On 2015 Sep 03, Lydia Maniatis commented:

It seems generally to be the case that when a darkish-looking surface grows in area, it appears to lighten. There is no doubt that this occurs, and also no doubt that the effect is subject to a large degree of variability between experiments and between individuals. Boyaci et al (2014) report, as their main finding, that the effect occurs in the context of "computer-rendered scenes" (their stimuli) as well as in the context of "real scenes" (referring to the rarefied laboratory set-up of Radonjic and Gilchrist (2014)).

As a way of describing experimental conditions - whose results, moreover, were subjected to technical analyses to determine which datasets fit a linear model, a quadratic model, a one-phase model, a two-phase model, and used to form hypotheses about the neurophysiological underpinnings of performance and to judge conceptual models - the crude distinction between "computer-rendered" and "real" seems rather insufficient. Indeed, when it comes anything more specific than the general consistency with the "larger gets lighter" effect, all bets are off. Noting that their results were different from those of Radonjic and Gilchrist (2014), the authors speculate simply that: "The disagreement between our studies is likely to be because of differences between the stimuli." More specifically: "Whereas their stimuli (Radonji砦 Gilchrist, 2014) were real and as simple as possible, ours were computer generated and relatively more complex." Like the unqualifed "real" vs "computer-generated" distinction, the "simple vs complex" distinction is too vague for the purpose. Very subtle and neurophysically complex perceptual operations can be triggered by geometrically simple stimuli. Surely, mathematical modelling and theoretical/neurophysiological extrapolations should follow, not precede, qualitative understanding of a phenomenon and the effects of conditions. Otherwise, such speculation is strictly ad hoc.

The "area rule" as originally proposed said more than that the darker of two areas will lighten with area. It said that this lightening of the darker area will "anchor" cause the appearance of the lighter area such that it brightens and eventually becomes luminous. The proposal was made in order to rationalize the luminosity observed in the disc in classic disc/annulus experiments, without having to ascribe a role to figure-ground organization. The predicted changes in the lightness of the lighter area qua area apparently haven't been corroborated and are no longer referred to or tested for. As an isolated phenomenon not affecting surrounding surfaces, it's hard to see the theoretical importance of a highly-variable and condition-sensitive tendency toward lightening of relatively darker surface with increases in area, or the value of making fussy models of this tendency, tailored to particular stimuli.

I would also like to take issue with the choice of references the authors chose to support their opening assertion that "The lightness of a surface depends not only on its luminance but also on its geometry and the context within which it is viewed (Boyaci, Doerschner, Snyder, & Maloney, 2006; Gilchrist, 2006; Kingdom, 2011; Maloney, Boyaci, & Doerschner, 2005; Maloney, Gerhard, Boyaci, & Doerschner, 2010)." The case for this was made in classic literature and experiments well before 2006. The assertion is so fundamental and uncontroversial that references are not even necessary.

On 2015 Sep 03, Lydia Maniatis commented:

It seems generally to be the case that when a darkish-looking surface grows in area, it appears to lighten. There is no doubt that this occurs, and also no doubt that the effect is subject to a large degree of variability between experiments and between individuals. Boyaci et al (2014) report, as their main finding, that the effect occurs in the context of "computer-rendered scenes" (their stimuli) as well as in the context of "real scenes" (referring to the rarefied laboratory set-up of Radonjic and Gilchrist (2014)).

As a way of describing experimental conditions - whose results, moreover, were subjected to technical analyses to determine which datasets fit a linear model, a quadratic model, a one-phase model, a two-phase model, and used to form hypotheses about the neurophysiological underpinnings of performance and to judge conceptual models - the crude distinction between "computer-rendered" and "real" seems rather insufficient. Indeed, when it comes anything more specific than the general consistency with the "larger gets lighter" effect, all bets are off. Noting that their results were different from those of Radonjic and Gilchrist (2014), the authors speculate simply that: "The disagreement between our studies is likely to be because of differences between the stimuli." More specifically: "Whereas their stimuli (Radonji砦 Gilchrist, 2014) were real and as simple as possible, ours were computer generated and relatively more complex." Like the unqualifed "real" vs "computer-generated" distinction, the "simple vs complex" distinction is too vague for the purpose. Very subtle and neurophysically complex perceptual operations can be triggered by geometrically simple stimuli. Surely, mathematical modelling and theoretical/neurophysiological extrapolations should follow, not precede, qualitative understanding of a phenomenon and the effects of conditions. Otherwise, such speculation is strictly ad hoc.

The "area rule" as originally proposed said more than that the darker of two areas will lighten with area. It said that this lightening of the darker area will "anchor" cause the appearance of the lighter area such that it brightens and eventually becomes luminous. The proposal was made in order to rationalize the luminosity observed in the disc in classic disc/annulus experiments, without having to ascribe a role to figure-ground organization. The predicted changes in the lightness of the lighter area qua area apparently haven't been corroborated and are no longer referred to or tested for. As an isolated phenomenon not affecting surrounding surfaces, it's hard to see the theoretical importance of a highly-variable and condition-sensitive tendency toward lightening of relatively darker surface with increases in area, or the value of making fussy models of this tendency, tailored to particular stimuli.

I would also like to take issue with the choice of references the authors chose to support their opening assertion that "The lightness of a surface depends not only on its luminance but also on its geometry and the context within which it is viewed (Boyaci, Doerschner, Snyder, & Maloney, 2006; Gilchrist, 2006; Kingdom, 2011; Maloney, Boyaci, & Doerschner, 2005; Maloney, Gerhard, Boyaci, & Doerschner, 2010)." The case for this was made in classic literature and experiments well before 2006. The assertion is so fundamental and uncontroversial that references are not even necessary.