J. J. Gibson’s most radical idea: The development of a new law-based psychology

A new law-based psychology for agents in their environments

Ecological psychology (Gibson, 1966, 1979) is a law-based approach to psychology. This is reflected both in Gibson’s works and in almost 50 years of ecological research. Seminal works in ecological psychology explicitly appeal to ecological laws of perceiving and acting (Turvey, 1992; Turvey, Shaw, Reed, & Mace, 1981) and the most successful models within the ecological tradition take the form of law-like differential equations (Haken, Kelso, & Bunz, 1985; Lee, 2009; Warren, 2006). Regarding its law-based nature, nevertheless, ecological psychology is different from other lawful psychophysical and physiological approaches. Ecological psychology aims to find lawful regularities between perception and action at the level of animal–environment interactions—the ecological scale. Moreover, Gibson aimed to do so by postulating neither “mental laws,” such as laws of association or Gestalt laws of sensory organization, nor internal mechanisms such as to those used by the dominant paradigms in contemporary cognitive science.

In his autobiography of 1967, Gibson talked about the kind of scientific explanations he preferred:

As the reader may gather, I prefer radical solutions to scientific problems whenever possible. General explanations are always preferable to piecemeal explanations (“models” as they are nowadays called), and this is all that is meant by a radical theory. (1967/1982a, p. 14)

According to Gibson, the radicality of an explanation has to do with its generality, but he also held a specific idea of what “to explain” is: while referring to his interests in politics and social issues, Gibson claimed that “to understand, to be able to explain and predict, entails the knowing of laws” (1967/1982a, p. 15), and right after that he chastised psychologists for their inability to find psychological laws for the social realm.

The difficulty of developing a law-based explanation for political and social issues caused Gibson to reluctantly abandon the field of social psychology. His interests focused on perception, but he maintained his conviction regarding scientific explanations. For example, when in the late 1950s he adopted the term ecology from Egon Brunswik and Gordon Walls, he nevertheless rejected Brunswik’s idea of probabilistic correlations on the ecological scale. According to Edward Reed (1988) in his biography of Gibson, this is because:

Like Einstein, Gibson was willing to accept statistical models for lack of better descriptions of complex phenomena, but he believed that lawfully deterministic explanations should be forthcoming. (p. 207)

The belief in lawfully deterministic explanations distinguishes Gibson from Brunswik in a relevant sense. Although Gibson (1957) acknowledged that Brunswik’s probabilistic functionalism was insightful, its probabilistic character made it “imperfectly lawful” (p. 33) and required the perceptual system to act as an “intuitive statistician” (p. 33). Gibson rejected Brunswik’s proposal as a fruitful law-based explanatory strategy because of these two reasons. Gibson’s preference for deterministic explanations accounts for the fact that he regarded an imperfectly lawful system as problematic. His skepticism in considering perceptual systems as intuitive statisticians, however, reveals another important aspect of his understanding of what a law-based explanatory strategy for psychology should look like.

In the case of Gibson, a distrust in the proposals that made use of mental or internal resources to explain those phenomena—for example, laws of association, computational processes, probabilistic/statistical engines, etc.—complements the preference for general, law-based explanations of psychological phenomena. He believed that laws captured the causal determinations or correspondences within the phenomena of interest at the scale of those phenomena, as it seems to occur in the physical sciences, for instance. In this sense, placing mental or internal resources mediating between the experience of the world and the world itself was a symptom of lacking those causal correspondences in psychology and therefore a lack of a proper explanation:

We [J. J. and Eleanor Gibson] have no patience with the attempt to patch up the S-R formula with hypothesis of mediation. In behavior theory as well as in psychophysics you either find causal relations or you do not. (Gibson, 1967/1982a, p. 12)

Importantly, the unwillingness to posit mediatory mechanisms in order to explain psychological events does not entail a commitment regarding the non-lawful nature of the theories that posit them. Many theories that posit mediatory mechanisms may be considered instances of some form of law-based explanation. For example, Gestalt psychology posited mental laws of sensory organization and the computational mechanisms posited by cognitivist theories are lawful at many different scales (e.g., physically lawful, computationally lawful, and so on). Classic psychophysical and physiological approaches to psychological events—including Gibson’s (1933, 1937) own research into adaptation to curved and tilted lines—could also be understood in a lawful way. However, according to Gibson, none of them were lawful at the proper scale of psychology. None of them describes lawful causal regularities between perception, action, and environmental states: their lawfulness has to do either with the activity of sub-personal mechanisms or with simple bits of stimulation and physiological reactions and sensations. The uniqueness of Gibson’s law-based explanatory strategy for psychology was the search for lawful relations at the scale of the perceptual events that constitute one psychological aspect of the interaction between organisms and their environments.

Such a quest for a general, law-based explanation of perception with reference to no mediatory resources was Gibson’s main project. In doing so, he had to re-evaluate his own convictions regarding sensation and perception many times. He also had to reject classic psychological conceptions and introduce new ones. However, his conviction regarding what counted as a good explanation remained constant as he was developing ecological psychology. He never felt the need to justify his law-based explanatory strategy, but he needed to create one that accounted for perceptual events at their own scale. The historical development of ecological psychology may be seen as a process to fulfill such an aim. I turn to that process now.

Until 1950: Dissident psychophysics

J. J. Gibson started his professional career in 1928, when he was hired to teach experimental psychology at Smith College (Northampton, MA). Before this, Gibson received a PhD in psychology at Princeton under the supervision of H. S. Langfeld. Gibson was actually a philosophy student interested in American pragmatism who took a class on experimental psychology with Langfeld in his senior year and who was offered an assistantship afterwards. During Gibson’s assistantship, former Harvard faculty member E. B. Holt came to Princeton and became his fundamental long-lasting influence (see Charles, 2010, 2012; Heft, 2001). Holt was a radical empiricist who embraced a behaviorist methodology and developed molar behaviorism: a lawful approach in which the relata are not stimulus and response but objects of the world and behavior considered as a coordinated totality. From Holt’s (1915b) point of view, “the fairly accurate description of [psychological] activity will invariably reveal a law (or laws) whereby this activity is shown to be a constant function of some aspect of the objective world” (p. 370). ³ Gibson, who always defined himself as a “Holtian philosophical behaviorist” (Gibson, 1967/1982a; Heft, 2001), likely took from Holt the conviction of the pertinence of a law-based explanatory strategy in psychology and the idea of the mutuality between organisms and environments that would be central for his later ecological approach. Also, Holt’s conception of wish as a purposive kind of action ⁴ was, along with Woodworth’s conception of motivational perception, the basis for Gibson’s active account of perceptual events.

Gibson was, thus, a radical empiricist when he first arrived at Smith College and met Kurt Koffka, the prominent Gestalt psychologist. Gibson was not especially interested in Gestalt psychology during these years (E. J. Gibson, 2001, p. 27), but the few years he spent with Koffka and the later exposure to Gestalt theory and Phenomenology influenced him to a great extent (Käufer & Chemero, 2015). By the end of his life, in 1979, Gibson would claim that his own ecological approach was “a sort of ecological Gestalt theory” (1979/1982g, p. 112). We will see, however, that Gibson’s relation to Gestalt psychology was in constant tension during his life.

In the 1930s, Gibson started gaining relevance as a perceptual psychologist with different studies of, for example, the Gibson effect (Gibson, 1933, 1937) or the relation between perception and driving (Gibson & Crooks, 1938). Due to this, he was recruited by the US Army Air Force in 1942. In the following years, Gibson studied the nature of spatial perception in aviation and developed tests and training tools for prospective pilots. The need for applying perceptual theories to a practical environment convinced him of the insufficiency of classic theories to account for spatial perception. ⁵ Gibson then started developing his own perceptual theory.

The efforts to provide a better theory for spatial perception during the 1940s culminated in his book The Perception of the Visual World (Gibson, 1950). In the book, Gibson proposed what he defined as perceptual psychophysics. To do so, Gibson embraced the psychophysical explanatory strategy:

The policy of searching for a stimulus variable with which some quality of experience may prove to be in correspondence is the policy which underlies psychophysical methods in psychology. It is the first step in the explanation of experience. (Gibson, 1950, p. 8)

However, while classic psychophysics searched for correspondence between stimulation and sensation, Gibson was looking for correspondence between stimulation and perception: “There are laws relating perception to physical stimulation as well as laws relating it to physiological processes. Explanation is a matter of lawfulness” (1950, p. 8).

Gibson’s strategy in The Perception of the Visual World reveals some aspects of his perceptual theory. First, it was a law-based perceptual theory. As it should be clear by now, Gibson took explanation to be a matter of laws. In the case of perceptual events, he aimed to unveil the lawful correspondence between stimulation and perception. Such kinds of lawful correspondence would be labelled as specification in future works. Second, perceptual psychophysics required a new description of stimulation. Classic psychophysics searched for the stimulus–sensation lawful correspondence in an atomistic sense: punctuated bits of stimulation corresponded to punctuated sensations. Gibson’s perceptual psychophysics searched for lawful correspondence (or specification) between stimulation as a whole and perception. To achieve that aim, Gibson had to introduce the idea of high-order patterns of stimulation, both in the spatial and the temporal sense (e.g., texture gradients and ordinal stimulation), that lawfully corresponded to (or specified) features of the environment perceived through them (e.g., distance to an object, forms of objects, edges, and so on). ⁶

A third aspect of Gibson’s perceptual psychophysics is that it permitted avoidance of the appeal to mental or internal resources to account for the characterization of perceptual processes. In other words, as Gibson offered a theory that traced lawful correspondences between stimulation and perception, no mental or neural mechanism was needed to connect them. In this sense, Gibson avoided both classic theories of perception—received from Modern philosophy and Helmholtz, according to whom a kind of unconscious inference was necessary in order to get that connection—and Gestalt theory of unconscious, automatic neural laws of organization of stimulation:

Why is the perceived form specifically related to the retinal form? Only if there is a specific relation is the perception good for anything, since only thus can it be related to the outside world. The Gestalt theorists… assumed that [perception] could not be wholly specific to the retinal image (like other theorists before them) and went on to account for this discrepancy by a theory of dynamical processes in the brain. The laws of sensory organization were the expressions of such processes. (Gibson, 1950, p. 192)

Finally, the last aspect The Perception of the Visual World reveals from Gibson’s first theory of perception is that he maintained one classic assumption: that the retinal image is the basis of perception. Gibson achieved a theory based on the lawful specification of the environment in the retinal stimulation by introducing new concepts such as texture gradients and ordinal stimulation. However, his reliance on retinal images produced several theoretical and experimental anomalies in his first theory. Then, he realized that maintaining the classic assumption entailed that his theory was unable to capture the phenomenon of perception in a law-based manner at the perceptual scale he was aiming for. Gibson spent the next 15 years trying to shape his theory to meet such explanatory requirements.

From 1950 to 1966: From stimulus–response to perception–action

When The Perception of the Visual World appeared in 1950, J. J. and Eleanor Gibson had already moved to Cornell. The book was well received and James Gibson spent the next few years testing some of the hypotheses and ideas he developed in it. Some anomalies regarding the retinal image persuaded him of the need to abandon the idea as it had led to “all sorts of insoluble problems” (Gibson, 1967/1982a, p. 20; see also E. J. Gibson, 2001, pp. 75–76). The full list of anomalies was analyzed by Gibson himself several years later (1970/1982f), but the possibility of having stimulation without perception (e.g., Metzger’s experiments on the Gandzfeld), the dissimilarities between retinal patterns of activation and actual pictures or images, and the non-lawful correspondence between retinal textures and the perceived slant of environmental surfaces are among the chief ones. Also, Gibson noticed that a framework in which the features of the perceptual world were lawfully correlated with the retinal image did not account for the effects of object motion in the latter (Gibson, 1957/1982b). The problems with the perception of the slant of surfaces and the perception of motion were in direct contradiction with his idiosyncratic conviction regarding the causal and law-based foundations of scientific theorizing (Reed & Jones, 1982, pp. 24–25).

In the mid-1950s, Gibson started developing one of the central concepts of ecological psychology: ecological optics. Ecological optics allowed for abandoning the retinal image as the foundation for visual perception and constituted a decisive step towards the law-based explanatory strategy of ecological psychology. ⁷ The first instances of the concept of ecological optics (Gibson, 1960/1982c, 1961/1982d) and the correlated concept of optic array were based on the thesis of the existence of an optical structure introduced a few years before (Gibson, Purdy, & Lawrence, 1955). In an illuminated environment, light coming from any source (i.e., sun, bulbs) reverberates off the surfaces it reaches and is structured by their layout, material, and color. The result of this process is an optical structure that entails an optic array: a solid angle of 360º at any point of observation whose structured arrangement of light lawfully corresponds to the layout of the environment. Such a lawful correspondence is described in terms of the laws of ecological optics (Gibson, 1961/1982d, p. 69) and is the key for the optic array to be informative of the environment. ⁸ Importantly, the structure of the optic array provides information about the environment but cannot be considered sensory stimulation anymore: information is not imposed on the perceiver, there can be stimulation without information (e.g., a foggy environment or the case of Metzger’s Ganzfeld), and there can be information without stimulation (e.g., an occluded object).

The scope of ecological optics is wider than it might seem prima facie. Ecological optics is not just a way to describe the lawful correspondence between the environment and the structure of the ambient light. By virtue of such a lawful correspondence, ecological optics is Gibson’s proposal to avoid retinal images as the foundation of perception, in particular, and to replace classic psychophysics and the perceptual psychophysics he developed in The Perception of the Visual World (1950), in general. Neither classic psychophysics nor Gibson’s perceptual psychophysics were able to provide a law-based account of the relation between environmental stimulation and perceptual events at their own scale. On the one hand, classic psychophysics needed mental or neural resources to account for the transition from the stimulus–sensation lawful relation to perception. On the other hand, perceptual psychophysics offered anomalous, inexact results regarding some basic perceptual facts of surfaces and motion. Ecological optics, nevertheless, provided a law-based account of the relation of the layout of the environment and the structure of the optic array. Moreover, the perceiver could be in (perceptual) contact with the optic array just by being situated in some point of an illuminated environment (i.e., point of observation) with open eyes. Such is the way ecological optics was able to substitute any kind of psychophysics and its related concept of the retinal image:

The pattern of the excited receptors [retinal image] is of no account; what counts is the external pattern [optic array] that is temporarily occupied by excited receptors as the eyes roam over the world, or as the skin moves over an object. (Gibson, 1966, p. 4)

The relevant aspect of the optic array regarding perception is that it is informative of the environment; but its informative character is detached from stimulation. This permitted Gibson to avoid the appeal to the retinal image as the foundation of perception. However, there was an open question: if the senses are not receptors of stimulation and channels of the resulting sensations to the brain, what are the senses and how can they be sensitive to information in the optic array?

Gibson answered this question in The Senses Considered as Perceptual Systems (1966). The main thesis of the book is that senses are not (only) channels of sensation but active detectors of perceptual information. This is the sense in which senses are perceptual systems: self-organized functional assemblies of different organs able to actively detect the information available in different ambient energy arrays—aka the theory of information pick-up (Gibson, 1966, 1979; Michaels & Carello, 1981). The visual system, for example, consists of the eyes, the optical nerve, and the brain, but also of the movable head in a movable body, etc. The whole self-organized functional assembly, including its movements and activities, is the perceptual system.

The description of the senses as perceptual systems has two main consequences for the Gibsonian theory. First, the fact that perceptual systems are active entails a complete rejection of the classic stimulus–response formula. Perceptual events are better characterized if they are understood in terms of perception–action loops in which perception is “intrinsic to the flow of activity, not extrinsic to it; dependent on it, not independent of it” (Gibson, 1966, p. 31). ⁹

And second, the description of the senses as perceptual systems entails a new way to understand physiology. Classic physiology was concerned with the lawful relation between the physiological reaction of sensory receptors and sensations, leaving open the question of how these sensations become integrated into perceptual events—by laws of association, or laws of sensory organization, or computational processes, etc. The Senses Considered as Perceptual Systems (Gibson, 1966), however, offered a way to break with classic physiology. The physiological reactions of sensory receptors, nerves, and other organs are vicarious of the function of perceptual systems as wholes (Gibson, 1966, pp. 4–5), and the function of perceptual systems is to extract information from the ambient energy arrays, not to provide sensations or to build perceptual events out of them (1966, p. 251). An adequate physiological account of perception, thus, must abandon the atomistic approach based on simple sensations triggered by the activation of sensory units and must explore the lawful regularities in the relation between the information available in ambient sensory arrays and the activity of perceptual systems. In other words, psychology must offer a law-based explanation of perceptual events in which they are lawfully determined by the availability of environmental information and vice versa.

Until 1979 and beyond: Ecological psychology

At the end of The Senses Considered as Perceptual Systems (1966), Gibson anticipated the following steps of the ecological approach to perception and action, highlighting the relation between perceptual events and the control of behavior:

The activity of orienting and that of exploring and selecting—the commonsense faculty of attending—is seen to be one that extracts the external information from the stimulus flux while registering the change as subjective feeling. This feedback system also, of course, controls the performatory activity of the body, the executive systems of behavior proper as distinguished from perception [emphasis added], but that aspect of proprioception lies outside the scope of this book. (Gibson, 1966, p. 320)

On the one hand, Gibson restated the central thesis of his second book. The physiology of perceptual systems must account for their functional activity as information detectors and the way such an activity is lawfully related to the perceptual information available in the environment. This is a thesis about what kind of activity perception is. On the other hand, Gibson pointed out that such a perceptual activity also controls action (behavior). By rejecting the stimulus–response formula, Gibson already noted the importance of action for perception. Now, at the end of his second book, he was noting that perception also plays a control function for action. In other words, that perception and action are coupled in a constant loop. The law-based explanation of such a loop at its proper scale is the aim of ecological psychology (see also E. J. Gibson, 2001, p. 80).

The coupling between perception and action in terms of loops in which actions enable perception and perception allows for the control of actions has its roots, at least in the case of Gibson, in American pragmatism and functional psychology. The idea is similar to Holt’s molar behaviorism and his notion of wish (Heft, 2001; Holt, 1915a), and is also present in Dewey’s famous critique of the reflex-arc in psychology (Dewey, 1896). Gibson himself had defended a similar notion in an earlier work on automotive locomotion (Gibson & Crooks, 1938) and definitely proposed an elaborated version of it in his 1958 article, “Visually Controlled Locomotion and Visual Orientation in Animals” (Gibson, 1958/1982h). In that article, Gibson analyzed the way behavior generates transformations or flows in the available optic structure and the way those changes (i.e., optic flow) are used in return to visually control behavior. However, in 1958, he had not fully abandoned the idea of the retinal image as the foundation for perception. He was still developing the concepts of ecological optics and needed to take some steps to fully articulate The Senses Considered as Perceptual Systems (1966). After this book, however, he already had the conceptual tools to re-elaborate his perceptual theory and the intrinsic relation between perception and action. In other words, he had the tools to construct ecological psychology. Ecological psychology was presented in his last and most famous work, The Ecological Approach to Visual Perception (Gibson, 1979), which begins with a statement of intent:

This is a book about how we see. How do we see the environment around us? How do we see its surfaces, their layout, and their colors and textures? How do we see where we are in the environment? How do we see whether or not we are moving and, if we are, where we are going? How do we see what things are good for? How do we see how to do things, to thread a needle or drive an automobile? Why do things look as they do? (p. 1)

Gibson made it clear that it was a book about our perceptual contact with the environment, but also a book on how perception helps us to perform and control other activities. Right after this statement, Gibson claimed that his last book was a sequel to The Perception of the Visual World (1950). However, it was not based on the retinal image, but on an ecological approach to perception: an approach based on searching for “relations between events and things in the world and the perceiver” (E. J. Gibson, 2001, p. 80).

Ecological psychology aims to find lawful or law-like relations between agents and their environments in terms of perception–action couplings. To achieve such an aim for visual perception, Gibson first described the environment in terms of what he sometimes called ecological physics: the lawful description of the layout of environmental surfaces with reference to the agent (Gibson, 1979, p. 19). Then, he addressed the relation between such a layout and the ambient optic array; namely, he described ecological optics. And finally, he addressed the way information is detected by perceptual systems to control behavior. This final step was mostly developed in a later chapter of The Ecological Approach to Visual Perception (1979) entitled “Locomotion and Manipulation” and constituted the last theoretical element needed to have a new law-based explanatory strategy in psychology.

During his career, and especially in the 1970s, Gibson was exposed to Phenomenology and, more concretely, to the work of Maurice Merleau-Ponty (1942/1963, 1945/2012). ¹⁰ In The Structure of Behavior (1942/1963), Merleau-Ponty famously stated that behavior is a Gestalt (i.e., a form), meaning that behavior is a total process whose “properties are not the sum of those which the isolated parts would possess” (pp. 45–50; see also pp. 130–131). In this sense, behavior is a self-organized functional activity whose features are very similar to those Gibson attributed to perceptual systems. ¹¹ Such a characterization of the phenomenon—which is analogous to Holt’s (1915b) conception of behavior as a coordinated totality—permitted Gibson to put the information of the ambient optic array and behavior itself in a lawful relation. In a famous passage in “Locomotion and Manipulation,” he stated:

Locomotion and manipulation are neither triggered nor commanded but controlled … And they are controlled not by the brain but by information, that is, by seeing oneself in the world. Control lies in the animal–environment system … The rules that govern behavior are not like laws enforced by an authority or decisions made by a commander; behavior is regular without being regulated [emphasis added]. The question is how this can be. (Gibson, 1979, p. 215)

First, Gibson claimed that behavior—in this case locomotion and manipulation—is not controlled by the brain, but by information. Information is about the animal–environment system; that is, information is about the environment the animal faces but also about the animal’s own states within such environment. In this sense, control is an ecological phenomenon that must be explained on the ecological scale—i.e., control is a phenomenon that occurs at the scale of the relations between animals and environments and so it has to be explained. And second, behavior is regular without being regulated. That is, behavior is not the product of a brain controlling all the motor components of the body of the animal, but a self-organized functional activity that is controlled by information at the ecological scale. The rules or laws that govern behavior are not enforced by a central controller (e.g., the brain) but they are ecological laws. Ecological laws capture the regularities in the relationship between behavior and the information of the environment (e.g., information available in the ambient optic array) without the need for postulating sources of internal control (Turvey et al., 1981).

The explanatory strategy of ecological psychology is, thus, a law-based explanatory strategy that searches for the lawful relations between perception and action or, more concretely, between the perceptual activity of picking-up environmental information and the behavioral activities that both enable and are controlled by such a perceptual activity. The activity of picking up environmental information is played out in terms of the detection of invariant properties of the organism–environment relation (e.g., patterns of optic flow) that are specific to the affordances ¹² of the environment and of the behavior of the organism (e.g., a centrifugal pattern in the optic flow is specific of forward locomotion). ¹³ Thus, the control of action (behavioral activities; e.g., forward locomotion) is lawfully related to perception (information pick-up; e.g., centrifugal optic flow) and the task of psychology as a science is the discovery of these lawful relations.

With ecological psychology, Gibson finally found the framework for a law-based account of perception and action he was after during his whole academic life. As we have seen, he was convinced that, in order to really explain a phenomenon, scientists should be able to provide laws that capture causal regularities at the proper phenomenal scale. During his life, he became more and more convinced that classic theories of perception were unable to do so and, for that reason, started developing his own theory. In order to achieve a law-based theory of perception at the proper perceptual scale, first he had to reject classic psychophysics. Later, he had to give up his own perceptual psychophysics (Gibson, 1950) and entirely redefine the field of physiology (Gibson, 1966). And, finally, he had to propose an ecological approach to perception that accounted for the relations between perceivers and their environments in terms of the lawful regularities of perception–action couplings (Gibson, 1979). Gibson criticized classic concepts of psychology like the retinal image or the stimulus–response formula and proposed new concepts and fields of inquiry such as affordances or ecological optics. He also changed his mind regarding his own theory of perception several times. However, he maintained an idea at the roots of his research: that to explain, to do science, entails the knowing of laws, so we need a law-based explanatory strategy. Such is the strategy of ecological psychology and such is the strategy that permits us to understand its historical development. Such is Gibson’s most radical idea.