Koffka on the Phenomenological Questions of Perception Science

Koffka (1955) claims that the psychological theory of perception has to address the question “why do things look as they do?” This question epitomizes the phenomenological commitment of experimental research. Koffka maintains that a valid theory of perception should answer this question to be true to perceptual experience, and is thus committed to finding adequate theoretical constructs and a suitable level of analysis of perception. Phenomenology pre­vents unproven commonsensical, theoretical, scientific assumptions to bias the design of experimental conditions and the interpretation of evidence. He rejects two strategies to answer the question. The first strategy suggests that things look as they do because of the material properties they have, hence it reduces perceiving things to recording physical distal properties. The sec­ond strategy suggests that things look as they do because of the properties of the stimulation they bring about on nerve endings, hence it reduces perceiv­ing things to taking notice of the physiological proximal stimuli, from which the likelier seeming of external things has to be reconstructed. Both answers resort to explanatory constructs that do not fit the phenomenal scale at which subjects use perception to understand the environment, to recognize things successfully and to act effectively. Koffka points out that counterexamples to both strategies exist.

According to the first strategy, being physically a unitary object is the nec­essary and sufficient condition for it to occur as perceptual unity. It is yet an observable fact either that perceptual unities may not correspond to the physi­cal objects that are believed to exist, or that the material boundaries of things may not correspond to perceptual ones, as in masking and camouflage, which alter the phenomenal connection of the boundaries of things. According to the second strategy, the stimulations are the necessary condition for sensa­tions to occur. Their synthesis into constellations of sensations is the sufficient condition for giving rise to perceptual unities that reconstruct the external things, assuming that the constancy of single sensations with respect to physi­cal causes is preserved in the whole constellation. Yet a white surface keeps on appearing white for a long while, even if the illumination is reduced to a very low intensity. Besides, if the constancy assumption cannot be satisfied, this strategy requires introducing ad hoc hypotheses on the synthesis of sensa­tions to explain that the reconstruction of physical objects is still obtained. For example, a surface that appears white at low illumination stimulates retinal cells far less than a surface that appears black at high illumination. The obser­vation that the surfaces still look like white and black is explained by figuring that the insufficient sensory base is offset by associative learning, memory of past experiences or high-level psychological functions. Koffka argues that this kind of hypothesis cannot increase the probability that the explanation is cor­rect. These explanatory constructs are not observable and their functional con­nection with the phenomena and environmental conditions derives from the theory on which they depend.

Koffka (1924) claims that the starting point of the psychology of perception is the environment that embeds the reactions of organisms. The environment is the phenomenal world, that is, the collection of the appearances of the ex­ternal world, which build the surroundings of organisms, hence the reactions of the organisms include their consciousness. The perceptual knowledge that the organism has of herself and the appearances of the environment are the proper “facts of psychology.” The description of these facts is essential to answer the phenomenological question. Koffka (1955) claims that the theory has to abstract the data and constructs from the description of direct experience. At this scale the perception is tuned to the environment and the appearances fix the reference to what is displayed as the reality to which the behaviour adapts. Koffka (1924: 155) acknowledges that any phenomenal reaction of the organ­ism to the environment is studied only through the theoretical constructs and the experimental devices. However, he claims that “the original reactions are to be studied just as they are and not merely under the aspect of what they will become when the analysis is applied to them. Only in this way shall we able to find their proper laws.” Therefore the phenomenological description com­mits the concepts and methods that construe the phenomena as the data of a scientific theory to fit the nature of appearances.

Koffka (1921) clarifies the sense in which this description is neither a triv­ial task nor differs from the customary practice of science. Just as any other science, the psychology of perception has to build its theoretical constructs. Constructs such as the wavelength of electromagnetic radiations and the prod­uct of mass times acceleration are hardly recognized by the layman as the same as colors and forces. The world of physics does not resemble the common- sense external world. This is also true of the methods of physics, whose instru­ments do not occur in nature, like the cathode ray tube, the Atwood machine or the Nicol prism. Nonetheless, this does not mean that physics is aimless or arbitrary. The instruments obey the same laws as the objects of nature; hence the phenomena observed through them become substitutes for the objects of nature. The experimental observations can be extended and the derived laws provide the explanation of nature (1921: 392). Like physics, the science of perception has to build its autonomous concepts and the methods through which the phenomenal data are abstracted as substitutes for the appearanc­es of experience. If this substitution is properly carried out, it allows for the discovery of previously neglected or discarded features, for instance the vo­calic quality of tones (infra 4.1). Indeed, the current physical knowledge of the properties of waves and musical practice have limited the research to pitch and loudness. Yet the phenomena studied on this basis are poor substitutes for tone appearances, while the vocalic character has to figure in the description and explanation of sound perception.

Koffka considers the substitution of appearances with phenomenal data as a conceptual transformation of appearances according to the method of science to make them observable and testable in controlled conditions. The substitution has to satisfy two demands. Firstly, it has to be an explanatory function for experience. For example, the view that one has of a diesel engine, if one does not know this kind of engine, cannot have the same value as the view after having acquired the knowledge of how it works. The appearance of the engine changes from the first to the second view. The latter conveys the perception of the engine components and the right connections thereof. It is a more suitable substitute for the appearance of the engine, for it contributes to understanding the engine functions more than the first, in which the engine looks like an unconnected heap of mechanical elements that are identified only through their geometrical forms (1921: 393). Secondly, the substitution has to be a description that is true to the nature of objects. In this sense, being more suitable means being more adequate to the structure of the object (1921: 394). The modes of appearances of something do not have the same value, and this difference relies on their inherent characteristics. The second view of the engine is an example of a “good phenomenon” because it permits one to see their parts and order. In general, good phenomena correspond to a com­plete description of objects that spells out the connection of their parts. Other phenomena may be more or less adequate according to the extent to which they deviate from this ideal description of the nature of objects towards the opposite ends of seeing in the object either a disconnected heap or a tangled web of pieces.

It may happen that the concepts and constructs do not remain true to the describable characteristics of perception and are applied beyond the bounds of the explanation for which they were designed. Koffka (1921: 389­390) holds that this is the case with the concept of sensation. If the constant physical properties had not been distinguished from the variable properties of sensations, the psychology of perception could not have been established as an autonomous science. The concept of sensation has allowed psycholo­gy to substitute the common-sense objects of daily life experience with the phenomena of sensory variability. The world considered under the respect of perception was then decomposed into the physical constituents and the sensory components, and the laws of sensations constituted the domain of psychology.

However, the concept of sensation became detrimental for the psychology of perception when it was extended beyond the limits of the explanation for which it was introduced, and was given the character of permanence that had been attributed to physical objects. The perceptual domain was constructed as the flux of sensations, so that the psychology of perception aimed at dis­covering the ultimate elements of sensations under the assumption that these elements are specified by the univocal, permanent and stable physiological correspondence with the stimulations of the sensory organs. Sensations have been intended as mental contents with the same constancy as physical objects but accessible through introspection. Koffka emphasizes that the ill-founded extension of the concept implies an inadequate substitution for perceptual ap­pearances with constructs lacking a descriptive foundation. If appearances are construed as subjective phenomenal contents, even the evident distinction be­tween the subjective and the objective side of direct experience is removed, so that “it seems impossible ever to explain in a satisfactory manner the objective character of our perception” (1924: 151b).

The complex requirement to trade the description off against the introduc­tion of new concepts shows that Koffka considers phenomenology a method of observing the facts of direct experience without a theoretical or scientific bias alien to the autonomous science of perception. The phenomenology of perception provides the research with “as naive and full description of di­rect experience as possible,” without which “we should not know what we had to explain” (1955: 73). The phenomenological description complements the experimental research. It abstracts the features of appearances that have to be transformed in the phenomenal data for the experimental investiga­tion and identifies which parts and connection of phenomena must be cap­tured by the theory. Therefore, the phenomenology clarifies the meaning of the explanatory constructs and may be useful to rule out alternative compet­ing theories.

Source: Calì Carmelo (2017), Phenomenology of Perception: Theories and Experimental Evidence, Brill.