal. Since there is some ( j u s t i f i a b l e ) controversy over whether Gibson has a single notion of information that can bear the theoret ica l weight he places upon i t , l e t the basic ecological claim be that an organism perceptually picks up information about i t s environment when, and only when, the appropriate invar iant is detected by some of i t s perceptual systems. One suspects that Gibsonians would l i k e to i den t i f y information about the environment with inva r ian t , ambient, s t ruc ture . Pending c l a r i f i c a t i o n of cer ta in apparent c o n f l i c t s , however, th is re la t ion should neither be assumed nor granted. Instead, I w i l l regard the invar iant- in format ion re la t ion only as a b icondi t ional one. Such a re la t ion would hold i f " informat ion" and " invar ian t " were i d e n t i f i a b l e . I t does not enta i l th is i d e n t i f i c a t i o n , however. Chapter Three The grand philosophical claims made by ecological theor is ts - - not merely that the Gibsonian approach is a d i rec t theory of perception of the environment, but also that the view provides a new basis fo r real ism, a new theory of cognit ion general ly, and even a new approach to epistemology* - - has drawn very l i t t l e reaction from mainstream f igures working in the relevant areas of philosophy. This is not surpr is ing fo r there is rare ly a generous flow of discourse between d i f f e ren t academic d i sc ip l i nes . The wr i t ings of some of the Gibsonians, moreover, are qui te inaccessible. 2 Fodor and Pylyshyn's commentary is a rare exception to the evident philosophical lack of in te res t . I t is even more rare because, since i t is r e l a t i v e l y recent, the a r t i c l e takes in to account The Ecological Approach To Vis ion, which contains the most developed and detai led version Gibson gives of his theory. The Fodor/Pylyshyn a r t i c l e , add i t i ona l l y , const i tutes one ha l f of a lengthy debate, having drawn a response from four prominent 3 Gibsonians. However, the debate is incomplete: f i r s t of a l l , subsequent to the or ig ina l two (lengthy) a r t i c l e s , and probably owing to the in t ran-sigence of the pa r t i c ipan ts , there has been no comprehensive fol low-up or r e b u t t a l . Second, Turvey, Shaw, Reed, and Mace (hereinaf ter TSRM) do not so much respond to Fodor and Pylyshyn's attack on Gibson as restate general ecologist qualms and redescribe Gibson's "theory of spec i f i ca t ion " . As a r e s u l t , the ecological side f a i l s to address the po in ts , good or bad, that Fodor and Pylyshyn make. The debate does, however, centre around the key issue which, in Fodor and Pylyshyn's terms, is how the ecological approach deals with i n t e n t i o n a l i t y in perception. The fol lowing chapters w i l l focus on the c r i t i c i s m of the Gibsonian view by Fodor and Pylyshyn, and TSRM's rep ly , with the underlying issue being th is one of in ten t ion-a l i t y (roughly put, whether or not Gibson's concept of information and theory of pickup could explain epistemic perception, how one perceives things as such). I t is not at a l l clear how TSRM's response, although i t is intended to handle the i n t e n t i o n a l i t y issue, even approaches the question. The task of th is chapter, however, w i l l be to discuss the Fodor/ Pylyshyn side of the debate. The i n i t i a l argument against Gibson made by Fodor and Pylyshyn is that Gibson has not provided appropriate constraints on his notions of "d i rec t pickup" or " i nva r ian t " . In other words, i t is not clear what l i m i t s , i f any, there are on that which may be perceived. They w r i t e : The main l ine of our argument w i l l go l i k e t h i s : Gibson's account of perception is empty unless the notions of ' d i rec t pickup' and of ' i nva r ian t ' are sui tably constrained. For, patent ly , i f any property can count as an invar ian t , and i f any psychological process can count as the pickup of an invar ian t , then the i d e n t i f i c a t i o n of perception with the pickup of invar iants excludes nothing.4 I w i l l show that th is par t i cu la r c r i t i c a l l ine is not a problem fo r Gibson. He does have constraints of a k ind, although they are not obvious, and he is not caught in the dilemma Fodor and Pylyshyn fashion from the i r argument. The ecological approach can be understood as claiming that a l l "genuine" ( in some sense) perception is d i rec t perception, and therefore invariant-based. The percept ib le , then, is constrained in terms of what invar iants there ( in fac t ) are, and what they ( in fac t ) specify. This w i l l be elaborated upon l a t e r . The f i r s t step the authors take is to pose what they ca l l a t r i v i a l -iza t ion problem, the purpose of which is to show Gibson cannot do without constra ints . They then suggest four d i f f e ren t constraining pr inc ip les (Fodor and Pylyshyn ca l l these "gambits") for which they claim Gibson shows some support in his w r i t i n g s . These gambits are that only ecological proper t ies, only p ro jec t ib le proper t ies, only phenomenological proper t ies, or only that to which 'perceptual systems' respond are d i r e c t l y perceived. Each is rejected in turn as being e i ther inherent ly unworkable or anyway incompatible with the basic st ructure of Gibson's approach. To Fodor and Pylyshyn, th is outcome is inev i tab le since they see Gibson as being caught in a d i lemia: he needs some constraining pr inc ip le but cannot introduce one without al lowing that some perception requires cognit ive processing ( in ference) . In shor t , Fodor and Pylyshyn argue that e i ther "d i rec t pickup" is vacuous or else Gibson must forsake his claim that perception only i n -volves the "pickup of invar ian ts " . They base th is claim par t ly on what they regard as a common assumption: Gibson and the Establishment agree that pickup and inference exhaust the psychological processes that could produce percep-tual knowledge; hence the more pickup is constrained the more there is l e f t fo r inference to do.* Since one of Gibson's fundamental tenets is that perceiving does not require inference, then his agreement with th is assumption would seem to place him in great d i f f i c u l t y . The dilemma Fodor and Pylyshyn attempt to set fo r Gibson, however, does not work because i t is based on some mis-understandings about the ecological approach. One is that the Fodor/ Pylyshyn dilemma requires an improper use of " i nvar ian t " . Another is that the authors f a i l to appreciate the extent to which the ecological approach to perception is an empirical theory. The f i r s t mistake allows the t r i v i a l i z a t i o n problem, which Fodor and Pylyshyn describe as fo l lows: Suppose that under certain circumstances people can cor rec t ly perceive that some of the things in t he i r environment are of type P. Since you cannot cor rec t ly perceive that something is P unless that th ing is P, i t w i l l always be t r i v i a l l y true that things that can be perceived to be P share an invar iant property: namely, being P. And since, according to Gibson, what people do in perceiv-ing is d i r e c t l y pick up an appropriate invar ian t , the fol lowing pseudo-explanation of any perceptual achievement is always avai lab le: to perceive that something is P is to pick up the ( invar ian t ) property P which things of that kind have. So, fo r example, we can give the fo l lowing disarmingly simple answer to the question: how do people perceive that something is a shoe? There is a cer ta in ( invar iant ) property that a l l and only shoes have - - namely the property of being a shoe. > ; Perceiving that something is a shoe consists in the pickup of th is property. For good measure, they add a second example of such t r i v i a l explan-ations using the alleged a b i l i t y of Bernard Berenson to " t e l l by looking" whether a paint ing is a genuine DaVinci. The trouble fo r Fodor and Pylyshyn rests in the fact that the kind of thing they want to f i l l in for "P" f a i l even to be candidates fo r the t i t l e of being an invar iant property. Optical invar iants are the features of and re lat ions between the faces and facets (explained in Ch.2) compris-ing a visual array that remain constant through various kinds of d is turb-ances of the st ructure of the 'a r ray . "Being a shoe", "having been painted by DaVinci", or the generic "being P", do not even remotely qua l i fy as the sor t o f th ing Gibson has in mind. The t r i v i a l i z a t i o n problem re l ies on a use of " invar iant property" according to which one can postulate the existence of an invar iant fo r v i r t u a l l y any dist inguishable class of environmental phenomena. Fodor and Pylyshyn take the process of v isual information pickup to be s imi lar to " t e l l i n g by look ing" . On t h e i r in te rp re ta t ion of Gibson, any feature of the environment an organism can ( learn to) be aware of by looking is cal led an " i nva r ian t " . The resu l t is arm-chair science: the p r o l i f e r a t i o n of "explanations" of how certa in things are perceived that need no backing of empirical t es ts . However, opt ical invar iants are empirical properties of visual arrays so the matter of what constant ra t ios and re lat ions of opt ica l s t ructure there are is only determinable by s c i e n t i f i c invest iga t ion . The empirical nature of invariants suggest one sense in which Gibson's view of the perceptible might be thought of as unconstrained. As an empirical s c i e n t i s t , he ( r i g h t l y ) leaves questions of which invar iants there are and what they specify r e l a t i v e l y open. He cannot draw conclusions as to the existence of spec i f ic invar iants p r io r to the appropriate tes ts . At the same t ime, one must of course grant that Gibson has some c lear , general views, both on the nature of invar iants and on the propert ies which w i l l prove to be spec i f ied. The views Gibson has can manage to' const i tu te constraints of a kind even though questions as to the existence of pa r t i c -ular invar iants and speci f icat ions are not known. Not even the question of types of invar iants fo r d i f f e ren t perceptual systems has been s e t t l e d , because Gibson comments tha t : There are also surely invar iants in the flow of acoust ic, mechanical, and perhaps even chemical s t imu la t ion , and they may prove to be closely re lated to the o p t i c a l , but I leave them to the readers' speculat ion. The study of invar iants is j us t beginning. ' I f invar iants were simply a rb i t ra ry propert ies one could postulate at w i l l , i t would be d i f f i c u l t to understand that of which the study of invar iants might consist . Even given a corrected in te rpre ta t ion of the Gibsonian concept of an opt ica l i nvar ian t , i t may s t i l l be tempting to think the ecological approach is susceptible to some form of t r i v i a l i z a t i o n problem. Consider-ing Gibson's descr ipt ion of v i sua l l y perceiving in terms of "looking at" o and "looking around", there is a reasonably strong basis fo r the Fodor/ Pylyshyn claim that visual perception is akin to t e l l i n g by looking. Te l l ing by looking can be regarded as the pickup of information by looking. The revised issue is whether or not Gibsonians are committed to saying any instance of the pick up of information by looking is based on the detection of the appropriate invar ian t . I f not , how does Gibson avoid supposing that some visual perception involves cognit ive processing a f ter a l l? I f so, then is not Gibson committed to the assumption of invar iants in f a i r l y un rea l i s t i c cases? The l a t t e r horn of th i s dilemma y ie lds a kind of t r i v i a l -i z a t i o n : although one is not postulat ing the existence of a par t i cu la r invar iant (as Fodor and Pylyshyn do with t h e i r "pseudo-invariants", "being a shoe" etc. . . . ) , one s t i l l assumes the existence of some invar iant or other fo r any example of information pickup by looking. An enormous number of instances of epistemic, visual perception ( fo r th i s expos i t ion, "seeing X as P", that may involve the visual system e i ther as a mere pa r t i a l determinant or not at a l l . Consider the statement "Daniel Ortega sees the p o s s i b i l i t y of a l i be ra l President as a chance to normalize re lat ions between the U.S. and Nicaragua". This statement refers to some non-perceptual judgement of Ortega's and does not make any use of the visual system at a l l . There are other cases, such as (from Fodor and Pylyshyn) Berenson seeing a paint ing as an authentic DaVinci or a meteorologist seeing black clouds over the ocean as an approaching hurr icane, which requires some form of expert knowledge, not generally possessed and perhaps not even learnable by human perceivers, in addit ion to the normal funct ioning of the visual system, for the pickup of cer ta in informat ion. One should th ink that cases such as these, cases in which the visual system only p a r t i a l l y determines the information which is acquired, would not be considered to be instances of d i rec t pickup. Yet, expert perceptions of meteorologists or a r t connisseurs, of which the perception of genuine DaVinci's or of approaching hurricanes are examples, are sometimes a mere matter of t e l l i n g by looking. I f one assumes that any information pickup by looking is the mere product of invar ian t -detec t ion , i t would seem that one must suppose that such perceptions are cases of d i rec t pickup, that there are invar iants to specify genuine Da V inc i ' s and the 1 ike. Examples of seeing X as P which do not involve the visual system at a l l ( that i s , where "sees" is being used in the sense of "understands", or " thinks o f " , f o r example) of course pose no problem here. I f seeing X as P does not involve the visual system at a l l , i t could hardly be described as an instance of t e l l i n g by looking. Cases in which the visual system is a pa r t i a l determinant are the d i f f i c u l t ones. The environment is f u l l of customary signs in v i r tue of which perceivers move from the contents of t h e i r f i e l d of view, from that which is manifest ly before t he i r eyes to information which goes well beyond t h i s . When th i s occurs unth ink ing ly , as i t so often does, i t const i tutes a case of t e l l i n g by looking - but i t is one in which the pickup of information is not exclusively determined by the visual system. For example, Wi l l may see that Abner is f i n a l l y plant ing his corn, when a l l there is for Wi l l to look at is Abner leaning on a fence post next to a f reshly furrowed f i e l d , bag of seed in one hand, c igaret te in the other. Abner actual ly put t ing seed in the ground is not before W i l l ' s eyes but the contents of his f i e l d of view is such t h a t , without any conscious computations, Wi l l picks up the information that Abner is in the midst of doing his p lant ing. Even though Wi l l undergoes no conscious process to get th is informat ion, i t would surely be correct to assume there are_ some kinds of mental processing involved. This is why i t would be a t r i v i a l i z a t i o n of a notion of "d i rec t pickup" or " invar iant " to suppose tha t , because Wil l picks up the information he does, there must be some invar iant specifying "man p lant ing" that W i l l ' s visual system detects th is when he looks at Abner rest ing against a fence post. I t sometimes seems that Gibson and his fol lowers are content to accept the in te rp re ta t ion that any instance of the pickup of information by looking is to be explained simply in terms of invar iant -detec t ion . They have occasion to express the expectation of the spec i f i ca t ion by invar iants of environmental phenomena which must appear to non-Gibsonians to be most unpromising candidates. Michaels and Carel lo, for instance, speculate on 9 the p o s s i b i l i t y of "very h igh- leve l " invar iants fo r aesthetic q u a l i t i e s . I t is important to understand such ind isc re t ions , however, as speculat ion, fueled pa r t l y by the enthusiasm that is often character is t ic of a r e l a t i v e l y new theoret ica l approach. A most in teres t ing case can also be made for suggesting that Gibson is not committed to the view that any case of information pickup by looking requires only the detection of the relevant invar ian t . Although a l l epistemic percept ion, according to the Gibsonian theory, is the pickup of informat ion, the reverse need not obtain. Hence, not a l l cases of i n fo r -mation pickup need be regarded as merely perception, and, in pa r t i cu la r , not every instance of information pickup by looking need be considered as merely the funct ioning of the visual system. Some information picked up by look ing, in other words, may not be visual perception, even though the visual system is a pa r t i a l determinant. The basic claim here is that only some information pickup by looking is the resu l t solely of the detection of some invar iant by an organism's perceptual system. On a Fodor and Pylyshyn in te rpre ta t ion of the ecolo-gical approach, th i s claim would raise two problems fo r Gibson. F i r s t , he would need to supply the pr inc ip le (s ) d i f f e r e n t i a t i n g between information that is d i r e c t l y picked up and information which is not d i r e c t l y picked up. Second, he would have to grant that such information as is not d i r e c t l y picked up involves cognit ive processing. The impl icat ion of the f i r s t point would be that Gibson lacks a coherent p r inc ip le of d i f f e r e n t i a t i o n , the impl icat ion of the second that he would have to concede some instances of perceiving involve mental processing a f ter a l l . The answer to the f i r s t problem begins with the assert ion that that which is d i r e c t l y picked up is that which invar iant properties specify. This asser t ion, which Fodor and Pylyshyn would no doubt f i nd inadequate, avoids being c i r cu la r or otherwise inherent ly defect ive because of a point mentioned previously. The question of what invar iant propert ies there are is an empirical one. I t is a question that is presumably to be answered through the study of natural perspective (with Gibson's emendations). I f the existence of invar iants is only s c i e n t i f i c a l l y determinable, then so too must be that which is speci f ied by invar iants and hence that which is d i r e c t l y picked up. Thus there i_s a general constraining p r i nc ip le : that which is d i r e c t l y picked is that which is speci f ied by invar iants . There is also a suggested method fo r determining par t i cu la r instances that explains why one cannot merely l i s t by re f lec t ing or int rospect ing the information about the environment that can be d i r e c t l y picked up: the determination of par t i cu la r invar iants and what they specify rests upon mutual tes t ing of perceptual and behavioral responses of various organisms and tes t ing fo r invar iant properties in structured energy. The a v a i l a b i l i t y of th is move can be missed fo r a couple of reasons. The main one is that Gibson does have, as part of his theory, very strong views on the terms in which the world is perceived. Affordance-perception w i l l prove to be basic according to Gibsonian doctr ine. This is the "new theory about what there is to be p e r c e i v e d " . ^ A key reason fo r the inclusion of a view about the terms of perception is that Gibson believes perceptual theory should take account of evolut ion. Hence, i t is a f a i r l y natural move fo r him to claim that the main categories of perception for an organism are l inked to i t s needs and a b i l i t i e s in re la t i on to i t s environment. Now the idea of perception being invariant-based makes some environ-mental propert ies bet ter candidates as possible objects of d i rec t percep-t ion than others. In p a r t i c u l a r , propert ies that can be l inked to a more or less res t r i c ted range of physical parameters are the only ones which hold the promise of proving to be associated with given pers is t ing ra t ios and re la t ions in the opt ic array. Because affordances are d isposi t ional properties objects possess re la t i ve to spec i f ic a b i l i t i e s and needs of organisms, they are j u s t such propert ies. I t could have helped Gibson's case i f he had mentioned th i s point but he unfortunately does not. Some examples of these "best case" properties would be "enclosure", "supporting surface", "opening", and " c l i f f " . In order fo r things in one's environment to be a place one can pass through or a place one can f a l l from, they have to meet some physical requirements. Where these require-ments are s u f f i c i e n t l y spec i f ic (some addi t ional l i m i t i n g pr inc ip les w i l l be mentioned l a t e r ) , there is i n i t i a l p l a u s i b i l i t y in the suggestion that a l l things having the affordance share some common invar ian t . Sameness in some tangible respect, constancy of s ize , pigmentation, or texture are also r e l a t i v e l y plausible candidates because i t is not too d i f f i c u l t to imagine 58 how opt ical invar iants might specify these. The predominance of properties character iz ing the environment in terms of u t i l i t y ( in funct ional terms) among the properties which are i n i t i a l l y plausible candidates for spec i f i c -at ion may help to render the Gibsonian move to "affordance-perception" more understandable . To r e i t e r a t e , in order fo r something to be an affordance fo r an organism with a given physical const i tu t ion depends on i t s meeting certa in physical spec i f i ca t ions . Everything of the affordance's kind w i l l there-fore share th is feature. Because there is some common, general feature amongst things which af ford a cer ta in kind of behavior fo r the organism in quest ion, there is something which could explain why they are a l l re lated to the same opt ica l s t ruc ture . Since there are best cases, there are also worse and worst ones - -propert ies picking out sets of objects for which i t is v i r t u a l l y impossible to imagine there being some associated invar iant . Many object names f a l l in to th i s category. Hence one could imagine a Gibsonian saying that "being a shoe" is not an ecological ly coherent category, meaning only that the extension of that property const i tutes so loose and ec lec t ic a co l lec t ion of items i t is un l ike ly in the extreme that there should be an invar iant 11 12 spec i f i c to "being a shoe". To adapt a point of Wi t tgenste in 's , there need be no fu r ther s igni f icance to a th ing 's having the property of being P than that i t is cal led a "P". Where th is is most l i k e l y the case, i t is most un l i ke ly that the set of phenomena should be speci f ied by some invar iant s t ruc ture . To repeat, the supposition that only some information pickup by looking is j u s t the resu l t of the detection of some invar iant (and hence const i tutes d i rec t pickup) raises two (related) questions. The f i r s t , as > discussed, is the issue of the pr inc ip le of d i f f e r e n t i a t i o n between i n fo r -mation pickup by looking which is d i rec t and that which is not. Only information that is in fact speci f ied by an invar iant property can be d i r e c t l y picked up. The second issue is that of how Gibson can avoid admitt ing some perception must involve cognit ive processing since some information pickup by looking is not merely the resu l t of invar ian t -detect ion. Fodor and Pylyshyn c lear ly would th ink Gibson could not f i nd a way around th i s problem. In considering how d i rec t pickup might be constrained, the authors suggest stronger and weaker forms of the ecological approach's basic thes is . The stronger corresponds to the claim that a l l information pickup by looking is d i rec t pickup. The weaker would correspond to the claim that only some information pickup by looking is d i rec t pickup. Fodor and Pylyshyn w r i t e : In shor t , the weak version of Gibson's claim is that there are some visual properties of the layout which are, to a f i r s t approximation, causally necessary and su f f i c i en t for properties of the l i g h t , which l a t t e r propert ies are themselves d i r e c t l y picked up. Our point has been that the Establishment theories say that too; in p a r t i c u l a r , the Establishment theories provide precisely that account-in the case of the sensory propert ies of the layout . . . Translat ing out of Fodor/Pylyshyn terms, th is passage has Gibson's "weaker claim" as being that only some features of the environment (proper-t i e s of the layou t ) , whose presence can be determined by looking, are speci f ied by invar iant properties in the opt ic array. Thus only some envir -onmental features are d i r e c t l y picked up. Fodor's and Pylyshyn's conten-t i on is that th i s weak, Gibsonian view amounts to a version of the Estab-lishment approach, at least wi th respect to the ro le of cognit ive process-ing in perception. In shor t , the c r i t i c i s m would be that Gibson simply claim that only some information pickup by looking is d i rec t pickup without caving in to the Establishment. Gibson does have a way of avoiding the envisaged problem, however, without saying that a l l information an organism picks up by looking is d i r e c t l y picked up. He can do th i s by d is t inguishing between information pickup by looking and visual perception (or visual information pickup: i t is convenient fo r my purposes to use th is expression as equivalent to visual perception, rather than as an equivalent of information pickup by looking) . The next move is to say that only cases in which there can be d i rec t pickup of information by looking can there be visual perception. There can only be d i rec t pickup of information by looking when there is some invar iant spec i f ic to the environmental feature. Awareness of the environmental phenomenon const i tutes the instance of information pickup in question. Gibson is^ committed to the claim that a l l visual perception is d i rec t pickup but th i s can be shown not to be the strong claim i t appears to be. Fodor and Pylyshyn, because they f a i l to see how d i rec t pickup can be constrained in terms of what ( in fac t ) invariants speci fy , take i t that anything about the environment that can be found out by looking is percep-t i b l e . Perhaps they side here with some notion of common usage: that which is perceptible is that which is commonly (co l loqu ia l l y ) said to be percept ib le. I f t h i s were Gibson's c r i t e r i a fo r deciding what is percep-t i b l e , then " a l l perception is d i rec t perception" would be the strong claim Fodor and Pylyshyn imagine. However, Gibson's c r i t e r i o n is that the features of the environment which are perceptible are those, and only those fo r which there are in fact invar iant propert ies. This may be regarded as a recommendation for a ( s t r i c t ) usage of "Perceives". In spi te of ordinary usage, only that which is d i r e c t l y picked up can be said to be perceived and only that which is speci f ied by an invar iant is perceptible (a matter to be se t t led by empirical inves t iga t ion) . Gibson's view is not, as Fodor and Pylyshyn perhaps th ink , that fo r everything one can o rd inar i l y be said to perceive, or for everything one can f i nd out by looking, there w i l l prove to be some associated invar iant . Hence the ecological approach avoids an absurdly strong reading of the kind Fodor and Pylyshyn suggest, without caving in to the Establishment by saying that some visual perception is d i r e c t , some not. One can make use here of the e a r l i e r point that some cases of informa-t ion pickup by looking, including ones which would sometimes be referred to as "seeing X as P", are only p a r t i a l l y determined by the use of a perceptual system. The d i s t i nc t i on between the involvement of the visual system as par t ia l and as exclusive determinant could be used as the basis of a d i s t i n c t i o n between those instances cal led "seeing X as P" which const i tu te "genuine" perception, so to speak, and those instances which const i tu te degenerate cases. Perceiving X as P is a genuine perceptual achievement i f and only i f an organism is perceptually aware of the very environmental feature, X's being P. This can obtain only when there is an invar iant which specif ies phenomena of the kind to which X's being P belongs. Degenerate cases are those in which one is perceptually aware of something e lse , such as X's being Q, from which awareness of X's being P is derived by some cognit ive process. Because the ecological approach is concerned with perception as such, a notion that produces semantic opaci ty , that an organism perceives X as P is no guarantee t h a t , even when "P" and "Q" are co-designative, i t therefore perceives X as Q. Moreover, the re la t ion between X's being P and X's being Q may vary. X's being P may be necessary and s u f f i c i e n t for i t to be Q in some cases. In others, X's being Q may be no more than some re l iab le or typ ica l ind ica tor . The epistemological distance in the association between X's being P and i t s being Q can vary immensely. Hence, some cases of being aware of X as P by being perceptually aware of X as Q w i l l involve del iberate inference/computations, making use of f a i r l y extensive special ized knowledge and t r a i n i n g . In other cases, the move w i l l be natural and more or less automatic. The degenerate instances are not to be dist inguished from the genuine on the basis of what information the theor is t feels he obtains automatically versus that which requires cognit ive e f f o r t . I f appropriate empirical tes t ing shows subjects' awareness of X as P can only be the resu l t of the detection of an invar iant spec i f ic ( fo r instance) to phenomena of X's being Q's k ind , then an organism can only be said to perceive X as P in some degenerate sense. In the proper, or s t r i c t sense of "perceives", the ecological theor is t should regard th is kind of "perception" of X as P as an i n t e l l e c t u a l , rather than a perceptual, achievement. TSRM contend, contrary to Fodor and Pylyshyn, that the only "gambit" 14 Gibson requires is the gambit of d i rec t perception. They fur ther claim that i t is the Establishment who i n s u f f i c i e n t l y constrain t he i r key notion 15 of perceiving. The charges may i n i t i a l l y appear to be a mere case of academic bravado but in l i g h t of the discussion of how Gibson constrains the percept ib le , TSRM's counterclaims may be understood. F i r s t , the gambit of "d i rec t perception" is j u s t the idea that only the information picked up by the mere detection of an invar iant counts as a genuine case of perception. Second, TSRM are suggesting that the Establishment, repres-ented by Fodor and Pylyshyn, o f fe r no workable c r i t e r i a of t he i r own fo r d is t inguishing between the perceptible and the non-perceptible. The basis of th is c r i t i c i s m would be that Fodor and Pylyshyn consider what TSRM would regard as degenerate instances of perception to be genuine ones. The par t i cu la r example which comes to mind is that of the perception of paintings as "authentic DaVincis". The general complaint could be that while t r a d i t i o n a l approaches acknowledge that some instances described as perceiving are not exclusively determined by the operation of a perceptual system, they have no agreed-upon, sat is fac tory account of the d i s t i nc t i on between such instances and those which are exclusively determined by the funct ioning of a perceptual system. There is a possible irony in th is understanding of the ecological approach. Fodor and Pylyshyn ca l l object recognit ion a "perceptual process par excellence" , ^ with perception of shoes (as such) being a per fec t ly good, typ ica l example of i t . However, i f the c r i t e r i o n fo r determining the l i m i t s and categories of the perceptible is that for which there are associated invar ian ts , Fodor and Pylyshyn's example of a perceptual process par excellence may well turn out to be an i n te l l ec tua l achievement, not a perceptual one. The prospect of discovering that which is genuinely perceptible via experimentation raises an important problem as we l l . This in te rpre ta t ion of Gibson c lear ly carr ies a recommendation fo r a l te r ing the use of "perceives" not in ordinary language but at least in theoret ica l communities. I t is a recommended change of usage in the sense of suggesting a c r i t e r i o n of d i f f -e ren t ia t ion that stands an extremely good chance of doing violence to conventional views of which environmental phenomena are genuinely percep-t i b l e and which are not. Embedded in a recommendation of the envisaged kind are promises to the e f fec t that acceptance of i t w i l l y i e l d some s i g n i f -icant benef i t , that i t w i l l do more good than harm. One of Gibson's spec i f i c promises is that counting only invariant-based perception as genuine w i l l not too severely l i m i t the class of environmental phenomena which actual ly turn out to be percept ib le. As U l r i c Neisser notes, the "cruc ia l invar iants have not yet been i so la ted " , and that "the claim that they ex is t is the largest outstanding promissory note in ecological o p t i c s " . 1 7 Another promise is that the invar iants that w i l l turn up w i l l be s u f f i c i e n t to explain the evolutionary success of organisms and the i r behav-i o r . Gibson's main complaints against the Establishment is that t he i r 18 theories pay i n s u f f i c i e n t regard to ecology and evolutionary theory. This is supposed to resu l t in theories which cannot explain adequately how perception resul ts in organisms having any knowledge about t he i r environ-ment at a l l . There must therefore turn out to be invar iants su f f i c i en t to explain the or ig ins of the knowledge various organisms evidently possess about t he i r environments. Also, given the concern fo r evolut ion and the eco log ica l , there is a promise that invar iants w i l l produce ecological ly useful information for organisms. There is a promise that they w i l l specify the environment in terms of i t s needs, a b i l i t i e s and vu lnerab i l -i t i e s . This point about ecological concern is another part of the explan-at ion fo r Gibson's in terest in affordance-perception). The ecological approach w i l l u l t imate ly stand or f a l l based on i t s a b i l i t y to show that there is an abundance of invar iants that specify the environment in useful terms to various organisms. ( i i ) To summarize, Gibson's notion of that which may count as the d i rec t pickup of information is constrained. Any instance of perception is one of d i rec t pickup of information. Direct pickup must be the mere resu l t of the detection of the relevant invar ian t property by an organism's percep-tual system. I t is very much worth keeping in mind, however, that th is constraining p r inc ip le says nothing at a l l about what is s p e c i f i c a l l y entai led by the process of "d i rec t pickup", or , a l t e r n a t i v e l y , that which const i tutes the mere detection of an invar iant . As matters were l e f t at the end of chapter two, the re la t ion between information about the environment and invar iant propert ies of the opt ic array was vague. As a r e s u l t , i t is not clear what must obtain in order fo r an organism to get from the detection of invar iant s t ructure in the array by i t s visual system to the pickup of some in fo r -mation (awareness of some feature of the environment as such). Fodor and Pylyshyn are among a number of c r i t i c s who have raised questions in t h i s area. The general suggestion is that Gibson's theory is 19 in some sense incomplete, or that i t leaves a "gap". On the Fodor/ Pylyshyn vers ion, i t is claimed that information about the environment is l i t e r a l l y in^ l i g h t only in the sense that informative properties ( that i s , invar iants) are in l i g h t . Informative proper t ies, in other words, are simply s t ruc tura l features of l i g h t . They w r i t e : So, fo r example, the frequency of the l i g h t can cause a state o f a detector, and the frequency of the l i g h t can be defacto informative about the color of re f l ec t i ng surfaces in v i r t ue of a cor re la t ion that holds between frequency and color . But the fac t that the frequency of the l i g h t is correlated wi th the color of r e f l e c t i n g surfaces cannot i t s e l f cause a state of a detector, and appeal to that fac t exhausts Gibson's construal of the notion that the l i g h t contains information about the color of surfaces. So we are back to the old problem: how (by what mental processes) does the organism get from the detection of an informative property of the medium to^the perception of a correlated property of the environment? *® Fodor and Pylyshyn's claim is that no a l te rnat ive to "perceptual in ference" , as they ca l l i t , has yet been suggested. This apparent gap in Gibson's theory, as the authors go on to explain i t , consists in the absence of an account of the i n t e n t i o n a l i t y in (epistemic) perception. This issue quite r i g h t l y takes up the bulk of TSRM's reply to Fodor and Pylyshyn. They argue that Gibson's "theory of spec i f i ca t ion" provides the basis fo r a new theory of i n t e n t i o n a l i t y . In remaining chapters, I w i l l consider TSRM's attempt to defend the ecological approach, arguing that th is defense is not successful. Their discussion nevertheless suggests a way of understanding the Gibsonian approach which shows i t can accomodate some account of i n t e n t i o n a l i t y . This w i l l involve construing Gibson as what might be cal led a "perceptual behavior is t " . I use th is t i t l e to describe theor is ts who analyze the perceptual awareness of organisms in terms the acquis i t ion of behavior-states. Two general forms, the more sophist icated of which might more properly be considered a var ie ty of funct ional ism, w i l l be described in Chapter s i x . Chapter Four As I noted at the end of las t chapter, Fodor and Pylyshyn are among a group of c r i t i c s who suggest that Gibson's account of perception does not explain something important. Di f ferent attempts have been made to make i t clear what is supposed to be missing, but none of these have found much sympathy from Gibson's fo l lowers. Fodor and Pylyshyn i n i t i a l l y explain the gap they feel as the lack of an explanation of "how an organism gets from the detection of an informative property of the medium" to the per-ceptual pickup of information about the environment. They la te r develop t h e i r c r i t i c i s m of Gibson as the claim that he has no "theory of in ten t ion-a l i t y " , or no account of the in tent ional component in epistemic perception:. Everybody has to face the issue about i n t e n t i o n a l i t y somewhere. For Gibson, push comes to shove with the question: What is i t fo r an event (a conf igurat ion of the l i g h t , e tc . ) to specify a property? To say that Gibson has no theory of i n t e n t i o n a l i t y is to say that he has no answer to that question. Or, to put i t the other way around, the f a i l u r e of Gibson's theory of spec i f ica t ion is no minor flaw in his theory. I t marks the precise point at,which Gibson's treatment of i n t e n t i o n a l i t y proves to be bankrupt. D.W. Hamlyn expresses essent ia l ly the same concern by w r i t i n g : . . . when an object in a given context af fects a perceptual system in such a way that information is derived about i t because of the structure of s t imula t ion , the perceiver is enabled to see the object in a cer ta in way, as a such and such. I t is impossible fo r something to see something as X unless i t has some idea of what i t is fo r something to be an X. To say th i s is to say that i t must have in some way, and to some extent , the concept of X. Thus to speak of i t as obtaining information is not i ^ f a c t to ru le out as unnecessary any reference to concept. The point which Fodor and Pylyshyn make and which Hamlyn makes is roughly the same. There are only two di f ferences. F i r s t , Fodor and Pylyshyn re fer to representations where Hamlyn refers to concepts. Second, Fodor and Pylyshyn are s l i g h t l y more conc i l ia tory than Hamlyn. Hamlyn appears to suggest that the ecological approach simply cannot do without concepts/representations. Fodor and Pylyshyn argue only that the Gibsonians have in fact given no a l te rnat ive means of accounting fo r the in tent ional component. They leave open the p o s s i b i l i t y that the ecological approach could come up wi th some a l te rna t i ve . To repeat, the underlying theme in both Hamlyn's and Fodor and Pylyshyn's remarks is the same. I f Gibson were merely taking account of the extensional concept of simple perception, his view would be in no d i f f i c u l t y over the i n t e n t i o n a l i t y issue. The ecological theory is in p r inc ip le a per fec t ly adequate explanation of the extensional notion of how perceptual systems detect s t ructure in the ambient array, for instance. The perceptual pickup of informat ion, which is perceptual awareness of environmental features as such is c lear ly an in tent ional not ion. One way to put th is is that perception of X as P requires something to mean "P" to the 'perceiver . The alleged gap in Gibson's theory may now be thus explained: the ecological approach explains the detection of an invar iant property by an organism's perceptual system, but not how a case of detec-t i o n manages to mean anything to the perceiving organism. The substant ial reply given by TSRM to Fodor and Pylyshyn's c r i t i que of the ecological approach consists predominantly of an attempt by the authors to c l a r i f y Gibson's view so as to i l l u s t r a t e why i t is not suscep-t i b l e to the "gap" problem. A considerable amount of philosophical machinery is brought to bear on th is issue by the four Gibsonians, but there are essent ia l ly two notions upon which t h e i r re-explanation of Gibson's view res ts . These are "natural kinds fo r animals" and "natural law". I w i l l argue that the philosophical terms and theories which are meant to ground TSRM's explanation, however, do l i t t l e or no work because they are misapplied or the authors hold other points inconsistent with them. The theories to which TSRM appeal are Hi lary Putnam's theory of 3 4 natural kinds and Fred I . Dretske s theory of natural laws. Because the appeal to these two theories is not the kind of thing TSRM need to explain how information can be picked up by mere invar iant detect ion, they f a i l to explain how the ecological approach avoids the "gap" problem. However, i t is possible to in terpre t Gibson as avoiding t h i s problem by accounting for that which the detection of an invar iant means to an organ-ism in terms of behavior states of some sor t . Such an understanding of the ecological approach is suggested, perhaps only un in ten t iona l l y , by parts of TSRM's discussion. Fodor and Pylyshyn suppose that perception of X as P has an intent ional component that needs explaining. They give th i s explanation in terms of a perceiving organism's re la t ion to a (mental) representation of X's being P ("concept o f " , " in tens ion" , or "meaning" could a l l be subst i tuted fo r "representation of" as i t is used here), suggesting that any perceptual theory needs to subscribe to th i s view of i n t e n t i o n a l i t y , or to another that does the same work. So, for example, an organism's perception of X as P would require the detection of the appropriate corresponding invar iant which, though cogni t ive processes of cer ta in sor ts , eventuates a represent-at ion of X's being P ( in some sense, the apprehension by the organism of the intension that X (something) is P). The a l te rna t ive which seems to be open to the ecological t heo r i s t s , which is by no means novel, is to make a reduct ion is t move: that the detection of some invar iant means "X is P" to an organism is to say only that i t s detection produces cer ta in sorts of behavior, or behavior-producing states in the organism. As I w i l l l a te r suggest, the ecological approach to perception is very l i k e D.M. Armstrong's so-cal led "Bel ie f - theory" (where "be l ie f " is analyzed as "states apt fo r the production of cer ta in sorts of behavior") . Further deta i l of th is w i l l be given a f te r an evaluation of TSRM's own attempt to avoid the gap. ( i i ) To begin, there are some t roubl ing points about the logic of portions of TSRM's commentary. Part of t h i s is a t t r i bu tab le to a less than adequate grasp of terms as they would t y p i c a l l y appear in philosophical discussion. Their d i f f i c u l t i e s are also a t t r i bu tab le to TSRM's f a i l u r e to appreciate the impl icat ion of statements they make. The l a t t e r is evident in a c r i t i c i s m directed at Fodor and Pylyshyn, for.example. TSRM spend a considerable length of tiTfie. c r i t i c i z i n g what they re fer to as Fodor and Pylyshyn's "argument from the philosophy of science". Since th is is a t i t l e of TSRM's choosing, and since there is no spec i f ic reference given fo r the Fodor and Pylyshyn a r t i c l e , i t is not clear what TSRM take the argument from the philosophy of science to be. What the argument may be, TSRM's c r i t i c i s m is that i t is based on notions which are too cont rovers ia l : I t is pa in fu l l y obvious . . . that no substantive argument can be b u i l t from the notions of k inds, p ro jec t ib les , laws, and counterfactuals given the current state of the a r t . In the p h i l -osophy of science these notions are notor iously opaque and notor-iously uneven in t h e i r usage and are commonly recognized as such without undue embarrassment . . . insofar as Fodor and Pylyshyn have chosen to ground t h e i r argument in the philosophy of science, we feel i t incumbent upon us to show that that foundation is porous - les t the philosophical wool.,J>e pulled; over the eyes of the non-philosopher.? This c r i t i c i s m is a hopelessly general one of a kind that TSRM cannot af ford to make. There is nothing wrong wi th making use of not ions, when theor iz ing in one area of philosophy, that are s t i l l subject to debate in other areas: TSRM ought to show spec i f i ca l l y how the contentiousness of a given term actual ly bears on Fodor and Pylyshyn's c r i t i c i sms of Gibson. At best, they show only that some equivocation may be occurring in Fodor's and Pylyshyn's use of " law", and of "natural k inds". These points are unconvincing and are not developed. The incredib le part about TSRM's argument is that t h e i r own defence of the ecological approach and t h e i r objections to the Establishment are based on some of the very notions which they claim provide too porous a foundation for "substantive argument". Add i t iona l l y , Shaw, Turvey and Mace have elsewhere attacked t r a d i t i o n a l views and defended t h e i r own in terms of the notion of a proposit ion and of a possible wor ld .^ These two ideas are subject to no small amount of controversy in t h e i r own r i g h t . I f Fodor and Pylyshyn must be discounted fo r t h e i r use of concepts whose meaning and the nature of whose referents is not at a l l decided, then so too should be TSRM. In the in terest of avoiding closure of t h i s , and most other philosophical discussion, the kind of general complaint TSRM make ought to be dropped. ( i i i ) Anyone whose background is predominantly in philosophy (perhaps non-ecological psychology as wel l ) is bound to be somewhat confounded by the language found in TSRM's a r t i c l e . There are some words, "proper ty" , " i n tens ion" , "extension", and "natural kind" being chief instances, which often f ind a place in philosophical discussions and are also employed by TSRM. Their use of these words are of course the subject of some contro-versy. Nevertheless, there are accepted general usages for each, however rough. In places, i t is not clear at a l l that TSRM are fo l lowing any general p rac t i ce , but are instead usurping i t by employing reasonably fami l ia r terms in ways that merely su i t the author's purposes. Also, in at least one instance TSRM are f l a t l y mistaken in t h e i r employment of one of these not ions. Before proceeding, i t is necessary, fo r purposes of subsequent discussion, to b r i e f l y describe a study mentioned in TSRM's a r t i c l e . The study in question concerns the visual perception of plant stems by marsh per iwinkles. TSRM are fond of re fe r r ing to i t as an i l l u s t r a t i v e example fo r various points they make. The marsh periwinkle is a small snai l found only in upper i n t e r t i d a l zones containing vegetation. For the most par t , the per iwinkle 's a c t i v i t i e s consists of making i t s way, on the ground, amongst and around the plant stems inhabi t ing i t s hab i ta t . When the t ide advances, however, the snai l climbs the plant stems to avoid water-bound predators. In a we l l - con t ro l led study, Paul V. Hamilton f inds convincing evidence to suggest t ha t the per iwink le 's a b i l i t y to locate plant-stems is v i sua l l y cont ro l led : Snails released on the substrate amid plant stems, just p r io r to the release area being inundated by the advancing t i d e , moved in a r e l a t i v e l y s t ra igh t path and usually (67%) ascended the closest plant stem. The propensity of these snai ls to ascend the closest stem was shown not to resu l t from co l l i s ions resu l t ing from snai ls t r a v e l l i n g a s t ra igh t path in a random d i rec t i on . Field experiments involv ing transparent and black rods of equal size showed7a s ign i f i can t tendency for snai ls to move toward black rods. The high percentage of instances in which the periwinkle selects the nearest plant stem suggests a non-random process of o r ien ta t ion . The high propor t ion, in the subsequent f i e l d study, of selections of black rods ( f o r t y - f o u r black, s ix transparent, with ten snai ls reaching the perimeter of the plexiglass arena in which the rods were placed, out of s i x ty t r i a l s ) suggest that i t is v is ion which guides the per iwinkle. TSRM re fer to th is study as suggesting that when the periwinkle selects and climbs a plant stem, i t is v isua l l y perceiving something as climb-upable. When i t moves around the stems (when the t ide is not advancing), i t v i sua l l y perceives them as a ba r r ie r . Hamilton himself does not speculate on the spec i f ic content of the sna i l ' s perception of i t s environment. In p a r t i c u l a r , he makes no mention of perception of ba r r ie rs . This causes no special problems fo r TSRM but i t should be made clear that Hamilton's study, which TSRM do not describe, receives a cer ta in amount of embellishment in order to serve the ecological theor i s ts ' purposes. One use to which the authors put the marsh periwinkle example is in an object ion to the Establishment approach to perception. TSRM claim that the Establishment has no coherent explanation of why the periwinkles climb plant stems in one instance and avoid them in others. They claim th i s is because Establishment views take an "extensional" view of perceiving (such that i f one perceives some property, one perceives any-th ing coextensive with i t ) . Why th i s is supposed to be a feature of t r a d i t i o n a l approaches generally is not c lear. In the marsh periwinkle s i t u a t i o n , the idea is t h a t , in the i n t e r t i d a l zone in which the snai l is found, climb-upable things and impassable things (bar r ie rs) happen to be co-extensive. TSRM assume, because the Establishment view allows the subs t i tu t ion of co-extensive terms (apparently the authors mean in general ) , a marsh periwinkle always perceives "c l imb-upab i l i t y " and " impass ib i l i t y " together. That i s , i t s perception of things in i t s environment is i nd i f f e ren t between perception of things as climb-upable and perception of them as ba r r ie rs . The f i r s t mistake TSRM make is in t h e i r understanding of coextension-a l i t y . They w r i t e : Let us say that for a thing to be a bar r ie r i t must have the propert ies p,q, r . That i s , (p ,q , r ) is the intension b of " b a r r i e r " . And l e t us say that fo r a thing to be a climb-upable thing i t must have the properties s , t , u , v . That i s , ( s , t , u , v ) is the intension c of "cl imbable". The extension of b ( in the i n t e r t i d a l zone) is the plant stems and other sna i ls . The extension of c ( in the i n t e r t i d a l zone) is the plant stems (other snai ls being unwi l l ing and too short to comply). Thus c is coextensive with b.° The same idea is repeated a l i t t l e l a te r in a table which purports to 9 make the "Establishment analysis" of the marsh periwinkle case e x p l i c i t . The in tent is to show t h a t , on t r a d i t i o n a l views, the inputs to the sna i l ' s visual system is not spec i f ic to "climb-upable" and "ba r r i e r " . The ecol-ogical theor is ts conclude that the sna i l ' s a b i l i t y to climb plant stems sometimes and avoid them at other times can only be explained by the Establishment by a t t r i b u t i n g inappropriately sophist icated cognit ive proc-esses to the per iwinkle. F i r s t , i t is f a i r l y obvious that "c l imb-upab i l i t y " and "bar r ie r " are not coextensional: TSRM's example contains a mistake. The extension of c in the i n t e r t i d a l zone, since i t is comprised only of plant stems is included in the extension of b, comprised as i t is of both plant stems and other sna i l s , j u s t as the class of spouses includes but is not coextensive with the class of husbands. I f TSRM have some va l id point to make against the Establishment, the example could, of course, be a l te red. However, that TSRM should make such a mistake over the r e l a t i v e l y simple notion of coextensional i ty , while using i t to make a point they c lear ly regard as important, is rather a blow to the authors' c r e d i b i l i t y . One cannot help but be a l i t t l e suspicious of TSRM's understanding of the other, more d i f f i c u l t phi losophical concepts that they employ. In add i t ion , the c r i t i c i s m TSRM are attempting to make carr ies l i t t l e weight. The universal p r inc ip le of s u b s t i t u t i v i t y of coextensive terms 75 that the authors a t t r i b u t e to the Establishment is fa r too strong. V i r t u a l l y any philosopher who has thought about questions of subst i tu t ion recognizes the existence of some opaque contexts in which coextensive terms may not be subst i tu ted. One of the more common of these contexts, and i t i s one which Fodor and Pylyshyn e x p l i c i t l y recognize, is that of epistemic perception. Therefore, i f a blanket p r inc ip le of s u b s t i t u t i v i t y of co-extensive terms is taken by TSRM to be some essential tenet of the Establishment pos i t i on , they are object ing to a straw man, to a theory to which few, i f any, would be committed. Another d i f f i c u l t y ar is ing in TSRM's a r t i c l e is the manner in which the authors jumble together the terms "property" (and "parameter") and " in tens ion" . For example, they wr i te that " . . . the in t imat ion is that a law relates intensions (propert ies and magnitudes) rather than extensions (domains of propertyless ind iv iduals) and also that "the predi lec-t i on fo r extensionalism is sustained, one is t o l d , by the f a i l u r e to provide a c r i t e r i o n by which two proper t ies, two intensions, can be judged the same". 1 1 There are legi t imate and i l l e g i t i m a t e ways of using " intension" as an al ternate fo r "property" . I t is legi t imate to equate the two terms when "property" is taken in the abstract or universal sense, as that which is ins tant ia ted by some set of ind iv idua ls . Another way of def ining an abstract property ( to use an example of Fodor and Pylyshyn's) is as that which is expressed by an open sentence, such as " is a shoe". An intension can be very generally understood as that which is expressed by a given form of words. I t could thus be considered a property in the general sense when-ever the form of words expressing i t is an open sentence. However, the term "property" is sometimes also used in a non-general sense of property-instances or , put another way, as par t icu lars that possess. Properties as par t icu lars cannot sensibly be i d e n t i f i e d with intensions. Yet, t h i s is what TSRM must do since, in spi te of using "property" and " intension" interchangeably, they deny the existence of abstract , universal proper t ies , saying there are only "propert ied th ings" . This denial of TSRM's w i l l come up again in connection with t h e i r endorse-ment of Dretske. The absence of a proper d i s t i n c t i o n between d i f f e ren t notions of "property" seems to lead TSRM to a t t r i b u t e the view that the world only contains bare, or "propertyless" indiv iduals to Establishment theo r i s t s . According to TSRM, t r a d i t i o n a l views consider the intension ("c") of "c l imb-upab i l i t y " to be a mental representation or concept. This is at least t rue of Fodor and Pylyshyn. Since TSRM. also equate properties ( in any sense) with intensions, they conclude that the t rad i t i ona l approach must consider properties to be mental representations. TSRM would then proceed to argue that these in ternal representations are a t t r ibu ted to the environment and are not features, hence not perceptible features, of i t . Therefore, the Establishment is committed to the notion that only bare, propertyless indiv iduals are perceived. This is a very strange claim given that there is an abundance of examples of t r a d i t i o n a l perceptual theories on which propert ies are not only percept ib le, but serve as the basis of perception. In defense of the Establishment, two points ought to be kept in mind. F i r s t , there are two d i f f e ren t notions of perception: simple and epistemic. Second, acceptance of universal propert ies does not automatical ly preclude the postulat ion of some p a r t i c u l a r i s t i c sense of "property" (as instances or as ins tant ia t ions of the un iversa l , fo r example) as we l l . Now in the simple sense of "perceives", anyone, whether he is an ecological theor is t or not, who allows that there are propert ies as par t i cu la rs may grant that organisms (simple) perceive propert ies. The theor is t may do so independently of whether or not he add i t iona l l y believes there are propert ies in some general sense. The simple sense of "perceives", however, is not the sense which should be of concern to TSRM. On the other hand, in the epistemic sense of "perceives" (which is i n t e n t i o n a l , unl ike the simple sense), organisms are not related to par t i cu la r instances of properties anyway. To perceive something as climb-upable is to recognize a par t i cu la r thing as being of a k ind, so the re la t ion is between a perceiving organism and the general property of " c l imb-upab i l i t y " . Essent ia l ly , TSRM need to be much more clear on the i r conception of propert ies. They need to develop a better understanding of what can be sensibly claimed about propert ies as par t icu lars (they are not intensions and they reside in the environment) and what can sensibly claimed about universal propert ies (they are intensions and they do not reside in the environment). ( i v ) Because TSRM confuse two d i f f e ren t notions of "property" and take the misleading resu l t to be interchangeable with " in tens ion" , they tend to s h i f t between ta l k about simple perception of propert ies and ta l k about epistemic perception of things as having some property. One reasonably clear example of the d i f f i c u l t i e s that can ar ise occurs when TSRM favour-ably re fer to an argument of Gibson's about affordances. Gibson reasons: . . . i f there is information in l i g h t fo r the perception of surfaces, is there information the perception of what they afford? Perhaps the composition and layout of surfaces con-s t i t u t e s what they a f fo rd . I f so, to perceive them is to perceive what they afford.13 Gibson goes on to claim that th i s is a "radical hypothesis" according to which values or meanings are d i r e c t l y perceived. There are a couple of separate issues raised by th i s passage. F i r s t , Gibson's reasoning in the conclusion of the foregoing passage, to perceive the composition and layout of surfaces of things is to perceive what they a f f o r d , requires a notion of "perceives" that allows t ruth-preserv ing sub-s t i t u t i o n . Only the simple sense allows t h i s , and i t does so because i t i an extensional sense of the term. Perceptual information pickup is in ten-s iona l . I t does not admit of the requis i te kind of subs t i t u t i on . More-over, TSRM e x p l i c i t l y assert that the ecological approach does not want subst i tu t ion of coextensive terms in perception contexts. With regard to Gibson's overal l theory, th is argument therefore re l ies on an equivocation of which TSRM are also g u i l t y because they endorse the Gibsonian argument. In addit ion to the slippage in uses of "perceives", there is a second sort of equivocation in TSRM's a r t i c l e about which one must be care fu l . I t sometimes appears as though TSRM are taking affordances themselves (which are d isposi t ional propert ies) to be objects of perception, instead of taking the (actual) things having the affordance to be the object. Taking affordances to be "poss ib i l i t i es " fo r ac t i on " , they w r i t e : . . . the ecological approach, with i t s commitment to r e a l i s m . . . , focuses on real p o s s i b i l i t y ; for i t takes p o s s i b i l i t y to be an ontological category . . . Poss ib i l i t i es fo r action or , more prec ise ly , things wi th p o s s i b i l i t i e s for ac t ion , are among the kinds of things that populate an animal's niche and are, therefore , things to be seen or heard or smelt e t c . W This move between " p o s s i b i l i t i e s fo r act ion" and "things wi th pos-s i b i l i t i e s " is no i nd i f f e ren t matter. The claim that some non-occurent property, the p o s s i b i l i t y of being climbed, is i t s e l f simply perceived is rather d i f f e ren t from the claim that things having the p o s s i b i l i t y are perceived (and in such a way that the organism may thereby be said to perceive the object as having the property in quest ion). The assertion that a marsh periwinkle perceives (the intension of) "c l imb-upabi l i ty " is a qui te misleading way of saying that i t perceives some environmental thing (a stem) having the affordance as a climb-upable th ing . The s l ide between these two claims, which is no doubt encouraged by the confusion of "proper ty" , in the sense of some kind of pa r t i cu la r , and " in tens ion" , should be avoided. One is owed some explanation of how the detection of an opt ica l (or other) invar iant associated with things which are climb-upable fo r the snai l in the i n t e r t i d a l zone manages to mean "climb-upable th ing" to a per iwinkle. I t would be easy to lose sight of t h i s fac t i f the ecological theor is ts are permitted to t a l k too f ree ly as though the meaning ( intension) of "c l imb-upab i l i t y " was i t s e l f simply perceived. This second issue raised by Gibson's argument is bas ica l ly the issue of what sense is to be made of the assertion that (actual) composition and layout const i tu te (d ispos i t iona l ) affordances. There are two d is t ingu ish-able concerns. One is how th is claim may be turned in to a reasonable view about affordances (qua d isposi t ional proper t ies) . The second, which w i l l mainly be dealt wi th in the next chapter, is how th i s view explains how things having par t i cu la r affordances are perceptible as such. TSRM's view of how environmental substances and layout of surfaces could be understood as const i tu t ing affordances amounts to a f a i r l y stan-dard sort of modal real ism. Even though th is in te rp re ta t ion makes Gibson's use of "const i tu tes" a very queer one, TSRM's view f i t s very well with Gibson's discussion. TSRM's view of affordances is that the affordances that things possess they possess in v i r tue of t he i r actual manifest (occurrent) s t ruc ture . In shor t , the non-occurrent properties a thing holds are grounded in i t s occurrent propert ies. I t has been pointed out to me that t h i s p r inc ip le may not be universal ly t rue , that i t may not hold fo r subatomic pa r t i c l es . I t does, however, work at the macroscopic l e v e l , and since the ecological approach is confined to considering the physical world at the environmental l e v e l , i t w i l l not matter that the p r inc ip le is not generalizable to include basic, physical pa r t i c l es . A neu t ra l , physical example that TSRM use to indicate what they mean by saying that d isposi t ions (there is no relevant d i s t i n c t i o n to be made here between "d i spos i t i on " , "affordance", and " p o s s i b i l i t y " ) are grounded in the occurrent is that of the s o l u b i l i t y of s a l t . Salt has the property of s o l u b i l i t y because i t is comprised of l a t t i c e s of ions, bonded by e lec t r i ca l a t t r a c t i o n s , the values of which are substant ia l ly reduced by the high d i e l e c t r i c constraints of cer ta in l i q u i d s . TSRM's account of affordances is along the very same l ines : A climb-upable thing must possess a cer ta in r i g i d i t y , a cer ta in surface area, a cer ta in height, a cer ta in textual q u a l i t y , e t c . , to support the climbing of the snail and the snail must be of a cer ta in mass, i t s mucous of a cer ta in v i scos i t y , i t s ventral . f i surface of a cer ta in f l e x i b i l i t y , e t c . , to e f fect the cl imbing. A stem's being climb-upable fo r the s n a i l , in other words, consists in i t having the appropriate occurrent physical dimensions. There is only one complication in the notion of affordances grounded in occurrent propert ies over the explanation TSRM giveof d isposi t ional propert ies in general. This is that the common, physical structure among members of the affordance must somehow also produce an invar iant speci f ic to i t in order fo r the affordance to'be perceived as such. (v) Affordances. const i tu te what TSRM ca l l natural kinds fo r animals. The notion i t s e l f f a i r l y s t ra ight forwardly serves to fu r ther r e s t r i c t the classes of environmental phenomena which must be considered as speci f ic to be a given, i nva r ian t , as I w i l l explain a l i t t l e l a te r . I t is l i k e l y to produce a certain amount of perp lex i t y , however, because the use of "natural kind" by TSRM is an atypical one. Moreover, the authors claim several times to be using the term in accordance wi th the views of Hi lary Putnam when i t is neither clear what they take to be Putnam's use of "natural kind" nor that what Putnam in par t i cu la r says is relevant to the i r case. As is far too common in the TSRM a r t i c l e , there is no proper reference to indicate what is said by Putnam with which the authors f ind agreement. The only concrete a t t r i b u t i o n to Putnam occurs when TSRM w r i t e : Freeing the conception of a natural kind from the requirement of inclusion in a natural law is the tack taken by Putnam (1970a, b ) : a natural kind term merely serves to draw commonalities among things that are s u p e r f i c i a l l y d i f f e r e n t ; i t is a s c i e n t i f i c convenience and an in ten t iona l l y temporary one at that.1? Three ideas are a t t r i bu ted to Putnam in th is quote: 1. Natural kinds need not be included in natural laws 2. Natural kind terms only serve to draw at tent ion to commonalities amongst the s u p e r f i c i a l l y d iss imi la r 3. Natural kind terms are " i n ten t iona l l y temporary" In " Is Semantics Possible?" (TSRM's reference is "(Putnam, 1970a)"). Putnam introduces his discussion of natural kinds by wr i t i ng tha t : An important c lass, phi losophical ly as well as l i n g u i s t i c a l l y , is the class of general nouns associated with natural kinds — that i s , with classes of things that we regard as of explanatory importance; classes whose normal d is t inguishing character is t ics are "held together" or even explained by some deep-lying mechan-isms. jgGold", "lemon", " t i g e r " , "acid" are examples of such nouns. Some examples of "deep-lying mechanisms" are "proton donor" for acids, a pa r t i cu la r DNA structure for t igers or lemons, atomic weight and number for chemical elements l i ke gold. TSRM say that on Putnam's view natural kinds are not required to be included in natural laws. There seems to be two possible in terpretat ions of th is claim. The f i r s t is that what counts as a natural kind is not determined by that about which the current s c i e n t i f i c community have (discovered or postulated) natural laws. Since natural kinds are taken by Putnam to be classes held together by non-superf ic ial s t ructure (mechanisms) which sc ien t i s ts may or may not have discovered, th is in te rpre ta t ion of 1) is very l i k e l y t rue . TSRM want to postulate natural kinds ( in some sense) about which there are presently no expressed natural laws, so there is some basis for th ink ing the authors do_ mean that a kind is not a natural one in v i r tue of having laws postulated about i t . The a l te rnat ive i n te r -pretat ion is that there need not ex is t any natural laws (discovered or not) concerning a given natural kind. The kinds of deep-lying mechanisms which would make natural kinds sui table subjects fo r laws. Thus t h i s second i n t e r p r e t a t i o n , that natural kinds need not be in fac t included in natural laws, is surely not Putnam's view. Regarding 2 ) , TSRM seem to be mistakenly taking the existence of "super f i c ia l di f ferences" among members of a natural kind to be some key feature of natural kinds. Perhaps the authors have not qui te grasped the 19 points which Putnam attempts to make? Putnam uses d i f f e ren t examples (green lemons, three-legged or non-striped t ige rs ) in an attempt to stress the point that since members of natural kinds could be, and sometimes are observably d i f f e r e n t , i t is implausible to hold that membership in natural kinds is determined by s u p e r f i c i a l , observable propert ies. This point 20 is made in support of a view about the meaning of natural kind terms. Whether or not members of the extension of a general noun do happen to be s u p e r f i c i a l l y d iss imi lar is quite i r re levant to the question of whether that noun picks out a natural k ind. The meaning of a natural kind term (and the class i t picks out) therefore cannot be f ixed in terms of s u p e r f i c i a l , observable s i m i l a r i t i e s between th ings. The a t t r i b u t i o n of 3) to Putnam may embody the very same kind of confusion. To say that the use of some term is " i n ten t iona l l y temporary", one must assume, means only that i t s meaning is subject to change. I t is part of Putnam's view about natural kind terms that t h e i r meanings can_ change but i t is not Putnam's view that changes in meaning are e i ther a common or d is t inguishing character is t ic of natural kind terms. Putnam says that which is conveyed by a natural kind term is i t s extension and 21 an associated stereotype. The meaning of such a word can change because fo r example, s c i e n t i f i c developments can bring about a change in the assoc iated stereotype (and hence that which is thought to be a member of the extension). Putnam is wr i t i ng in contrast to those who would hold that th meaning of a natural kind term, unl ike the meanings of other general nouns is permanently f i x e d . The large puzzle, however, is why TSRM bother to re fer to Putnam's par t i cu la r theory at a l l . The posi t ive use that is made by the ecolog-ica l theor is ts of "natural kind" does not obviously require acceptance of the Putnam account. In f a c t , TSRM's use of "natural kind" may not even be the notion over which philosophers debate. Their use is of "natural kinds fo r animals", explained as fo l lows: Construed as natural kinds for animals, affordances do not require grounding in occurrent properties that sa t is fy the explanatory s t r i c tu res of science but in occurrent properties that sa t i s fy the pragmatic c r i t e r i a of successful a c t i v i t y in a res t r i c ted universe of p o s s i b i l i t i e s , v i z . , an ecological world.22 "Occurrent proper t ies" , as TSRM mean th is could be taken as the same as what Putnam means by "underlying mechanism" and therefore one can speculate that the authors have simply h i t on th is par t i cu la r feature of Putnam's view. But what is to be made of the reference to "pragmatic c r i t e r i a " ? The point TSRM are attempting to make turns out to be consid-erably simpler than the obscure phrase would suggest. In the phrase "natural kind for an animal", the word "natura l " is given a rather d i f f e r e n t emphasis than is customary in ta l k about natural kinds. The affordances environments hold for perceiving organisms are grounded in the manifest structures of the fu rn i tu re of the environment, as noted previously. Anything at a l l meeting certa in physical spec i f i c -ations could af ford certain kinds of behavior. The general class of things which could af ford some par t i cu la r a c t i v i t y , though, w i l l almost invar iab ly be larger than the subclass of those things that natura l ly or t y p i c a l l y occur in an organism's local environment ( "habi tat " is a word that corresponds most nearly to TSRM's use of "environment"). So, for example, anything at a l l of a certain minimum height, r i g i d i t y , of a certain textual q u a l i t y , and so could be climbed by a marsh periwinkle as long as i t sa t i s f i es some general, mostly minimum requirements. This general class of things the snai l could climb may well include objects such as penci ls , p las t i c plant stems or swizzel s t i cks . As TSRM use the phrase, none of these are natural kinds for the per iwinkle. Swizzel s t icks and the l i ke are not members of "climb-upable th ing" because they are non-typical features of the habitat (the upper i n t e r t i d a l zone) in which the species resides and in which ( th is is an important point to the ocological theor is ts ) i t evolved. The idea TSRM therefore want is not so much that of a natural kind as that of a res t r i c ted kind: "cl imb-upable t h i n g " , as a perceptible category fo r the per iwink le, consists only of the usual consti tuents of i t s environment (a subclass of plant stems) that a f ford climbing fo r members of that species. To c l a r i f y TSRM's conception, the per iwinkle 's environment is upper i n t e r t i d a l zones containing certain kinds of vegetation. This local environment and i t s natural contents const i tutes the " res t r i c ted universe of p o s s i b i l i t i e s " in which the species and i t s perceptual systems evolved. The so-cal led natural kinds f o r the marsh per iwink le, or for any other organism, are defined by TSRM in terms of the things in i t s environment that a f ford a cer ta in kind of behavior. In the present case, "climb-up-able th ing" has as i t s extension plant-stems f a l l i n g w i th in some set of physical dimensions represented in TSRM's discussion by the set s , t , u , v . These const i tu te the occurrent proper t ies, TSRM's corre late of Putnam's "underlying mechanisms", which ground or explain the a b i l i t y of things with the physical character is t ics of the periwinkle to climb a subclass of plant-stems. I f an affordance, which is a natural kind fo r an animal, is r e s t r i c -ted to items in an organism's environment that allows a par t i cu la r kind of a c t i v i t y , then i t is w i th in th is context that some invar iant property must be spec i f ic to environmental features that a f ford an a c t i v i t y . The idea would then be that i f periwinkles do perceive climb-upable th ings, there is some opt ica l invar iant in the per iwinkle 's habitat associated wi th a l l and only things having s , t , u , v (plant stems of cer ta in minimum dimensions). These dimensions are the ones in v i r tue of which plant stems are climb-upable. The per iwinkle 's responding c l imbingly , as i t were, only to the detection t>f a p a r t i c u l a r opt ical invar iant works in i t s local environment in the sense ttvat a successful bout of climbing i s , other things being equal, the inev i tab le resu l t . Outside the res t r i c ted universe, matters may be quite d i f f e ren t . This includes the removal o f the organism from i t s habi tat and also the in t roduct ion of some non-typical element in i t s habi ta t . In e i ther case, the periwinkle could be expected to pass up things i t could in fact climb and t r y out many "non-climbables". The l a t t e r would occur i f there were phenomena outside the habitat which happens to produce the same opt ical invar iant associated with climb-upable plant-stems. The former holds because the opt ica l invar iant to which periwinkles respond is not necess-a r i l y produced by anything the organism could climb. In pa r t i cu la r , i t would not be produced by things which the periwinkle could as a matter of fact c l imb, but does not possess the structure s , t , u , v . Hamilton's study would seem to ind icate , fo r example, that (if_ the per iwinkle 's perception is in fact invariant-based) the relevant opt ica l s t ructure is not produced by clear plexiglass rods. To summarize, the extension of a natural kind fo r an animal consists of the natura l ly occurring features of the environment in which i t s species evolved and l ives which af ford some kind of a c t i v i t y . Two general points need to be made about t h i s . F i r s t , TSRM's account of affordances has the advantage of reducing the size of the classes of things to which invar iants must be spec i f ic from anything that would af ford a cer ta in a c t i v i t y fo r an animal to the usual consti tuents of i t s environment which do so. Thus narrowing the f i e l d , so to speak, makes i t a l l the more easy to envisage a par t i cu la r opt ica l invar iant being uniquely associated with the class of phenomena an organism perceives as af fording some kind of behavior. At the same t ime, i t is worth observing that the marsh p e r i -wink le , because i t s habitat is qui te r e s t r i c t e d , provides a misleadingly simple case. As one moves on to consider more adaptable and mobile organ-isms (humans, fo r instance), i t w i l l become increasingly d i f f i c u l t to c lear ly i den t i f y t he i r pa r t i cu la r " res t r i c ted universe". I t w i l l conseq-uently become ever more d i f f i c u l t to iden t i f y perceptual categories and t h e i r extensions for those organisms. The second general point is to emphasize the fact that TSRM's sole c r i t e r i a fo r determining the content of the per iw ink le 's , and other organisms', perception is i t s behavior. The basis fo r the claim that periwinkles perceive certa in plant stems as climb-upable things is that the snai ls in fac t respond cl imbingly in cer ta in circumstances to those stems. The step from using behavior as the c r i t e r i a for determining what organisms perceive things as to analyzing what they perceive things as in terms of behavior (or behavior-states) is a f a i r l y short one. In th is chapter, I have indicated that TSRM's discussion is bedevil led by some confusions about cer ta in phi losophical ly common notions. They allow themselves to make mistakes about extensional i ty and in tens iona l i t y and about the re la t ion between properties and intensions. These lead to confusions about commitments the Establishment must make, TSRM also have a very queer understanding of "natural kind" which leaves i t unclear as to why they are so favourably disposed toward the views of Putnam. Their concept of natural kind turns out to be a concept of a " res t r i c ted k ind" . The kinship with Putnam seems to consist only in the idea of underlying mechanisms that "ground" cer ta in propert ies. The next chapter w i l l continue the c r i t i c a l discussion of TSRM's a r t i c l e . This continuation w i l l f i n a l l y focus on TSRM's attempt to show how Gibson's theory can account for in ten-t i o n a l i t y , which w i l l largely involve TSRM's adoption of Dretske's theory of natural laws. Chapter Five The idea of natural kinds for animals and a par t i cu la r view about natural laws is the basis of TSRM's descr ipt ion of the Gibsonian theory of spec i f i ca t ion . TSRM attempt to use these notions answer the charge that the ecological approach has no account of i n t e n t i o n a l i t y in perceiving. The basic idea is that the ecological approach somehow manages to account for i n t e n t i o n a l i t y in perceiving because there are lawful connections between ecological propert ies (natural kinds fo r animals or any other environmental features the ecological theor is ts take to be percept ible) and occurrent physical propert ies of the environment, and between invar iants and the manifest environmental ones. According to TSRM, f a i l i n g to recognize the point that there ex is t lawful connections of cer ta in kinds is the crucia l mistake Fodor and Pyly-shyn make in t h e i r c r i t i c i sms . They claim that Fodor and Pylyshyn repea-tedly a t t r i b u t e a "weak corre la t ional view" to Gibson, meaning, one would suppose, a view on which re lat ions between ecological propert ies and invar iants are correlated merely in some s t a t i s t i c a l sense.* However, i t is not clear that Fodor and Pylyshyn do_ a t t r i bu te th is weak view to Gibson or that i t would a f fec t t h e i r central objections in any case. That is to say, i t w i l l prove unclear as to how, through the postulat ion of lawful re la t ions between occurrent "underlying mechanisms", ecological proper t ies, and inva r ian ts , TSRM can show how the theory of spec i f i ca t ion has an account of i n t e n t i o n a l i t y . Now a central feature of TSRM's supposed a l te rnat ive account of in ten-t i o n a l i t y in perception is t h e i r adoption of Dretske's intensional theory 2 of natural laws. The most reasonable way to proceed with th is discussion there fore , is as fo l lows. F i r s t , TSRM's version of the theory of specif-i ca t ion should be explained. They c lear ly indicate that the notion of a natural law on. a more conventional understanding of laws, is of no help. I w i l l then examine TSRM's claims in l i g h t of the Dretskean theory to see why TSRM might th ink they derive some advantage from embracing th is theory in pa r t i cu la r . ( i ) TSRM's version of the theory of spec i f ica t ion consists of four large claims. The f i r s t two of these were elaborated in the previous chapter but i t is usefu l , fo r c l a r i t y , to l i s t a l l four of them here. Given some occurrent, physical s t ruc tu re , P, and opt ical invar ian t , 0, both of which are associated with the ecological property, A: 1. In the context of an organism's environment, P is a necessary condit ion fo r A 2. A must be grounded in P, in the sense tha t , in the organism Js environment, A's presence is explained by the presence of P 3. Among the typ ica l consti tuents of an organism's environment, things having 0 are uniquely correlated with things that have A 4. The environmental constituents having A and those having 0 are uniquely correlated in an organism's environment because the same physical structure which explains the presence of A also explains the presence of 0 1) and 2) j u s t serve to indicate that A const i tutes a natural kind fo r some organism, and that the re la t ion between P and A is non-accidental 3) states t h a t , under cer ta in r e s t r i c t i n g condi t ions, 0 is spec i f ic to things having A. The main r e s t r i c t i n g condi t ion, mentioned in chapter four , is that the re la t ion only needs to hold between occurrences of 0 and the presence of A in the organism's natural habi ta t . Outside th is envir-onment, things may have A but f a i l to exh ib i t 0 because the property A can be produced by occurrent propert ies other than P (other than the only occurrent s t ructure in the organism's environment which in fac t produces i t ) . S im i la r l y , things may produce 0 but f a i l to have A because, outside the organism's environment, 0 may be produced by occurrent physical structures other than P. Another l i m i t a t i o n TSRM mention is that 0 only needs to be spec i f ic to A from the "perspectives" which an organism normally takes on i t s environment. By "perspectives" the authors mean the posit ions from which, and the conditions under which members of a given species natura l ly observe features of t h e i r environments. Hawks and f i e l d mice may be found largely in the same kinds of regions, but the perspectives they would have on t h e i r t e r r e s t r i a l environment would be quite d i f f e r e n t . The claim must be that the invar iants avai lable to each species would s im i la r l y d i f f e r . 4) states that the manifest property that grounds A is the very one that produces 0. Since the manifest property, P, produces 0, the connec-t ion between the two is non-accidental. S im i la r l y , since P produces A, the connection between these two properties is non-accidental. TSRM use these points that they make in order to claim that the re la t i on between A and 0 is also non-accidental . The argument used fo r th is purpose w i l l be examined shor t l y . Assertion 4) places an obvious r e s t r i c t i o n on the ecological approach' conception of which par t i cu la r invar iants could specify which features of 91 the environment. The r e s t r i c t i o n is that the very same manifest physical property responsible for the presence of some ecological property must also be the very one which produces the invar iant which specif ies the ecological property in question. V isua l , spa t ia l l y -o r ien ted examples seem to lend themselves readi ly to th is idea: A s i t -onab le , climb-upable, or cut-withable thing for a human is so because of the size and angles between the surfaces (as well as the r i g i d i t y and perhaps other th ings ) , f o r instance. I f there is an invar iant spec i f ic to any df these a f fo rd -ances, i t is l i k e l y to be s im i l a r l y due to such physical parameters as size and angular arrangements. The ecological approach is supposed to be generalizable to other sorts of ecological properties and other percep-tual modali t ies ( including those which are not analyzable in terms of surfaces), though. These other instances may not prove to be so promising. There is no par t i cu la r reason, fo r example, to expect that the very same underlying properties which make certain things edible are the very ones which could produce the invar iant that would specify i t as such. To continue with the main top ic , TSRM are very clear in i ns i s t i ng that only an intensional view of laws w i l l help the i r case. They are quite committed to the view that the ecological approach gains no advantage by assert ing that spec i f ica t ion of ecological properties is lawfu l , given some extensional conception of laws. One can make a reasonable guess as to why the authors believe th is based on what they have to say about Establishment views. They claim that such t r a d i t i o n a l views as Fodor and Pylyshyn defend can only explain an organism's a b i l i t y to d i f f e r e n t i a t e between two perceptible proper t ies , A and B, that happen to be coextensive by appealing 4 to concept acquis i t ion and possession and the cognit ive processing. Where such appeals can not work, TSRM claim, the Establishment must simply concede that the perceiver in question does not d i f f e ren t i a te between A and B, even when the organism's behavior indicates that i t can. The problem is alleged to ar ise because of a committment to an extensional or empir-i c i s t , theory of natural laws on the part of the Establishment. TSRM (mistakenly) suppose that an extensional is t view takes statements' status as laws survive subst i tu t ion of coextensive terms. On such a view, any laws in which A par t ic ipates are ones in which B equally par t ic ipa tes . The authors would then go on to reason that i f some law explains an organ-ism's perception of A, i t also explains i t s perception of B. The lesson which TSRM wish to draw from th i s problem they envisage for the Establishment is that a bet ter explanation of how organisms d i f f e r -ent ia te between coextensive properties must be devised. The prominence of the notion of natural laws in t h e i r discussion suggests very c lear ly that TSRM th ink the postulat ion of cer ta in kinds of lawful connections w i l l y i e l d the explanation they want. In general, t h e i r account seems to come to the fo l lowing: even though propert ies A and B are coextensive, A is lawfu l ly re lated to a d i f f e ren t invar iant than is B. The organism some-times behaves in ways appropriate to the perception of something as A because i t detects the invar iant which is lawfu l ly related to A. Because A and B are coextensive, B is always and only present when A's invar iant is present, but B's re la t ion to that invar iant is "merely co r re la t i ona l " . The organism responding as though i t perceives something as A when i t s perceptual systems detect a par t i cu la r invar iant is supposed to be explained in terms of the fact that the re la t ion between invar iants of that sor t and the ecological property A is non-accidental, while i t s re la t ion to B is accidental . As an example, one might consider again TSRM's case of the marsh per iwink le , amending the i r descr ipt ion so that "climb-upable th ing" and "bar r ie r " are genuinely coextensive in the sna i l ' s habitat ( that i s , other snai ls and plant-stems are barr iers i f and only i f they are climb-upable). TSRM imagine that certain manifest proper t ies, s , t , u , v , are the ones which make something climb-upable. I f th is is a perceptible property, s , t , u , v (or perhaps some subset of them) must be the properties which produce the invar iant the periwinkle detects whenever i t perceives something as climb-upable. TSRM are assuming that some other manifest proper t ies , p ,q , r , are the ones which make something a bar r ie r fo r the per iwinkle. Assuming that i t is not plausible to a t t r i bu te concepts, cognit ive processes, or the l i k e to sna i l s , TSRM claim that the Establishment has no explanation of how periwinkles d i f f e ren t i a te between "climb-upable" and "co l l i de -w i thab le " . I t is also assumed, of course, that the creature's behavior is s u f f i c i e n t evidence that they do so d i f f e r e n t i a t e . The Establishment, so the argument goes, does not have cognit ive processing avai lable to explain the d i f f e ren t ia t i ons made by the per iwinkle. Since "climb-upable th ing" and "bar r ie r " are coextensive, they are equally correlated with the invar iant which s , t , u , v produces. Therefore, neither can the Establishment claim that the per iwinkle 's perceptual systems have picked up on some association between a property of the opt ic array and "climb-upable" that does not ex is t as well between that opt ical property and bar r ie rs . TSRM want to conclude that t r a d i t i o n a l theor is ts such as Fodor and Pylyshyn must therefore disregard the per iwinkle 's evident a b i l i t y to d i f f e r e n t i a t e between barr iers and climb-upables. TSRM claim to solve such puzzles in terms of the lawful connections they postulate: c l imb-upab i l i t y , grounded as i t is in the properties s , t , u , v is lawfu l ly related to the invar iant produced by s , t , u , v (ca l l t h i s invar iant 0 ( c ) ) . C o l l i d e - w i t h a b i l i t y , on the other hand, is only accidental ly related to s , t , u , v since i t is grounded in other manifest propert ies (namely p , q , r ) . Thus the invar iant s , t , u , v produces lawful ly speci f ies "climb-upable t h i n g " , and does not lawfu l ly specify " c o l l i d e -withable th ing" ( b a r r i e r ) . Therefore, TSRM would conclude, the periwinkle perceives something as climb-upable, and not as a b a r r i e r , whenever i t detects 0 ( c ) . Two points should be made about th is suggested explanation of how organisms d i f f e ren t i a te between instances of ecological propert ies. F i r s t , in defence of the Establishment, i t is not at a l l clear why they should be caught in the d i f f i c u l t i e s that TSRM seem to be suggesting. One should th ink the problem of explaining differences in the per iwinkle 's behavior toward the very same class of physical things (plant stems) could be handled in terms of a notion such as readiness or expectancy sets: given one set of background conditions (receding t i d e , for example) which the periwinkle can sense, the detection of a given invar iant const i tutes the perception of something as a ba r r i e r . That i s , the periwinkle moves around the objects perceived. Given a d i f f e ren t set of background conditions that the periwinkle can sense (advancing t i d e ) , detection of the very same invar iant const i tutes the perception of something as climb-upable. Per-ceived background conditions prime the per iwink le, as i t were; they cause i t to be in an expectancy set such that the snai l responds "c l imbingly" to the detection of the invar ian t . Second, i t is not at a l l clear how, on any conventional understanding of natural laws, they w i l l do what TSRM require of them. That 0 (c) law-f u l l y speci f ies climb-upable things but is merely invar iably correlated wi th perceived barr iers does represent a di f ference between the two eco-logical proper t ies , but the key question is whether or not th is is a d i f -ference of which perceptual systems of organisms could take advantage. The two ecological properties are correlated to precisely the same degree in the per iwinkle 's environment. In order for the dif ference to be usable, the visual system of the organism must somehow be sensi t ive to the re la t ion between 0 (c) and c l imb-upab i l i t y . However, the visual system does not detect the re la t ion in question: i t detects the opt ica l invar ian t , 0 ( c ) , i t s e l f . Given t h i s , and the fact that the presence of 0 (c) is equally and invar iab ly correlated with the two ecological proper t ies, i t should be a mystery as to why a species that has evolved in th i s environ-ment should have evolved so as to perceive things as climb-upable (but not as co l l ide-wi thab le) whenever i t detects the invar iant opt ica l s t ruc ture , 0 ( c ) . In other words, a general appeal to natural laws does not explain how organisms manage to perceptually d i f f e ren t i a te between coextensive ecological propert ies. For th i s reason, and also because TSRM take the acceptance of a par-t i c u l a r theory about natural laws to be of fundamental importance, i t is reasonable to assume that TSRM's account of i n t e n t i o n a l i t y in terms of the ecological theory of spec i f i ca t ion could not possibly work without Dretske's intensional view of natural laws. Since acceptance of Dretske's par t i cu la r theory is to be regarded as essential to TSRM's case, therefore any one of the fo l lowing three points would show that t h e i r account of in ten-t i o n a l i t y does not work: 1. Dretske's theory is mistaken 2. The intensional view of laws affords TSRM no advantage which would not have been avai lable to them on a conventional, empi r ic is t conception of laws 3. TSRM cannot consistent ly accept the Dretskean theory because of other views the authors hold. My discussion of the Dretskean theory in re la t ion to what TSRM say about i t w i l l reveal reasons for both 2) and 3) . ( i i ) There are a number of confusions which TSRM appear to make about Dretske's view that a natural law is a re la t ion between intensions. The largest one might be the idea that embracing th is intensional theory of laws w i l l somehow grant t h e i r theory of spec i f icat ion the a b i l i t y to re la te perceivers to (some conception of) intensions. In any case, a thorough discussion of the mistakes about Dretske's account should show that i t is doubtful TSRM have any proper understanding of Dretske's theory at a l 1 . Dretske's account is wr i t ten against what he ca l ls "empi r ic is t " con-ceptions of natural laws. The empi r ic is t view is taken to be basical ly the view that what is expressed by a law-statement is the very same kind of thing which is expressed by any ( t rue) universal ly general statement of f a c t , plus an add i t i ona l , special ingredient: This response to the alleged uniqueness of natural laws is more or less standard fare among empir ic is ts in the Humean t rad -i t i o n . Longstanding (= venerable) epistemological and ontological commitments motivate the equation: law= universal t ru th + X. There is disagreement among authors about the d i f f e r e n t i a X, but there is near unanimity about the fact that laws are a species of universal t r u t h . 5 Dretske says that that which is expressed by some universal cond i t iona l , "A l l F's are G", is a re la t ion between extensions: every member of the extension of F is also a member of the extension of G. He takes the empir-i c i s t models of laws to s im i la r l y consist of t h i s extensional sort of r e l a t i o n . The puzzle, then, is to f ind the special ingredient X which explains the necessity of the r e l a t i o n . One must f i nd whatever i t is that confers on laws the pecul iar functions and feature that are not possessed by j u s t any universal general izat ion. Dretske's a l t e rna t i ve , which he neatly encapsulates by saying a law is a singular re la t ion between intensions, is introduced thus: To say that i t is a law that F's are G is to say that "A l l F's are G" is to be understood ( in so fa r as i t expresses a law), not as a statement about the extension of the predicates "F" and "G" but as a singular statement describing a re la t ionsh ip between the universal propert ies F-ness and G-ness. In other words, ( C ) is to be understood as having the form: ( 6 ) F-ness—> G-ness.6 ( " (C ) " , in t h i s passage, denotes " I t is a law that F's are G".) There are two points which need to be made about th is quote. F i r s t , a footnote is attached to i t in which Dretske indicates that the arrow used in ( 6 ) is to be taken as a "dummy connective", meaning that he does not intend i t to indicate a given re la t ion (material or causal imp l i ca t ion , for example). Second, the theory Dretske is proposing is a condit ional one. If_ there are any natural laws, then they are singular re la t ions between universal propert ies. The idea is that the acceptance of the existence of natural laws enta i ls a form of Platonism.^ The f i r s t point relates to the argument by which TSRM attempt to show that the re la t ion between an ecological property and i t s associated invar iant is non-accidental ( l a w f u l ) . Using Dretske's nota t ion , TSRM t r y to argue by an appeal to t r a n s i t i v i t y : Thus we have two laws re la t ing propert ies: "o-ness—> c-ness" (between occurrent property and affordance) and "e -ness -> o-ness" (between opt ical property and occurrent environmental property) . By t r a n s i t i v i t y we have: " e - n e s s - > c-ness". That i s , theregis lawful spec i f i ca t ion of an affordance by an opt ica l property. q Since Dretske "attaches no s igni f icance" to the arrow, i t is a l i t t l e premature fo r TSRM to endow i t wi th spec i f i c logical propert ies without saying what re la t ion they s p e c i f i c a l l y have in mind. They need to defend a par t i cu la r view of the re la t ion involved in laws on which i t is t r a n s i t i v e . TSRM also need to give some explanation of why the o ' s , c ' s , and e's in t h e i r argument occur in j u s t the order they do. I f the order matters ( that i s , the re la t ion is not symmetrical), then TSRM need to give an account of lawful re la t ions which make sense, for example, of why "o -ness—> c-ness" and " e - n e s s o - n e s s " , rather than "c-ness - > o-ness", or "o-ness—> e-ness" That i s , i t seems a l i t t l e odd that the occurrent physical st ructure (which produces both the ecological property and the invar iant by which i t is per-ceived) should be the f i r s t term in the one re la t ion and the second term in the other. I f the order in which the o ' s , c's and e's occur does not matter then TSRM need to provide some conception of lawful re lat ions on which they are symmetrical.*^ Unti l these conditions have been met TSRM's attempt e i ther to derive or to explain lawful re la t ions between ecological propert ies and invar iants in terms of t h e i r " t r a n s i t i v i t y argument" is qui te without substance. The point that the be l i e f in natural laws, given Dretske's view, requires one to postulate abstract , universal proper t ies, is a feature of the intensional theory of laws that TSRM miss e n t i r e l y . They w r i t e : Let us ecological r e a l i s t s put our major ontological cards on the tab le : ( i ) there are no bare par t icu lars ( ind iv iduals) and there are no pure forms. The nominalist claim that universals are co l lect ions of indiv iduals is denied as is the Platonist claim that indiv iduals in themselves are clusters of universals (Bunge, 1977). There are no universals in themselves but there are propert ies that are invar iant across a given co l lec t ion of evolving indiv iduals . . . ( v i i ) Properties are not a separate category of i n d i v i d u a l , f o r there are only propert ied things (Bunge, 1977). ^ I w i l l not speculate on whether or not there is a viable posi t ion in what TSRM say. The important point here is t h e i r re jec t ion of the idea of Platonism about propert ies. They are f i rm ly and c lear ly committed to the existence of natural laws, and also to Dretske's view of them. But according to Dretske, his intensional model requires, i f there are any natural laws, the sort of Platonism that TSRM re jec t . A law cannot be a singular re la t ion between certain kinds of existents i f there are no existents of that k ind. Something has to be sacr i f i ced here. Since TSRM's en t i re defence of the Gibsonian approach is grounded in the notion that there are laws connecting organisms' perceptions and the i r environment, TSRM can hardly be expected to abandon t h e i r commitment to ecological laws. The authors must therefore e i ther give up the advantage, whatever i t i s , they are supposed to derive from the Dretskean model, or they must take up a meta-physical view about propert ies on which there w i l l be re la ta fo r t h e i r r e l a t i o n . As i t spec i f i ca l l y a f fects the ecological approach, i t does not seem to matter p a r t i c u l a r l y which a l te rnat ive one adopts. On the one hand, the issue of whether there are universals is qui te i r re levan t to the acceptance of Gibson's, as opposed to any other, approach to per-ception. On the other hand, there are only three imaginable reasons sug-gested by TSRM's a r t i c l e as to why they should think they require the intensional theory of laws. None of these prove to be genuine grounds fo r adoption of the theory by the ecological theor i s t s . These poss ib i l -i t i e s are as fo l lows: A) TSRM mistakenly believe a l ternat ives to the Dretskean view allow subst i tu t ion of coextensive predicates in law-statements. In re la t ion to the periwinkle discussion, TSRM claim: The Establishment/extensional analysis goes through on the fo l lowing assumptions: ( i ) that the general izat ion of law or o f fac t i s in the form of a syntact ic universal . . . ; ( i i ) the s u b s t i t u t i v i t y of coextensive predicates . . . Assumptions ( i ) and ( i i ) fo l low from the t rad i t i ona l conception of law. They are both rejected in the view of law advanced by Dretske (1977), a view which sustains the ecological / in tensional analysis.12 The "eco log ica l / in tens iona l " analysis referred to here is the asser-t ion that the per iwink le 's d i f f e r e n t i a l behavior toward plant stems is to be explained in terms of the detection of d i f f e ren t invar iants which lawfu l l y speci fy , respect ive ly , "climb-upable th ing" and "ba r r i e r " . The imp l ica t ion , in the foregoing passage is that the empir ic is t theories of laws (upon which, TSRM would c laim, Establishment accounts of perception must re ly ) hold that subst i tu t ion of coextensive predicates in law-statements w i l l y i e l d another statement which expresses a law. This "p r inc ip le of s u b s t i t u t i v i t y " is not part of t r a d i t i o n a l theories of laws and Dretske in no way suggests (as the TSRM quote implies) that i t i s . The authors appear to have mistaken something which Dretske regards as a puzzle fo r something he is claiming to be a tenet of other theor ies. Dretske introduces his descr ipt ion of empi r ic is t accounts and his own view of natural laws by making the point that subs t i tu t ion of a coex-tensive predicate in to a statement of law does not always produce another law-statement. He puts th i s more simply by noting that subs t i tu t ion of coextensive terms wi th in the scope of the functor " i t is a law that . . . " 13 is non-truth-preserving. This kind of opacity is one important way of showing the di f ference between statements which express laws and those which merely express universal general izat ions. The purpose is not to suggest that the empir ic is ts are incognizant of dif ferences between laws and simple universal general izat ions, but that t h e i r approach to exp-la in ing the differences is wrong from Dretske's point of view. No-one, however, whether they are empi r ic is t or not , imagines that the subst i tu -t i on of any coextensive predicate into a statement of law w i l l always resu l t in another statement o f Taw. TSRM's adoption of the intensional model of laws might be explained by th i s mistake of a t t r i b u t i n g a blanket p r inc ip le of s u b s t i t u t i v i t y to Dretske's r i v a l s . TSRM want cer ta in l i n g u i s t i c contexts that are relevant to perceptual theory to be opaque. Any theory which interpreted " i t is a law that . . . " as r e f e r e n t i a l l y transparent would allow subst i tu t ion in the contexts TSRM take to be opaque. Since they think empi r ic is t theories about natural laws are offenders in th i s respect, TSRM perhaps think they must adopt Dretske's theory in par t i cu la r . However, an empir ic is t theory of laws does not contain the p r inc ip le of s u b s t i t u t i v i t y TSRM have in mind, but would regard f a i l u r e of subst i tu t ion as a puzzle, as Dretske does. TSRM thus gain no advantage here by taking up the intensional model. B) Another possible source of TSRM's conclusion about laws is not based on anything they d i r e c t l y say, but i s instead the product of a body of circumstant ial evidence. F i r s t , Dretske l i kes to describe laws as singular re la t ions between intensions, where an intension is some kind of abstract , universal prop-e r t y . Hence, Dretske sometimes uses "property" ( in the abstract sense) in place of " in tens ion" . Second, TSRM both endorse Dretske's notion of a law, and s i m i l a r l y take up the practice of using "property" and " intension" interchangeably. TSRM e x p l i c i t l y re jec t the idea of abstract , universal 14 proper t ies , however. Th i rd , the important c r i t i c i s m Fodor and Pylyshyn make against Gibson may be thought of as the claim that perception as mere invar iant -detect ion lacks an a l ternat ive the Establishment's re la t ion between a perceiver and an intension ( in the sense of some sort of i n te r -im nal representat ion). F ina l l y , as noted e a r l i e r , TSRM, wi th e x p l i c i t support from Gibson, tend to think of meanings or values as being (simply) perceived. Gibson asserts that affordances are d i r e c t l y perceived, and that to perceive what things a f ford is to perceive what they mean. 1 6 Given these points as background, i t is possible that TSRM think that by accepting Dretske's view about laws, they can re late perceivers 102 and intensions (of some kind) without recourse to the a t t r i b u t i o n of mental representations to perceivers. The lawful spec i f ica t ion of some ecological property, A, by some invar ian t , 0 , is (qua lawful re la t ion) a re la t ion between the intensions "A" and " 0 " . So TSRM might then be t ry ing to argue that when an organism's perceptual system detects 0 , i t perceives the " intension" A. In other words, the idea would be to conceive intensions so that they would somehow be in the world as par t i cu la r feat -ures of objects. I f th is happens to be what TSRM have in mind, they gain nothing at a l l from Dretske's view. The foregoing l ine of reasoning rests on the r e l a t i v e l y obvious fa l lacy of confusing "property" in the abstract.sense wi th "property" in the sense of an instance. Dretske's re la t ion is between abstract e n t i t i e s (universals) . When.an organism's perceptual system funct ions, i t detects par t i cu la r instances of invar iants . This does not re late a perceiver to an intension on any theory of natural laws. Adoption of the intensional theory of laws, to put i t b r i e f l y , does not in any way change the conception of what organisms are related to when they perceive. I t should be emphasized, once again, that TSRM may not be committing the suggested fa l l acy . Working out the motivations behind a view is not exactly l i k e showing what i t e n t a i l s : log ic is often not on one's s ide, and th is is especial ly so when the author(s) being studied appears to have been led astray. Therefore, in spi te of the uncanny coherence in a cer ta in body of evidence, i t would indeed be unfa i r to simply conclude that TSRM are g u i l t y o f the suggested confusion. One can only hope that they are not and note that i f t he i r reason fo r adopting Dretske's model of laws is an attempt to re late perceiving organisms with " intensions" TSRM are quite misguided. Any advantage the ecological approach might therefore hope to gain from the intensional construal of natural laws does not l i e here. C) F ina l l y , TSRM might be adopting the Dretskean view because they think i t allows them a notion of laws which are res t r i c ted in scope. They evident ly want such a notion in order to have natural laws that hold only in cer ta in environments. Once again, they take Dretske to be expounding the point they wish to make: The universal scope of laws of nature should not be taken to mean that the same laws apply everywhere and everywhen, fo r laws can apply only where they are ins tan t ia ted . The laws governing electron o rb i ts are un iversa l , but no one expects them to operate in the solar nacleus, where atoms are deprived of t h e i r electron shel ls by the intense play of other forces. Following Dretske, we take laws to be par t i cu la r statements about properties that are more or less widely d is t r ibu ted in space- t ime.^ There is a great deal to puzzle about in th is quotat ion, which makes i t a f a i r sample of TSRM's very b r i e f and very d i f f i c u l t discussion on the scope of natural laws. The authors want a re la t ion between invar iants and ecological properties such t h a t , in a res t r i c ted natural environment, an invar iant and an ecological property are the resu l t of the same under-l y i n g , occurrent property, and the invar iant is found in the avai lable opt ical (or other) st ructure always and only when the ecological property is present. Why TSRM should think they need Dretske's theory of laws to postulate th is re la t ion is mysterious. A couple of points about the foregoing quote should be made. F i r s t , the statement "laws can only apply where they are ins tant ia ted" could only have one of two in terpreta t ions here. I t could be meant as "laws can only apply where they apply". A l te rna t i ve l y , i t could mean that a law can apply only where the properties which are the subjects of i t are ins tan t ia ted . The former in te rpre ta t ion is t r i v i a l while the l a t t e r is p la in ly fa lse and does not fo l low from anything Dretske says. 104 The example of the electron o r b i t s , as well as the comment that laws are about propert ies " that are more or less widely d is t r ibu ted in space-time" suggest that TSRM are taking the fa lse in te rp re ta t ion . They seem to claim that laws whose subjects are atoms with electron shel ls do not hold where there are no atoms with electron she l ls . However, one of the main problems in analyzing laws, and one of the reasons Dretske gives the f o r -mulation he does is precisely that i f "F-ness — ^ G-ness" is law, i t 18 obtains even when and where there are no F's. A c losely related point pertains to a logical mistake that TSRM make in the foregoing passage. I t cannot be the case both that laws about electron o rb i t s are universal and that they f a i l to obtain at the solar nucleus. I f the l a t t e r were t rue , the laws would f a i l to hold un iversa l ly . On the other hand, i f the laws apply un iversa l ly , then the laws apply to the solar nucleus as well as anywhere else. I f they are un iversa l , the laws apply counter factual ly : i f atoms could have electron shel ls at the solar nucleus, the electrons would conform to the relevant laws concerning the i r o r b i t s . I t j us t happens that there can be no atoms in the nucleus that possess the i r electron she l l s , which is only to say that in cer ta in regions, there are no subjects of laws about electron o r b i t s . This f i t s with Dretske's conception of laws j us t as much as empir ic is t ones. There are also two mistakes evident in TSRM's discussion with re -spect to t he i r view of Dretske's theory. F i r s t , they claim tha t , on the intensional theory, laws are "par t i cu la r statements", which is fa lse because they are not conceived of as any kind of statement at a l l . This is an instance of a confusion against which Dretske s p e c i f i c a l l y warns, between statements of law and the laws expressed by them. Unfortunately, Dretske cautions at the same time that he sometimes speaks " i n d i f f e r e n t l y " 19 between the two. The other mistake TSRM make here again has to do with the conception of a property. "Proper t ies" , in the sense in which these are the subjects of s ingular re la t ions Dretske conceives, are not "more or less widely d is t r ibu ted in space-time". In f a c t , they are not d is t r ibu ted in space-time at a l l . To r e i t e r a t e , Dretske is using "property" in the sense of universal propert ies. The only sense TSRM have avai lable is that of prop-er t ies as pa r t i cu la rs . Part of the d i f f i c u l t y in TSRM's discussion of the scope of laws is that they do not appear, once again, to understand the posi t ion against which they are (through Dretske) arguing. They suppose that t r ad i t i ona l accounts of laws are based on the assumption " that laws must be expressed as universal ly quant i f ied statements about extensions" and that th is 20 " implies to many that the scope of any law is un iversa l " . Of course no-one thinks that laws must be expressed as universal condit ionals (un i -versal ly quant i f ied sentences) although there are undoubtedly those who believe they must be expressible by some universal condi t iona l . Dretske's own posi t ion is that " law- l i ke statements are singular statements of fac t describing a re la t ionship between properties or magnit-21 udes", hastening to add in a footnote: I am not denying that we can, and do, express laws as simply "A l l F's are G" (sometimes th i s is the only convenient way to express them). Al l I am suggesting is that when law- l ike s tate-ments are presented in th i s form i t may not be clear what is being asserted: a law or a universal g e n e r a l i z a t i o n . " TSRM c lear ly cannot be get t ing at the heart of the matter by complain-ing about views which allow law-statements to be "universal ly quant i f ied statements". TSRM c i te "A r i g i d object with a sharp dihedral angle, an edge, affords c u t t i n g , i t i s a kni fe" as a typ ica l example of an ecolog-ical law. They go on to note that i t is not the kind of example of a law that philosophers would give because " i t does not f i t the schema (x)(Fx — > Gx)". There is absolutely nothing for TSRM to gain by making th is claim. According to Dretske, the form a statement has or lacks does not determine whether or not i t can be a law-statement. This is j us t as well fo r TSRM because, unless t h e i r example of an ecological law is meant to re fer to a pa r t i cu la r r i g i d object , i t i_s_ a universal general izat ion and does f i t the schema ( x ) ( F x — ^ G x ) . Again TSRM appear to be confused over the dif ference between making claims about cer ta in kinds of statements and claims about what these express. Dretske's denial of "un iversa l i t y " of laws comes to the denial that fo r a statement to express a law is fo r i t to express a universal t ru th ( In th is context, "universal, t r u t h " should be taken to mean simply a true universal genera l iza t ion) . This is not to deny that the same form of statement by which universal generalizations are expressed might also express laws. In f a c t , Dretske even holds that a given statement that expresses some law can also express a universal general izat ion. He claims only t h a t , when a statement does express a Taw, th is is not a matter of i t s expressing a special kind of universal general izat ion. This point exhausts Dretske's denial of "un ive rsa l i t y " of laws. The point that TSRM want is that lawful spec i f ica t ion may be r e s t r i c -ted in scope. That is cer ta in invar iant properties are present only when a pa r t i cu la r ecological property is present in spec i f ic natural environ-ments. This is because, in these l im i ted condi t ions, only the underlying manifest property which grounds the ecological one produces the par t i cu la r invar ian t . TSRM's denial o f " u n i v e r s a l i t y " , in other words, comes to the claim that cer ta in unique, non-accidental correspondences hold between invar ian t , eco log ica l , and occurrent properties when the frame of reference is confined to par t i cu la r local environments. TSRM's point is thus that some lawful re la t ions hold only in certain places ( i e . , some non-accidental corre lat ions hold only in cer ta in places) Dretske's claim is that a statement does not express a law in v i r tue of expressing a true universal general izat ion. I t should be readi ly apparent that there is no connection between the two. As such, there is no par t i c -ular reason to think that TSRM must adopt Dretske's, as opposed to the more conventional empir ic is t view of laws in order to make t h e i r point about res t r i c ted scopes. F ina l l y , there is a b i t of irony in TSRM's use of the Dretskean view to develop some concept of non-universal, lawful re la t ions . One of the chief problems Dretske claims against the empi r ic is t formula is that the resu l t cannot be universal enough. Using the analogy of legal imperatives he wr i tes : I f a law was to be interpreted as of the form: "For a l l x, i f x is (was or w i l l be) President of the United States, then x must ( lega l l y ) consult Congress on matter M", i t would be incom-prehensible why Sally B ick le , were she to be President, would have to consult Congress on matter M. For since Sal ly Bickle never was and never w i l l be President, the law, understood as an imperative applying to actual Presidents (past, present, and f u t u r e ) , does not applyT^** Dretske's claim here, in re la t ion to the empi r ic is t model, is that his view allows laws to apply to addi t ional things ( to which they should apply) : not merely actual F's, but things that would be F in d i f fe ren t counterfactual s i t ua t i ons , or things that are F in d i f f e ren t possible worlds. Hence, i f Dretske's assessment is cor rec t , the intensional view of laws makes them, in one respect, more universal than other accounts of natural laws. To summarize, TSRM cannot accept Dretske's theory because they hold that there are natural laws while denying the existence of universals. Their f a i l u r e to appreciate the role of universal properties in the intensional view, as well as some of the other very odd mistakes TSRM make regarding Dretske's theory, suggest very strongly that t h e i r grasp of the theory is tenuous at best. Three possible reasons have been proposed as to why TSRM imagine they need the intensional theory o f laws in pa r t i cu la r , and none of these do in fact give the ecological theor is ts any advantage over an in terpre ta t ion of t h e i r theory of spec i f ica t ion on a conventional view of laws. Dretske does not need to have any d i f fe ren t view regarding the s u b s t i t u t i v i t y of coextensive predicates in law-statements. No-one with a theory in th is area would claim that the subst i tu t ion of any coexistensive predicate in to a statement governed by " I t is a law that . . . " would preserve the t r u t h -value of the statement. Second, adoption of the Dretskean view of laws does not allow organisms to simply perceive intensions nor does i t allow t h e i r perceptual systems to detect them. This would be to confuse abstract propert ies with t h e i r instances. Th i rd , the intensional theory of laws is not required in order to postulate unique, non-accidental correspondences that obtain only in cer ta in animal-environments. 109 Chapter Six The important point which TSRM wish to make to explain how the ecological approach accounts fo r i n t e n t i o n a l i t y in perceptual awareness is that the theory of spec i f ica t ion postulates lawful re la t ions between invar iants and ecological properties (especial ly natural kinds fo r organ-isms). TSRM take i t as central to t h e i r case that one understands the natural laws holding between invar iants and ecological propert ies on the Dretskean, intensional view of laws. The authors regard th is point as key, and i t is anyway qui te unclear how tine postulat ion of natural laws on a more conventional conception of these could help t he i r case. Therefore, l e t i t be assumed that i f TSRM's e f f o r t s to adopt the Dretskean theory f a i l s , so too does t h e i r explanation of how the so-cal led theory of spec i f ica t ion manages to account fo r i n t e n t i o n a l i t y in perception. There are three grounds upon which TSRM's defence in terms of the intensional conception of natural laws could f a i l : 1. Dretske's view is mistaken 2. The intensional view affords TSRM no advantage which would not have been avai lable to them on a conventional, empi r ic is t conception of 1 aws 3. TSRM cannot consistent ly accept the Dretskean theory because of other views the authors hold I have endeavoured to show both 2) and 3 ) . That I do not attempt to show 1) as well is no ind icat ion of acceptance of Dretske's theory of laws, however. There may well be good reasons fo r re jec t ing i t . However, since the advantage TSRM imagine themselves as gaining by endorsing Dretske's theory in par t i cu la r is so mysterious, i t seems u n f r u i t f u l to go on to c r i t i c i z e the theory as we l l . One should note that there are d i f f e r e n t , more or less sophisticated versions of perceptual behaviorism as I intend to use th i s not ion. The least sophisticated would be c lass i ca l , stimulus-response behaviorism: an organism is pkrceptual ly aware of some phenomenon when, and only when, par t i cu la r behavior is e l i c i t e d as a resu l t of the detection of a cer ta in kind of st imulus. More sophisticated variants would not require actual behavior: an organism is perceptually aware of some phenomenon when, and only when, i t comes to be disposed to behave in some way as a resu l t of the detection of some stimulus. The forms of perceptual behaviorism become increasingly sophisticated as one attempts to account adequately fo r d i s -pos i t ions, relevant behavior, and the in teract ion between states of per-ceptual systems and other states of the organism. The resu l t of a l l such considerations is a fami ly of views. Some of these may more properly be cal led versions of funct ional ism. I t is convenient fo r me to re fer to the ent i re family as forms of perceptual behaviorism. What makes these theories a family is that each is an attempt to explain perceptual awareness of organisms with reference to the output behavior produced by the funct ioning of t he i r perceptual systems. There is a great deal of controversy in the philosophy of mind and perception about whether analyses of cogni t ive or perceptual awareness in terms of output behavior (manifest or d ispos i t iona l ) are p lausib le. On the side of those who do not th ink any such theory is plausible is Dretske, 1 a p h i l -osopher with whom TSRM are f a m i l i a r . Treating the Gibsonian theory as a form of perceptual behaviorism I l l allows the approach to avoid a serious i n i t i a l d i f f i c u l t y : the charge . that i t has no account of how the detection of some invar iant manages to mean anything to a perceiving organism. This is managed by giv ing an answer to the question of what i t is for the detection of an invar iant to mean something to an organism (or , a l t e r n a t i v e l y , what i t is for an organism to be perceptually aware of some feature of i t s environment as such). Indeed, one might speculate that the reason certa in philosophical c r i t i c s do not see how Gibson could account fo r i n t e n t i o n a l i t y in knowing perception is that they have a d i f f e ren t answer to the "what"-question, a d i f fe ren t view on what i t is for the deliverances of a perceptual system to mean something to an organism, in mind. Now there is a cer ta in amount of evidence fo r th ink ing that Gibson and his fol lowers might actual ly subscribe to perceptual behaviorism (as opposed, that i s , to holding a theory which happens to be compatible with i t ) . Ecological theor is ts in general are at pains to stress a t i g h t con-nection between perception and a c t i v i t i e s of organisms. Both perceiving and knowing are conceived of more as acts than as re lat ions of an organism 2 to e i ther descript ions or proposit ions. While commenting on affordances on the c l a s s i f i c a t i o n of objects Gibson wr i tes that " I f you know what can be done with a graspable detached object , what i t can be used f o r , you can 3 ca l l i t whatever you please". Evident ly, the possession of some pa r t i c -u lar descr ipt ion by a perceiver is not nearly so important as i t s being in a posi t ion to make some use of some environmental th ing . In TSRM's commentary, the main clues that the authors may have some form of perceptual behaviorism in mind occur in connection with t h e i r 4 view of perceptual er ror and wi th what they ca l l t he i r "semantic theory". They claim to adopt a semantic theory consist ing of three terms: the referent (or extension), the designation and the meaning. For example, i f a marsh periwinkle perceives a climb-upable t h ing , the referent is the plant-stem, the designation is the invar iant which speci f ies the climb-upable things in i t s environment ( I suppose the idea is that the invar iant "designates" the plant-stems which can be cl imbed), and the meaning is the descr ipt ion "climb-upable t h i n g " . The las t component is not avai lable to the s n a i l , but is ascribed to the perceptual s i tua t ion by some observer. From the sna i l ' s perspective, then, a l l that is involved in the s i tua t ion is the detection of some invar iant and subsequent performance of the appropriate behavior. This same idea is found in TSRM's very strange attempt to deal with 5 so-cal led errors in perception in normative terms. A number of examples are given, a l l of which have the same basic structure and moral: Suppose that one reproduces an invar iant by non-natural means so that i t does not have i t s usual physical accompaniment re la t i ve to some species' environment. A member of that species whose perceptual system detects the invar iant w i l l take i t that an instance of the usual environmental accompaniment is present. TSRM want to say the organism is not wrong because the detection of the invar iant ought to have yielded an environmental phenomenon of a cer ta in k ind. One of the examples used is that of a shark which is aware of some-thing as edible whenever i t detects a cer ta in kind of b ioe lec t r i c f i e l d (TSRM refer to i t as a "type F" e l e c t r i c a l f i e l d ) because th is f i e l d , in the shark's hab i ta t , is produced only by the species of f i sh which serve as i t s food. TSRM describe a case in which a type F f i e l d is reproduced by placing electrodes in the sand at the ocean bottom. The resu l t is that the sharks exh ib i t the same predatory behavior toward the electrodes as they would toward the f l a t f i s h they normally eat: 113 The shark digs tenaciously at the source of the f i e l d depart-ing from the s i te only when the act f a i l s to reveal an edible thing (Kalmijn, 1971). Now there is no i n t e l l i g i b l e sense in which i t can be claimed that the source ought to have appeared inedible i f the shark's perception were free from error and i f the shark's perceptions of affordances were d i r e c t . ° The crucia l issue to raise here is that of what TSRM could conceiv-ably mean by something appearing inedible or ed ib le, or what they could mean in predicating (as they do) "takes to be an edible th ing" to the shark. Quite obviously, TSRM's attempt to handle errors in perception does not contact at a l l the epistemological issue of how one can wrongly take one-se l f to perceive some phenomena of one is supposed to d i r e c t l y perceive i t . TSRM's claim that the shark is somehow "not wrong" in taking the electrodes to be edible f l a t l y contradicts the simple epistemological fac t that i f one takes something to be edible that is in fac t ined ib le , one is mistaken. The shark can only be said to be correct in some normative sense when i t takes buried electrodes to be edible. That i s , given the cons t i tu t ion of i t s natural environment, the shark behaved as i t ought to have. TSRM are ins i s t i ng that the shark cor rec t ly perceives something as edible even though nothing in the given example is in fac t edible for sharks. The only way to combine these two points is i f one assumes that the shark's perception of something as edible is j u s t a matter of i t exh ib i t ing pred-atory behavior. This is to analyze what i t is fo r the detection of a type F e lec t r i ca l f i e l d to mean "edible" fo r a shark in behavioral terms. The shark exh ib i ts predatory behavior as a resu l t of the detection of an i n v a r i -ant which is speci f ic to edible things wi th in i t s environment. There are two general comments I wish to go on to make in concluding t h i s thes is . The f i r s t w i l l be that there are very close s i m i l a r i t i e s between D.M. Armstrong's so-cal led "be l ie f - theory" of perception and the ecological approach, especia l ly understood as a form of perceptual behav-ior ism. The kinship may be mutually bene f i c ia l . The second, and f i na l point w i l l be that Ecological theor is ts do not establ ish the strong hypothesis that a l l perception is the mere resul t of the detection of invar iant propert ies. ( i i ) There are s u f f i c i e n t super f ic ia l s i m i l a r i t i e s between Armstrong's be l ie f - theory and the ecological approach to i nv i te prima facie comparison. Further invest igat ion reveals s t i l l deeper resemblances. To develop the comparison, I w i l l b r i e f l y describe Armstrong's theory, which he introduces by w r i t i n g : I t is clear that the b io logical funct ion of perception is to give the organism information about the current state of i t s own body and i t s physical environment, information that w i l l assist the organism in the conduct of l i f e . This is the most important clue to the nature of perception. I t leads to the view that perception is nothing but the acquiring of true and false be l ie fs concerning the current state of the organism's body and environ-ment . . . Veridical perception is the acquiring of true b e l i e f s , sensory i l l u s i o n the acquiring of fa lse b e l i e f s . 7 In wel l - founded'ant ic ipat ion of resistence to the a t t r i b u t i o n of be l ie fs to many evident perceivers ( including p r e - l i n g u i s t i c humans), Armstrong is quick to subject his claim to q u a l i f i c a t i o n . His notion of b e l i e f is said to be a "sub-verbal" conception, so that l i n g u i s t i c a b i l i t y is not a prerequis i te to the possession of b e l i e f s . S t i l l no doubt concerned that some w i l l regard a be l i e f as too sophist icated a state to a t t r i b u t e to some perceiving organisms, Armstrong goes on to suggest the o word " informat ion" as an a l te rna t i ve . Part of the motivation fo r using " informat ion" in par t i cu la r is that i t makes good grammatical sense to t a l k about " fa lse" in format ion, whereas Armstrong notices there is no corre late to describe what he ca l l s sensory i l l u s i o n with other a l ternat ives that he enter ta ins . Subsequent to an elaboration of his basic theory, he then proceeds to analyze be l ie fs as "states apt fo r bringing about cer ta in 9 behavior". Armstrong thus has a view on which perceptual awareness w i l l receive some form of behavioral analysis. The be l ie f - theory can be considered as a form of perceptual behaviorism even though Armstrong (as, fo r that matter, does Gibson) disavows orthodox behaviorism. A number of super f ic ia l s i m i l a r i t i e s thus ex is t between the two theor ies: 1. Perception is fundamentally conceived as the acquiring of information 2. The main funct ion of perception is to allow organisms to serve t h e i r ecological needs (conduct t h e i r l i f e ) 3. That the service of ecological needs is the chief funct ion of per-ception is taken to be the key to understanding the nature of i t 4. Organisms are said to be related t o , and to conduct t h e i r l i ves in an environment 5. The information organisms acquire is categorized as e i ther being about i t s e l f (p rop r iospec i f i c ) , or being about i t s environment (exterospeci f ic) 6 . Perception is treated as an epistemic notion These s i m i l a r a t i e s , of course, hardly suf f ice to show the two theor is ts are committed to the same view. The idea expressed in 2), as well as the concept of an "environment" are given f a i r l y extensive treatment on the ecological approach, but are undeveloped by Armstrong. He mentions 2) in ra t i ona l i z i ng the approach he takes, but i t does not pervade Armstrong's thought on such advanced issues as the terms of perception as i t does with Gibson. Gibson's use of "environment" is a technical one, whereas Arm-strong's need not be regarded as anything more than accidental . The s ig -ni f icance of 4) ia also affected by th i s point . 6 ) at least indicates that Armstrong and Gibson are endeavouring to give theories on the same topic (as opposed to authors who are wr i t ing about simple perception, or about the j u s t i f i c a t i o n , rather than the causation, of perceptual ly acquired knowledge). I have indicated two gen-eral forms of epistemic perception: "perceiving that" and "perceiving as". The main di f ference is that i t is always fa lse that one perceives that a par t i cu la r thing is P unless i t is true that the thing perceived i_s P. On the other hand, one may perceive something as P when the thing perceived is not in fac t P. Since Armstrong allows that perceiving may be the acquir ing of fa lse informat ion, his theory must be ( in spi te of what he evident ly th inks) based upon "perceiving as". Gibson occasionally makes' use of t h i s l ocu t ion , but he is unsett led on how to handle non-veridical perception. One's approach to non-veridical perception is the best indicator of a commitment to one locut ion or the other in the absence of an e x p l i c i t statement. On the handling of errors in perception there exists a clear possi-b i l i t y of d i f ference between Gibson and Armstrong, though. Many of the ecologists are incl ined to t r y to explain away apparent er rors . I t is sometimes suggested that when organisms appear to have made a perceptual e r ro r , they have simply f a i l e d to explore the i r immediate environments long or hard enough, and so have not picked up a l l the avai lable inform-a t ion . TSRM's curious normative view is another example of an attempt to explain away evident perceptual e r ror . Gibson himself, a f te r suggesting a var ie ty of examples he considers to be perceptual e r ro r , wr i tes : Optical misinformation enters in to each of these cases in a d i f f e ren t way, but in the last analys is , are they explained by misinformation? Or, is i t . .simply a matter of f a i l u r e to pickup a l l the information . . . ? Gibson would l i k e to explain away evident mistakes somehow but he is more conc i l ia to ry than some of his fo l lowers. That non-veridical percep-t ion is to be accounted fo r as the acquiring of fa lse information remains a p o s s i b i l i t y , and acceptance of such an approach would be tantamount to taking the "perceiving as" approach to epistemic perception. S imi lar ly 1) is crucia l since i t turns out to be a deeper resemblance than one might be i n i t i a l l y inc l ined to suppose. Armstrong equates "be l ie f " and"information", whereas the Gibsonians sometimes seem to be t ry ing to equate " informat ion" and " invar iant property". Armstrong is cer ta in ly not l i k e l y to suggest that informat ion, in his sense, is "simply ava i lab le" . On the other hand, some of the Gibsonians, Shaw, Turvey and Mace, expressly 12 deny a connection between perception and be l i e f . Their reasoning, however, i t both t e r r i b l y obscure and i r re levant to th is context. Once one has seen that Armstrong in terprets be l ie fs in terms of behavior-states, and that the ecological approach would lack some account o f i n t e n t i o n a l i t y in perception i f i t were not in terpreted as analyzing perceptual awareness in terms of behavior, one can see that the two approaches to perceptual information pickup bear a s t r i k i n g s i m i l a r i t y . In addi t ion to the s i m i l a r i t i e s in the two theor ies, there are two important differences that should be mentioned: 7. I t is not clear that the be l ie f - theory and the ecological approach should be seen as giv ing the same kind of behavioral analysis 8. The two theories d i f f e r on the terms in which organisms perceive t h e i r environments With respect to 8): on the issue of the categories of perception, Armstrong must be regarded as an Establishment theo r i s t . In the examples he chooses, and through his discussion of the problem of secondary q u a l i t i e s , i t is apparent that Armstrong accepts the basic categories of perception which have remained substant ia l ly unchanged since Locke's t ime. Armstrong shows no sign of deviating from the t rad i t i ona l formula that the environ-ment is perceived in terms of colours, shapes, tex tures, weights, and the l i ke (sensible q u a l i t i e s ) , and useful information is constructed out of these. Gibson, who takes affordance-perception as basic, would regard such propert ies as f e l t weight, sensed colour and perspectival shape as qua l i t i es of sensation. As such, they would neither be perceived nor the basis of perception of the environment. On a physical construal of sensible q u a l i t i e s , awareness of a thing as having a pa r t i cu la r colour, shape, or the l i k e , the Gibsonian l ine would presumably be that these are der i va t i ve , a product of the perception of ecological ly s ign i f i can t parameters of the environment. There is nothing in Armstrong's central view that is incompatible with the Gibsonian conception of the categories of percept ion, though. More-over, the be l ie f - theory could derive a couple of advantages from the idea that ecological properties are perceived. In the f i r s t place, although Armstrong takes the fac t that the chief purpose of perception is to provide an organism with information important to the "conduct of l i f e " to be a key determinant of the nature of perception, he does not make use of th i s basic point in subsequent discussion. This claim as to the purpose of perception is jus t the kind of point which motivates Gibson to adopt his theory about affordances. Gibson's approach thus shows how Armstrong could develop his claim. In add i t ion , Armstrong's theory has very l i t t l e to say about the behavior a given state is apt to bring about. What behavior, for example, is apt to be produced by the perception of some-thing as red, cub ica l , or soft? Beyond certa in obvious suggestions, such as verbal behavior or d iscr iminat ive behavior produced in a r t i f i c i a l tes t condi t ions, which have l imi ted appl icat ions nothing comes readi ly to mind. On the supposition that the environment is perceived in terms of what i t a f fo rds , i t is rather easier to give substantive descript ions of the behavior-states which const i tu te perceptually acquired informat ion. With regard to 7): there are both d isposi t ional and non-disposit ional ways of t r y ing to analyze perception. Armstrong is e x p l i c i t l y committed to a d isposi t ional view. In pa r t i cu la r , he maintains that to perceive is to acquire a b e l i e f , where a b e l i e f is analyzed as a d isposi t ional states apt for br inging about cer ta in behavior. A non-disposit ional account of perceptual information pickup in behavioral terms would be to claim that an organism perceives some phenomena when, and only when i t ac tua l ly ex-h i b i t s cer ta in behavior. TSRM might be held to a non-disposit ional view since in such examples as that of the shark or of the per iwink le, they make no mention of invar-iants producing disposi t ions to behave. I f the authors have any tendency to embrace such a view, they should abandon i t . The non-disposit ional form of perceptual behaviorism is very strong and not very p laus ib le . The main problem the non-disposit ional view raises is that every instance of a perceiver f a i l i n g to exh ib i t some sort of behavior must be counted as a f a i l u r e to perceive, even when circumstances make i t implaus-ib le to believe that the perceiver has f a i l e d to notice some environmental i tem. I f I look at a pen but do not pick i t up and wr i te with i t , must I have f a i l e d to perceive i t as wri te-withable? Does a shark, because i t only exhib i ts predatory behavior when i t is hungry, f a i l to perceive edible things unless i t is in th is state? Unquestionably, an animal's at tent ion is se lec t ive , and th is s e l e c t i v i t y is directed by present wants and needs. However, i t would be rash to suppose that an organism always f a i l s to notice features of i t s environment whenever they would not actual ly e l i c i t behavior of a cer ta in kind because the organism is in some s ta te . Various tac t i cs fo r avoiding such po ten t ia l l y embarrassing cases might be suggested. One is to expand the range of over t , physical behav-io r which const i tutes perception of the environment in pa r t i cu la r terms. For example, an opossum's running away, climbing a t ree , remaining motion-less, or defending i t s e l f may each const i tute perceptual awareness of a fox as a predator. There are l i m i t s to th i s t a c t i c , however, because the more one expands the range of relevant behavior fo r perception of some-thing as a such-and such, the more unclear i t becomes as to why th is should const i tu te perception as a such-and-such in pa r t i cu la r . As one expands the range of opossum-behavior that counts as perceiving something as a predator, i t becomes increasingly d i f f i c u l t to explain how th is d i f f e r s from perception of non-predators in cer ta in terms (other opossums as mates, competitors, and so f o r t h ) , or perception of foxes as things other than predators. I t becomes correspondingly d i f f i c u l t to explain why a cer ta in a c t i v i t y by the opossum is appropriately described as perception of a fox as a predator. Other possible tac t i cs would be to count verbal and/or i n te l l ec tua l behavior among that which const i tutes perceptual awareness. Both would obviously be l im i ted in terms of the organisms to which they could be appl ied, though. Moreover, i t would be d i f f i c u l t to argue that ca l l i ng phenomena such as " th ink ing of X as P" behavior is not merely as semantic t r i c k by which the perceptual behavior ist is allowed to smuggle in the 121 very representational component of perception that he purports to analyze away. However, philosophers attempting behavioral accounts of intent ional states abandon the non-disposit ional analysis rather than t r y to use my sugested compromises. I t is far more preferable to take up a disposi t ional account such as Armstrong suggests. Now i t is possible to sketch roughly a possible ecological analysis of perception. The f i r s t step is to characterize the detection of (or res-onation to) an invar iant by an organism's perceptual system as the acquired state that tends to produce a cer ta in kind of behavior. Given t h i s , an organism perceives X as P i f and only i f : 1. P is specif ied by some invar ian t , 0, according to the pr inc ip les of the theory of spec i f icat ion 2. An instance of 0 is detected by a perceptual system of the organism 3. The instance of 0 in question is produced by some underlying mechan-ism of X To repeat an ea r l i e r po int , Gibson does not elaborate on such func-t ions as detection or resonation. I t is therefore open to question whether these funct ions should be said to produce a behavior-state or to const i tu te the s ta te . The stronger in te rpre ta t ion that I use seems to be cal led fo r by TSRM's previously noted assertion that organisms pick up information when, and only when, t he i r perceptual systems detect/resonate to invar iants . With respect to the enumerated condi t ions, 1) s t ipu lates that P is a genuinely perceptible property. I f 2) f a i l s to obta in , the state of the organism is not the product of the normal funct ioning of i t s perceptual system. 3) is a simple causal condit ion that t i e s perception of X as P to X. The detection of 0 means "P" to the organism in that the resu l t is invar iab ly a d isposi t ion fo r behavior appropriate for i t toward things having 122 P. This is pa r t l y a theory of what i t is for perception of things to mean something to the perceiving organism. On t h i s d isposi t ional view, some fa i l u res to behave, in circumstances in which perception should be expected to occur, can be explained in terms of modif icat ions of the state which is produced by a given invar iant . The invariant-caused perceptual state may be modif ied, that i s , by other condi-t ions or states both of the perceiving organism and i t s environment. For example, the shark may detect a type F b ioe lec t r ic f i e l d , but f a i l to ex-h i b i t predatory behavior because i t is sick or not hungry, or because i t senses danger or some int rus ion into i t s surroundings. On the a l te rna t ive non-disposit ional analysis, such modifying states and condit ions must a l -ways be explained as causing fa i l u res to perceive rather than as mere f a i l -ures of the perceptual state to produce the usual behavior. However, there is a disadvantage to explaining the intent ional com-ponent in perception in terms of being in states apt fo r the production of cer ta in sorts of behavior. On the non-disposit ional view, i t is quite clear that nothing in the perceptual process could be regarded as representations of the environment in another guise. Given Armstrong's account of informa-t ion pickup in terms of the production of cer ta in behavioral states, i t would be incumbent upon an ecological theor is t to explain his notion of a state in a way which would show that i t should not be regarded as yet another form of internal representat ion. In pa r t i cu la r , he would have to con-strue these in some physical terms, as physiological states apt fo r br ing-ing about behavior, for example. Armstrong's own view is designed to leave open t h i s question of whether or not such states are mental or physical (Remember that Armstrong claims his notion of a state is compatible w i th , but does not enta i l mater ia l ism). The ecological theor is t would thus need 123 to give a convincing account of the reduction of cognit ion to physical phenomena, a task which has proven in t rac tab le . One of the problems that would have to be faced is the product of a d i s t i n c t i o n which is ignored on the ecological approach. I t is generally held that there is a d i s t i n c t i o n between the mere d iscr iminat ive behavior of organisms that is produced by the funct ioning of t he i r perceptual systems, and genuine epistemic states so produced. A marsh periwinkle successful ly climbing a plant stem as a resu l t of the funct ioning of i t s visual system would no doubt be regarded by many theor is ts as an example of mere d iscr iminat ive behavior and not as the product of perception in some intent ional sense at a l l . In f a c t , i t is a weakness of TSRM's de-fence of the ecological approach that the examples which they t rea t in any deta i l involve most unsophisticated perceivers (low-grade intent ional 13 systems in Daniel Dennett's terms ) , ranging from household scales to 14 gannets. Some philosophers, such as Karl P fe i fe r and C.B. Mart in , would undoubtedly even consider the use of such subjects as scales and bean plants in giv ing an account of the intent ional component in percep-t ion as showing that TSRM simply have the wrong c r i t e r i a fo r ident i fy ing i n t e n t i o n a l i t y . I f one were to account convincingly for epistemic perception in terms of manifest states apt fo r the production of behavior, one would need to give examples involving creatures which are capable of some reasonably high degree of novel and/or adaptive behavior as subjects. The tendency to suppose that something more is involved in perceptual information pick-up by human beings than by marsh periwinkles might u l t imate ly prove to be unfounded. Unt i l such time as i t does, the d i s t i n c t i o n between states apt to produce some behavior and genuine epistemic states, the d i s t i n c -t i on which underl ies the tendency to d i f f e r e n t i a t e between human and p e r i -winkle perceptions, has to be acknowledged. This is done by defending the view that the intent ional component in perception is accounted fo r in terms of d isposi t ional states of perceiving organisms using examples which are clear cases of epistemic perception (hence, cases which involve percep-t ion in some intent ional sense). The d i f f i c u l t y l i e s in the fac t that the ease with which one may achieve concensus on whether a given case of per-ceiving is an instance of epistemic perception is inversely proportional to the ease of giv ing convincing reasons for saying that case is analyzable in terms of the causation of some physiological state l i k e l y to produce cer ta in sorts of behavior. To summarize, i t appears that in the i r general st ructure the ecol -ogical approach and Armstrong's be l ie f - theory are qui te s im i la r . Moreover, the apparent d i f ferences, as well as the s u p e r f i c i a l i t y of some of the s i m i l a r i t i e s , are not insurmountable. They are no doubt pa r t l y a mere r e f l e c t i o n of the fac t t ha t , because Armstrong and Gibson come from d i f f e r e n t academic f i e l d s , they have somewhat d i f f e ren t p r i o r i t i e s and also ways of looking at problems. Thus there are points which both authors make that are subsequently developed in deta i l by one and not developed at a l l by the other. The main contr ibut ion Armstrong can make to the ecological approach is his idea that the operations of organisms' perceptual systems produces d ispos i t ional states of the perceiver apt for bringing about behavior. I t provides the ecological theor is ts with something they need: a way of showing that they avoid the i n i t i a l problem of lacking some account of the in ten-t ional component in perception. Also, i f i t is correct to in te rpre t the ecological approach in t h i s l i g h t , Gibson's view turns out to be somewhat less s ingu lar ly revolut ionary than some of his d isc ip les would wish. The good news is that there is a precedent fo r the theory which could make i t more accessible and understandable from a philosophical point of view. ( i i i ) To repeat, i t is s t i l l a matter of great controversy as to whether or not perception or cognit ion can be accounted fo r in terms of some kind of behavior-state. The p o s s i b i l i t y of such an analysis, however, shows tha t , contrary to Fodor and Pylyshyn's c r i t i c i s m , the Gibsonian theory is not completely lacking in providing some view about the intent ional component in perceiving. One job of the ecological theor is t is to pursue the i n t e n t i o n a l i t y issue along behavioral l ines by developing a v iable account in d e t a i l . However, there would s t i l l be at least two important reasons for reserving judgement on the general hypothesis of the ecological approach. This general hypothesis is that a l l perceptual information pickup is d i r e c t , in the sense, of being the mere resu l t of the detection of some invar iant by an organism's perceptual systems. The two reasons for remaining doubtful are as fo l lows: 1. As I have noted (along with U l r i c Neisser), i t remains to be seen whether or not enough invar iant propert ies (specifying environmental phenomena of the r i gh t sor t ) can be discovered. The view contains a c r i t e r i o n for d i f f e r e n t i a t i n g between genuine and degenerate cases which are commonly described as instances of percep-t i o n . The c r i t e r i o n is that only the cases in which the information an organism acquires is exclusively determined by the detection of an invar-iant by some of i t s perceptual systems should be considered as a genuine case. The acceptab i l i t y of t h i s c r i t e r i o n is contingent upon i t s a b i l i t y to f u l f i l l some very large promises. There is a general promise that common usage of "perceives" and/or commonsense conceptions of what is perceptible w i l l not be disrupted to too great an extent. There is the more spec i f ic promise that s u f f i c i e n t invar iants w i l l be discovered to sa t i s fy general conceptions of how much of one's environment is percep-t i b l e . Also, Gibson and his fol lowers object to t r a d i t i o n a l approaches largely by claiming that they cannot explain the or ig ins of knowledge of the environment, and that the basic terms in which such theories assume animals perceive t h e i r environments are not ecological ly usefu l . There are therefore promises to the e f fec t that such invar iants as w i l l be found in future research w i l l specify the environment in useful terms, and w i l l be of a sort to explain the knowledge various animals have of t h e i r environment. As long as i t remains for the ecological theor is ts to actual ly f i nd considerable numbers of the underlying invar iant structures which are supposed to serve as the basis of perceptual information pickup, these promises go u n f u l f i l l e d . 2. The chief general object ion to Establishment views, as Shaw, Turvey, and Mace put i t , is that they subscribe to "the doctrine of in t rac 15 table non-spec i f i c i t y " . This is j u s t to say that Establishment versions of the input to perceptual systems is not spec i f ic to features of an organ ism's environment. I t must therefore get the information via cognit ive processing. The ecological theor is ts claim th is requires p r i o r knowledge or concepts, the; means fo r representing the world in pa r t i cu la r terms. They argue that i f a l l perception is o f th is k ind , one cannot explain how an organism acquires information about i t s environment in the f i r s t place. The pickup of some information requires possession of p r io r informat ion, which requires s t i l l p r io r possession of informat ion, ad i n f i n i t um. The conclusion the ecological theor is ts draw is that perception should be conceived so that al1 perception is the d i rec t pickup of i n fo r -mation. In other words, no genuine perceptual information pickup involves the appl icat ion of p r io r knowledge (cognit ive processing) to the present deliverances of the perceptual systems. This conclusion, however, is too strong and does not fo l low from the premisses from which i t is derived. Given the i n i t i a l premisses, one can at best conclude that at least some of the genuine cases of perception cannot involve cognit ive processing. Hence, the ecological theory's most important argument agsinst Establ ish-ment views y ie lds only the conclusion which is a form of Fodor and Pyly-shyn 's Establishment pos i t ion : perceptual awareness of some features of the environment is d i rec t . Acceptance of the stronger, intended con-clusion is at least contingent at least upon the sa t is fac t ion of the promises indicated in point 1). ( i v ) Overa l l , Gibson's ecological approach does not contain the fundamen-ta l conceptual flaw suggested by certa in c r i t i c s since i t can be in te rpr -eted as analyzing the in tent ional component in perception in behavioral terms. This is a d i f f i c u l t posi t ion to argue, but i t is one with which I f i nd a great deal of sympathy. On the other hand, ecological theor is ts do not manage, by any means, to establ ish t h e i r hypothesis, par t ly because i t embodies a recommendation to adopt a c r i t e r i o n fo r determining genuine cases of perception which has not yet proven acceptable. Moreover, the "or ig ins of knowledge" argument does not y i e l d the strong conclusion desired of i t . In spi te of t h i s , the ecological approach contains many challenging ideas fo r perceptual theory, not a l l of which have been given the discus-sion they deserve in th is thes is . Nor do a l l of these ideas require acceptance of Gibson's general view. There are two such ideas that I have in mind. One is the novel conception of perceptual systems and t h e i r funct ion ing. One can only wish that Gibson had spent some time in expand-ing on such notions as the ex t rac t ion , resonating t o , and detection of invar iant s t ruc ture . The second idea is that an organism perceives i t s environment predominantly in terms of affordances. The theory of a f fo rd -ances const i tutes an in teres t ing and rare challenge to mainstream views on what is perceived. Assumptions in th is area of the categories of perception, especial ly in the philosophy of perception have too often gone unexamined. Footnotes Chapter One *The theory I w i l l ou t l ine is developed by Gibson mainly in The  Senses Considered as Perceptual Systems (Boston: Houghton M i f f l i n Co., 1966) and The "Ecological Approach to Visual Perception (Boston: Houghton M i f f l i n Co., 1979). Also helpful is James Gibson, Reasons for Realism, ed. Edward S. Reed arid Rebecca Jones (H i l l sda le , N.J.: Lawrence Erlbaum Assoc., 1982). 2 Claire F. Michaels and Claudia Carel lo, Direct Perception (Englewood C l i f f s , N.J.: Prentice-Hall Inc . , 1981), 115, 3 Jerry A. Fodor and Zenon W. Pylyshyn, "How Direct is Visual Perception?" Cognit ion, 9, no. 2 (1981): 141. 4 Gibson, Ecological Approach, 239. 5 Gibson, Ecological Approach, 75, 116; Gibson, Senses, 33-34, 200-201. Related terms include "propr ioceptor," "exteroceptor," and "p ropr iospec i f i c , " and "exterospeci f ic in format ion." ^Frank Jackson, Perception, (Cambridge: Cambridge Universi ty Press,, 1977). ^Jackson, Perception, 154-172. A classic discussion given in terms of proposit ional versus non-propositional perceiving may be found in Roderick Chisholm, Perceiving, I thaca, N.Y.: Cornell Universi ty Press, 1957. Q G.N.A. Vesey, "Seeing and Seeing As," Perceiving, Sensing and  Knowing, 68-83, ed. Robert J . Swartz (Berkeley and Los Angeles: Univers i ty of Cal i forr i ia Press, 1965), is an example of an analysis of "perceiving as" . D.M. Armstrong, A Mate r ia l i s t Theory of Mind, (London: Rout!edge and Kegan Paul, 1968), 208-244. Armstrong holds that perception is the acquir ing both of t rue and fa lse be l ie fs . 1 0 Fodor and Pylyshyn, "How Di rec t?" , passim. **S. Ullman, "Against Direct Perception," The Behavioural and Brain Sciences, 3 (1980): 375. 12 Gibson, Senses, 266. 1 3 Herman von Helmholtz, Epis temologicalWri t ings, Centenary ed i t ion of 1921 ed. Paul Hertz and Moritz Schl ick, new t rans, by Malcolm F. Lowe, ed. Robert S. Cohen and Marx W. Wartofsky, Boston Studies in the Philosophy of Science, 37 (Dordrecht, Holland and Boston: D. Reidel Publishing Co., 1977), 117 f f . *^John Locke, An Essay Concerning Human Understanding, 2 vo ls , co l la ted and annotated by A.C. Fraser (New York: Dover Publ icat ions, 1959), Book I I , chapter 9, sections 9-10. 15 Gibson, Ecological Approach, 238. 1 6 I b i d . , 251. 1 7 I b i d . . 1 8 I b i d . , 251-252. 1 9 I b i d . , 251-; Gibson, Senses, 275. "Gibson, Senses, 306 f f . Gibson, Ecological Approach, chapter one (especial ly 9-10, 12). The existence of levels of explanation and/or descr ipt ion in science of some kind is widely accepted in the philosophy of science, although there is considerable dispute as to the exact nature of these levels and as to the extent to which the hierarchy is col lapsable. I t is an in teres t ing part of Gibson's view that he believes t r a d i t i o n a l percep-tual theor is ts have d i f f i c u l t y because they have not established the proper level of explanation of psychological theories of perception and cogni t ion. One of Gibson's goals is to establ ish the proper l e v e l , mainly through his notion of perceptual systems and the parameters of physical energy to which these are sensi t ive. I w i l l not discuss the ecological approach in terms of levels of explanation and descr ip t ion. This in teres t ing issue is discussed by Lawrence Carleton in "Levels of Description and Explanation," Philosophy Research Archives, 11, 89-109 (H i l l sda le , New Jersey: Lawrence Erlbaum Assoc., 1986). 2 2 1 b i d . , 148. 2 3 1 b i d . , 148-149. 2 4 I b i d . , 220. 25 David Hume, Enquiries Concerning Human Understanding and  Concerning the Pr incip les of Morals, 3rd e d i t i o n , ed. P.H. Nidi i t ch (Oxford: Clarendon Press, 1975, a repr in t from the 1777 e d i t i o n , ed. e i t h in t roduct ion by, L.A. Selby-Bigge) 151-153. Hume's problem here is epistemological rather than psychological, though. Of. Gibson, Ecological Approach, 221. Gibson, Senses, 276. po W. Sch i f f , J.A. Caviness, and J . J . Gibson, "Persistent Fear Responses in Rhesus Mondeys to the Optical Stimulus of 'Looming'," Science, 136 (1962): 982-983. pq W. Sch i f f , "Perception o f Impending C o l l i s i o n , " Psychological Monograph, 79, no. 604. 30 Gibson, Ecological Approach, 175-176. 31 Michael T. Turvey et a l . , "Ecological Laws in Perceiving and Act ing: In Reply to Fodor and Pylyshyn (1981)," Cognit ion, 9, no. 3 (1981): 294 3 2 Gibson, Senses, 280. 3 3 R . E . Shaw, M7T. Turvey, and W.M. Mace, "Ecological Psychology: the Consequences of a Commitment to Realism," Cognition and the Symbolic  Processes, 2, ed. Walter B. Weimer and David S. Palermo (Hi l l sda le N.J.: Lawrence Erlbaum Assoc., 1982), 164-167. 34 Gibson, Ecological Approach, 253. 3 5 I b i d . , 3. O C Turvey et a l . , "Ecological Laws," 243. Similar descr ipt ion of the Ecological Approach in E.S. Reed, "Two Theories of the In ten t ion-a l i t y of Perceiving". 3 7 Gibson, Ecological Approach, 243, 262-263, 307. 3 8 I b i d . , 127, 238. Gibson, Senses, 267. 39 G.E. Moore, "Proof of an External World," Philosophical Papers (London: George Allen and Unwin L t d . , 1959; rep r in t ed. New York: Humanities Press Inc . , 1977), 127-130. 40 Gibson, Ecological Approach, 8-9. 4 1 I b i d . , chapter 2, 307. 4 2 1 b i d . , 307. 4 3 Turvey et a l . , "Ecological Laws," 237-304. 4 4 Gibson, Ecological Approach, 127. 4 5 Nelson Goodman, Fact, F i c t i o n , and Forecast, 4th ed . , with forward by H i l l a r y Putnam (Cambridge, Mass.: Harvard Universi ty Press, 1983), 40. 4 6 Gibson, Ecological Approach, 127, 137-141. Chapter Two •^Gibson, Senses, 50. 2 Gibson, Ecological Approach, 245. 3 I b i d . 4 Gibson, Senses, 50. 5 Gibson, Ecological Approach, 65-66. 6 I b i d . , 65. 7 I b i d . , 68-69. 8 I b i d . , 68. 9 I b i d . , 283-286. Gibson, Senses, 235-237. 1 0 Gibson, Ecological Approach, 73. 1 1 I b i d . , 87. 1 2 I b i d . , 247. 1 3 1 b i d . , 166-168, 169, 244. There may be others: t h i s is a favour i te point of Gibson's. 1 4 I b i d . , 78-86. Gibson, Senses, 203-206.. 15 Gibson, Ecological Approach, 165. 1 6 I b i d . , 310. 1 7 I b i d . , 87-91. 1 8 Michaels and Carel lo, Direct Perception, 21-22. 19 Gibson, Ecological Approach, 114. Gibson, Senses, 237. "Gibson, Ecological Approach, 208-209. Gibson, Senses, 284-285. pi Gibson, Ecological Approach, 95-100. 2 2 I b i d . , 10-12, 100-102. 2 3 I b i d . , 175. 24 David Lee, A Theory of Visual Control of Braking Based on Information about Time-to C o l l i s i o n , " Perception, 5 (1976): 431-459. o r Gibson, Senses, 280 ( in passing). 2 6 I b i d . , 237. 27 Locke, Essay, Book I I , chapter 9, sections 8-10. 28 Gibson, Ecological Approach, 310. 2 9 I b i d . , 63. 30 J U Gibson, Senses, 245. 31 Turvey et a l . , "Ecological Laws," 243. Chapter Three Michaels and Carel lo, Direct Perception, ch. 5; Shaw, Turvey, and Mace, "Ecological Psychology," passim; E.S. Reed and Rebecca Jones, "Gibson's Theory of Perception: A Case of Hasty Epistemologizing?" Philosophy of Science, 45, no. 4 (Dec. 1978): 526-529 are three examples, 2Fodor and Pylyshyn, "How Direct?" 139-196. 3 Turvey et a l . , "Ecological Laws," 240. 4Fodor and Pylyshyn, "How Direct?" 141. 5 I b i d . 6 I b i d . , 142. 7 Gibson, Ecological Approach, 311. 8 I b i d . , 203-205. g Michaels and Carel lo, Direct Perception, 178. 1 0 Gibson, Ecological Approach, 134-135, 240. **Turvey et a l . , "Ecological Laws," 261. 12 Ludwig Wit tgenstein, The Blue and Brown Books, (New York: Harper and Row, Pub., 1958; Harper Torchbooks, 1965). 1 3 Fodor and Pylyshyn, "How Direct?" 169. 1 4 Turvey et a l . , "Ecological Laws," 282-284. 1 5 I b i d . , 244-245. Fodor and Pylyshyn, "Ecological Laws," 188. 1 7 U l r i c Neisser, "Gibson's Ecological Optics: Consequences of Di f ferent Stimulus Descr ip t ion," Journal for Theory of Social Behaviour, 7, no. 1 (Apr. 1977): 24. 18 Robert E. Shaw and John Bransford, " In t roduct ion : Psychological Approaches to the Problem of Knowldge," Perceiving, Act ing, and  Knowing, ed. R.E. Shaw and J . Bransford ("Hillsdale, N.J. : Lawrence Erlbaum Assoc., 1977), 1-10. 19 John Heil uses t h i s term to describe the problem in Gibson's approach that (he suggests) both D.W. Garniyn and Ul r ic Neisser attempt to point out in "What Gibson's Missing," Journal for the Theory of Social Behaviour, 9, no. 3 (Oct. 1979): 265-269. ?o Fodor and Pylyshyn, "How Direct?" 167. Chapter Four ^ o d o r and Pylyshyn, "How Direct?" 192. 2 D.W. Hamlyn, Perception, Learning, and the Self (London: Rout!edge arid Kegan Paul, 1983), 30-42. 3 TSRM's references to Putnam are H i l l a r y Putnam, " Is Semantics Possible?" Language, BeTief, and Metaphysics, ed. H.E. Kiefer and M.K. Munitz (New York: SUNY Press, 1970), 50-63 and "On Propert ies," Essays  in Honour of Carl Hempel, ed. Nichol Rescher et a l . (Dordrecht, Holland D. Reidel Publishing Co., 1969), 234-254. 4 Fred I . Dretske, "Laws of Nature," Philosophy of Science, 44, no. 2 (June 1977): 248-268. 5 Turvey et a l . , "Ecological Laws," 255. Shaw, Turvey, and Mace, "Ecological Psychology," 182-185, 194-203. ^Paul V. Hamilton, "Dai ly Movements and Visual Location of Plant Stems by L i t t o r i n a I r ro ra ta (Mollusca: Gastropoda)," Marine Behaviour and Physiology, 4, (1977): 293-304. o Turvey et a l . , "Ecological Laws," 249. 9 I b i d . , 268. 1 0 I b i d . , 256. n i b i d . , 265. 1 2 I b i d . , 251. 13 Gibson, Ecological Approach, 127. Turvey et a l . , "Ecological Laws," 267. 1 4 Turvey et a l . , "Ecological Laws," 262. 1 5 B y G. Wedeking. Turvey et a l . , "Ecological Laws," 264. I b i d . , 255. 'Putnam, "Semantics?" 50. ' i b i d . , 51 , 53-54. ' i b i d . , 50. I b i d . , 58-59. Turvey et a l . , "Ecological Laws," 264. Chapter Five ^Turvey et a l . , "Ecological Laws," 273. 2 I b i d . , 265, 267-275. 3 I b i d . , 266. 4 I b i d . , 249-250, 284-295. 5 Dretske, "Laws," 252. 6 I b i d . , 252-253. 7 I b i d . , 267. Q Turvey et a l . , "Ecological Laws," 266. 9 Dretske, "Laws," 253 n. •^Turvey et a l . , in "Ecological Laws," 237, favourably note a comment of Gibson's that would indicate the lawful re la t ions they have in mind are causal ones. In the f i r s t place, causal re la t ions are not symmetrical, so some story needs to be to ld as to how TSRM ar r ive at the arrangement of the predicates in t he i r " t r a n s i t i v i t y argument" since i t does make a di f ference to the t r u t h of the premisses (because causal re la t ions are not symmetrical) as to how the predicates are ordered. In add i t ion , there is no causal re la t ion that f i t s the order TSRM use. Suppose the re la t ion is"causes": While i t makes sense to suggest that an occurrent, physical property is the cause of affordance (o-ness — > e-ness), i t is c lea r l y fa lse that an opt ica l property causes the occurrent property (e-ness — > o-ness). Suppose, then, that the re la t ion is " i s caused by": the resu l t is the reverse of the foregoing. I t could be true that an opt ical invar iant is caused by an occurrent property (e-ness — ^ o-ness) but i t could not be t rue that an occurrent property is caused by an affordance (o-ness — c - n e s s ) . Moreover, the affordance neither causes nor are caused by an opt ical invar iant . I f the re la t ion is causal, there-fo re , the order of the predicates o, c, and e, matter in TSRM's argument, and as they are presently arranged, at least one 135 of the premisses is bound to be fa lse no matter which way one t r i e s to express the causal r e l a t i o n . ^Turvey et a l . , "Ecological Laws," 260. 1 2 I b i d . , 268-269. 1 3 Dre tske , "Laws," 250. 1 4 Turvey et a l . , "Ecological Laws," 260. 1 5 Fodor and Pylyshyn, "How Direct?" 168, 188-193. ' * 6 Gibson, Ecological Approach, 127. 1 7 Turvey et a l . , "Ecological Laws," 274. 1 8 Dre tske , "Laws," 266. 1 9 I b i d . , 249 n. uTurvey et a l . , "Ecological Laws," 274. 2 1 Dre tske , "Laws," 253. 2 2 I b i d . , 253 n. 2 3 Turvey et a l . , "Ecological Laws," 274. 2 4 Dre tske , "Laws," 265. Chapter Six ^Gibson, Ecological Approach, 239; Shaw, Turvey, and Mace, "Ecological Psychology," 191-194. 2 Gibson, Ecological Approach, 134. Dretske, Knowledge and the Flow of Information (Cambridge, Massachusetts: MIT Press, 1981), 202-206. Dretske's main complaint is that he thinks one puts the matter backwards by analyzing (or giv ing "content" to) in tent ional states in terms of the behaviour or d i s -posi t ions to behave that are produced by them. He believes on the contrary, that i t is the content an intent ional state already has which explains the output. This is by no means an or ig ina l c r i t i c i s m of behavioural analysis of in tent ional states. 4Turvey et a l . , "Ecological Laws," 290. 5 I b i d . , 275-282. 6 I b i d . , 277. Armstrong, Material i s t , 209. I b i d . , 210. 9 I b i d . , 245-248. 1 0 I b i d . , 227-229. ^Gibson, Ecological Approach, 243. 12 Shaw, Turvey, and Mace, "Ecological Psychology," 177-178, 182-183. 13 Daniel Dennett, "Conditions of Personhood," The_ Iden t i t i es of  Persons, ed. Amelie 0. Rorty (Berkeley and Los Angeles: Universi ty of Ca l i fo rn ia Press, 1976), 179-180. * \ a r l P fe i fe r and C.B. Mart in, " I n t e n t i o n a l i t y and the Non-psychological ," Philosophy and Phenomenological Research, 46, no. 4 (June 1986): 531-554. Pfe i fe r and Martin argue that many typ ical accounts of I n t e n t i o n a l i t y apply to d isposi t ional propert ies of purely physical systems. They would take th i s to show that the c r i t e r i a for iden t i f y ing i n t e n t i o n a l i t y in question is wrong. Hence, TSRM's view about i n t e n t i o n a l i t y , which considers " s e n s i t i v i t y to four ounces or more" to be an intent ional state of a balance would be taken by Pfe i fe r and Martin as showing that the ecological theor is ts have the wrong c r i t e r i a fo r ident i fy ing intent ional states. Shaw, Turvey, and Mace, "Ecological Psychology," 164-167. Sources Cited Armstrong, David M. A Mate r ia l i s t Theory of Mind. London: Rout!edge and Kegan Paul, 1968. Carleton, Lawrence R. "Levels in Description and Explanation." Philosophy Research Archives, 11 (March, 1986): 89-109. Chisholm, Roderick. Perceiving: A Philosophical Study. I thaca, New York: Cornell Universi ty Press, 1957. Dennett, Daniel. "Conditions of Personhood." In The Iden t i t i es of  Persons, 175-196. Edited by Amelie Oksenberg Rorty. Berkeley and Los Angeles: Univers i ty of Cal i forn ia Press, 1976. Dretske, Fred I . Knowledge and the Flow of Information. Cambridge, Massachusetts?' MIT Press, 1981. . "Laws of Nature." Philosophy of Science, 44 (1977): 248-268. Fodor, Jerry A. and Pylyshyn, Zenon W. "How Direct is Visual Percep-t ion? : Some Reflections on Gibson's 'Ecological Approach'." Cognit ion, 9, no. 2 (1981): 139-196. Gibson, James J . The Ecological Approach to Visual Perception. Boston: Houghton M i f f l i n Company, 1979. . Reasons fo r Realism. Edited by Edward S. Reed and Rebecca Jones" H i l l sda le , New Jersey: Lawrence Erlbaum Associates, 1982. _______ The Senses Considered as Perceptual Systems. Boston: Houghton M i f f l i n Company, 1966. Goodman, Nelson. Fact, F ic t ion and Forecast. 4th ed. Foward by H i l l a r y Putnam. Cambridge, Massachusetts: Harvard Univers i ty Press, 1983. Hamilton, Paul V. "Dai ly Movements and Visual Location of Plant Stems by L i t t o r i n a I r ro ra ta (Mollusca: Gastropoda)." Marine Behaviour and Physiology, 4 (1977): 293-304. Hamlyn, David W. Perception, Learning and the Self : Essays in the Philosophy Of Psychology. London: Routledge and Kegan Paul, 1983. H e i l , John. "What Gibson's Missing." Journal for the Theory of Social Behaviour, 9, no. 3 (October 1979): 265-269. Helmholtz, Hermann von. Epistemological Wri t ings. Centenary edi t ion of 1921 edited by Paul Hertz and Mori tz Schlick. New t rans la t ion by Malcolm F. Lowe. Edited by Robert S. Cohen and Marx W. Wartofsky. Boston Studies in the Philosophy of Science, 37. Dordrecht, Holland and Boston: D. Reidel Publishing Co., 1977. Hume, David. Enquiries Concerning Human Understanding and Concerning  the Pr inc ip les 'o f Morals. Reprinted from the edi t ion of 1777. 3rd ed i t i on . Edited, with an In t roduct ion, by L.A. Selby-Bigge. Third ed i t ion edited by P.H. Niddi tch. Oxford: Clarendon Press, 1975. Jackson, Frank. Perception. Cambridge: Cambridge Univers i ty Press, 1977. Lee, David. "A Theory of Visual Control of Braking Based on Information about T ime- to-Col l i s ion . " Perception, 5 (1976): 431-459. Locke, John. An Essay Concerning Human Understanding. 2 vo ls . Collated and annotated, with prolegomena, by Alexander Cambell Fraser. New York: Dover Publ icat ions, 1959. Michaels, Claire F. and Carel lo, Claudia. Direct Perception. Englewood C l i f f s , New Jersey: Prentice-Hall Incorporated, 1981. Moore, George Edward. "Proof of an External World." In George Edward Moore, Philosophical Papers, 138-150. London: George Al len and Unwin L t d . , 1959; repr in t e d i t i o n , New York: Humanities Press Incorporated, 1977. Neisser, U l r i c . "Gibson's Ecological Opt ics." Journal fo r the Theory Social Behaviour, 7, no. 1 (Apr. 1977): 17-28. P fe i f e r , Karl and Mart in , Charl ie B. " I n t e n t i o n a l i t y and the Non-Psychological." Philosophy and Phenomenological Research, 46, no. 4 (June, 1986): 531-554. Putnam, H i l l a r y . " Is Semantics Possible?" In Language, Be l ie f , and  Metaphysics, 50-63. Edited by Howard E. Keifer and Mi l ton K. Munitz. Albany, N.Y.: SUNY Press, 1970. ; . "On Propert ies." In Essays in Honor_of Carl G. Hempel, " 234-254. Edited by Nicholas Rescher et a l . Dordrecht, Holland: D. Reidel Publishing Co., 1969. Reed, Edward S. "Two Theories of the I n t e n t i o n a l i t y of Perceiving." Synthese, 54 (1983): 85-94. Reed, Edward S. and Jones, Rebecca. "Gibson's Theory of Perception: A Case of Hasty Epistemologizing?" Sch i f f , W. "Perception of Impending C o l l i s i o n . " Psychological Mono-graphs, 79, no. 604 (1965). Sch i f f , W., Caviness, J .A. , and Gibson, James J . "Persistent Fear Responses in Rhesus Monkeys to the Optical Stimulus of 'Looming'." Science, 136 (1962): 982-983. Shaw, Robert E. and Bransford, John. " In t roduct ion: Psychological Approaches to the Problem of Knowledge." In Perceiving, Act ing ,  and Knowing. Edited by Robert E. Shaw and John Bransford. H i l l sda le New Jersey: Lawrence Erlbaum Associates, 1977. Shaw, Robert E., Turvey, Michael T . , and Mace, Will iam M. "Ecological Psychology: The Consequence of a Commitment to Realism." In Cognition and the Symbolic Processes, 2, 159-226. 2 volumes. Edited by Walter B. Weimer and David S. Palermo. H i l l sda le , New Jersey: Lawrence Erlbaum Associates, 1982. Turvey, Michael T . , Shaw, Robert E., Reed, Edward S., and Mace, Will iam M. "Ecological Laws in Perceiving and Act ing: In Response to Fodor and Pylyshyn (1981)." Cognit ion, 9, no. 3 (1981): 237-304. UHman, S. "Against Direct Perception." The Behavioral and Brain Sciences, 3 (1980): 373-415. Vesey, Got t f r ied N.A. "Seeing and Seeing As." In Perceiving, Sensing,  and Knowing, 68-83. Edited, wi th an In t roduct ion, by Robert J . Swartz. Berkeley and Los Angeles: Universi ty of Ca l i fo rn ia Press, 1965. Wit tgenstein, Ludwig. The Blue and Brown Books. New York: Harper and Row, Publishers, 1958; Harper Torchbooks, 1965. 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0097806/manifest

Comment

Related Items