The exact wording reads correctly "As soon as the global triangular for is recognized, the low level visual edges...appear at high [not higher] resolution..."

Now here is a point where a perceptual model of Gestalt phenomena can be validly challenged. For nobody can step inside the reviewer's head to determine whether or not he sees a high resolution perimeter around the camo triangle in figure 1 (a). All I can say is that I know that I can see that perimeter very clearly, although I perceive it as an amodal percept devoid of any explicit surface brightness component. (As with all illusory contours, including modal ones, the illusion tends to vanish under intense scrutiny. This does not however negate its perceptual reality under the conditions when it is clearly perceptible)

This dispute may simply reduce to the fact that the reviewer considers amodal contours to be "invisible", i.e. if he cannot see a modal contrast edge, then he reports seeing nothing at all. If that is his quibble, my answer is that this is a question of definitions. If we define the amodal contour as the kind of contour that is perceived in figure 1 (a) and figure 2 (a) and (c, upper transition), then, by definition, that contour is perceptible in those figures.

If on the other hand the reviewer denies seeing any contour-like percept in those figures, then maybe his brain operates differently than mine.

Now as to the question of whether modal illusory contours are indistinguishable from actual edges and surfaces, it has been shown that under closer scrutiny those edges can disappear, and therefore they are not completely indistinguishable from actual luminance edges. However on brief exposure one could easily be fooled into thinking that there was an actual brightness difference, especially if the task was to distinguish genuine Kanizsa stimuli from artificially constructed ones which actually possess a real luminance difference in the stimulus.

In any case, I maintain that it remains true that modal illusory contours are virtually indistinguishable from real contours, in that they are defined by a perceived brightness contrast across a visual edge, the very property that characterizes the perception of a real luminance edge. If the reviewer continues to quibble, perhaps he will at least concede that an image like figure 1 (c) is similar to the subjective experience of figure 1 (b), at least in terms of information content, and therefore this is the closest that a perceptual model can possibly come to representing that subjective experience.