Phenomenal Structures and Comparative Judgements in Phenomenology

Koffka (1917, 1922) provides a wealth of arguments to support Kohler’s phe­nomenological point. He provides an analysis of the perceptual judgements in the experimental setting and the structure of phenomenal data. If the task is the comparison between couples of colors, sounds and spatial elements, the constancy hypothesis assumes that there must be a single independent sensa­tion for each color, tone or line on which subjects are asked to make a compari­son under a particular respect. If subjects’ reports deviate from the expected one-to-one correspondence to the stimulation, two strategies are usually un­dertaken. The first is to posit unobserved additional sensations that interfere with sensing the physical properties of the stimuli. The second is to posit a threshold for a psychic function, like judgement or attention, that processes or lets the sensory material appear, under which the sensations that correspond to the stimulation are bound to remain unnoticed. For both strategies, the dif­ference from the predictions of the constancy hypothesis is interpreted not as an observation that falsifies the hypothesis, thus requiring a more comprehen­sive descriptive theory of perception, but rather as the alteration of the one- to-one correspondence due to sensory or psychic interference.

Koffka deals mainly with the second strategy, for it derives from an erro­neous interpretation of a key concept of psychology, namely the differential threshold. Koffka (1917: 1) reminds us that this concept has an equivalent in physics. Suppose that a 1 mg + 10^-n weight is put on a balance scale, where n is arbitrarily small and 1 mg is a quantity that is sufficient to make the scale tip.

If n is smaller than the observational error, the weight difference is not record­ed. Likewise the concept of a differential threshold implies that there must be a just noticeable difference between two stimuli to bring about two different sensations. Yet an unsound extension of this analogy may lead to flawed con­clusions. In general, the concept of the psychophysical differential threshold has no explanatory value if it is used to maintain that there is an unnoticed difference between sensations if the difference between the physical proper­ties of stimuli is under the noticing threshold. This interpretation brings in the inconsistent notion of a not sensed difference between sensations, which can­not provide an explanation, since it is equivalent to saying that something is not noticed if it is under the threshold for noticing something. In particular, it does not make sense to argue that if the predicted sensations are not reported, this is not an observation against the constancy hypothesis because the sen­sations are unnoticed since their effect is too small to surpass the threshold and be taken into account by judgement or attention. Such a claim makes the meaning of any observation undecidable.

In order to show the shortcomings of this interpretation of the constancy hypothesis, Koffka cites a paradox introduced by Stumpf (1883: 33-34). Con­sider the series of tones whose stimuli differ so smoothly in frequency as 430, 430.1, 430.2, 430.3 Hz and so forth to (n – 0.1), n Hz. If subjects are asked to compare the pitch of tone pairs, it is conceivable to pick up three stimuli A, B, C such that the perceptual judgements for the corresponding tones a, b, c do not satisfy the transitivity relation. For small differences in frequency such that A < B, B < C, A < C, the subjects may report three successive comparisons: a = b; b = c; a # c. Koffka claims that this interpretation of the paradox, like the ensuing solutions, is committed to the constancy hypothesis. The first solution is to use the concept of the differential threshold to hold that the difference between a and b corresponding to the frequency of the stimuli is unnoticed until a is compared with c. However, Koffka remarks that if A and B are pre­sented many times, another deviation from the constancy hypothesis may be observed, namely U(R (a, b)) where “U” stands for “it is uncertain that _” and “R” for “has a higher pitch than.” This solution does not distinguish between equality and uncertainty, hence it cannot be generalized. The second solution is to claim that a and b are unnoticed tone sensations and that their effect is so small that the difference falls under the threshold of the psychic function deemed to detect it (in fact, Koffka and Kohler interpreted Stumpf’s theory of perceptual judgements as a sophisticated variant of this solution; infra §7.3.1). Therefore, subjects make the wrong judgement when they compare a with b and b with c, actually mistaking two unnoticed different sensations for equal sensations. Koffka raises the objection that this solution brings an asymmetry to the account. The report that the first two pairs of tones are equal is con­sidered due to a wrong judgement; hence, its observation is dismissed. The report that a * c is considered a true judgement and its observation is valid but solely on non-perceptual grounds. Besides, the threshold regards a psychic function that by definition cannot be observed but in its presumed effects. Moreover, for many presentations of the stimuli, the inequality may appear as either a < c or a > c and this solution cannot account for the latter possible observation.

Finally, Koffka (1917: 3) claims that the paradox disappears if the posits of unnoticed sensations or unobservable psychic functions is given up and the difference, which is reported in the comparison judgements, is accounted as a real perceptual difference due to the fact that the value of tone appearances is dependent on the connection they hold with one another. The stimuli and the observations can be represented as follows:

A<B, B<C, A<C (the stimuli presented in three distinct series 1, 2, 3) ;   (1)

a₁ = b_v b₂ = c₂, a₃ < c₃ (the corresponding connections of appearances)           (2)

where the subscript denotes each appearance in function of the connection with the other, which is set by each comparison. Koffka holds that the paradox arises because the hypothesis is made that a₁ = a₃, b₁ = b₂, c₂ = c₃, for the same sensation must correspond to the same stimulus. Once it is hypothesized in­stead that appearances vary in these three connections, it is more reasonable to talk of three different perceptions of pitch rather than of sensations with differently noticed pitch.

Koffka (1922: 539) accounts for such connections in terms of ordering struc­tures of phenomenal attributes. For instance, if two grey cardboards are laid side by side on a black background, there is no observable reason to maintain that subjects sense single color sensations. They may report perceiving (1) a homogeneous grey oblong surface with a line in the middle that distinguishes it into two parts; (2) a brightness step: left darker than right; (3) the inverse brightness step: left brighter than right; (4) uncertain different colors at the left and the right. These observations point to the structure of phenomenal steps of brightness, whose direction and degree of conspicuousness vary according to the boundary conditions of the experimental setting and the subjects’ tasks. Brightness similarity and differences are perceived as the feature of each struc­ture in which the appearances of the cardboards are variously connected to each other for the given conditions, just as something is a step of a staircase only according to the position it holds in the whole staircase.

Koffka (1922: 540) makes the following analogy. From a physical standpoint a staircase is described as merely consisting of planks placed below and above one another. However, from the standpoint of experience a staircase is de­scribed as consisting of steps with definite properties. The steps are at differ­ent distances and have various heights so as to induce a steep or a smooth ascent. The same holds for brightness perception. There is no grey sensation to which the comparative judgement applies as a supervening psychic function. Rather, there are distinct features of brightness depending on the perceptual structure in which grey appearances are connected. Therefore, appearing brighter or darker is a function of the step that characterizes a particular struc­ture. This implies that a grey may become brighter or darker according to the upward or downward direction of the brightness change in the structure it be­longs to. Furthermore, as the order of the steps in a staircase makes it well bal­anced or poorly balanced, so the order of the cardboards makes the brightness change appear variously, for instance when a middle grey is on the left and a radiant white is on the right. Of course, walking up or down a staircase one can observe that the steps are planks, but that does not imply that “plank” has the same perceptual meaning as “step,” rather only that the planks are physi­cal constituents of the staircase. By analogy, Koffka claims that as regards per­ception, this is the standpoint of physics or physiology, which tries to isolate the responses that the nervous system gives to stimulation. Nonetheless, the perceptual constituents of the staircase are the steps that guide the subjects’ behaviour and consequently psychology cannot replace them with the planks if it is the science of perceptual experience.

As regards phenomenal structures Koffka states that a “pure description […] explains the comparison.” The comparison is not a separate psychic func­tion that takes into account distinct sensations. Grey appearances are not in­dependent of each other, rather they are parts of the phenomenal structure that displays an inner articulation. The appearance of brightness is the feature displayed by this structure, and the judgements can be traced back to the ac­quaintance with this articulation. Indeed, a phenomenal structure shows vari­ous degrees of non-independence of parts. The degree of non-independence explains the change of brightness as a perceptual change, rather than as the alteration of the same sensation due to a psychic function and the different truth-values of the comparative judgements (1922: 546). For example, a pure stepwise structure gives rise to the perception of the “crescendo” or “diminu­endo” of brightness. In this case, the whole structure is forced upon the sub­jects, who are likely to make a judgement on an undivided brightness step in which grey appearances occur solely as connected internal parts. The degree of non-independence may be lower and grey appearances may occur as distinct, relatively independent parts within the structure. In this case, the parts of the structure are its “limiting platforms,” namely they are forced as the step bound­aries upon the subjects, who make a judgement on the relation between them. Therefore, the comparative judgement takes the form of a relational judge­ment, for instance “a is greater than b.” This account predicts that in cases like Stumpf’s paradox subjects may report

1. perceiving a uniform or homogeneous structure when presented with two subliminally different stimuli like A and B as naive subjects;
2. being uncertain whether they perceive a uniform structure or a labile structure with an internal distinguishing feature when presented with two subliminally different stimuli like A and B as experimental subjects who have to accomplish a task that demands they look for even the finest perceptual difference;
3. perceiving the step of a feature, like pitch or brightness, when present­ed with two supraliminally different stimuli like A and C whose magni­tude and direction of difference depends on the occurring phenomenal structure.

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