On 2016 Jun 03, Lydia Maniatis commented:

The authors of this article perform a useful service: They experimentally falsify a claim that currently enjoys some favor. Falsifying (false) assumptions is a fundamental activity in science; a science incapable of rejecting false views is incapable of selecting true(er) ones, and thus incapable of progress.

This simple fact is so antithetical to the current “don't criticize” ethos in vision science that the authors have (either by choice or editorial demand) been forced to bury the lead and accentuate the mildly positive and/or vague. Thus the title of the article is vague, implying nothing and the abstract reports a mixed salad of findings. For a statement of any conclusions we're directed to the article itself.

From the intro, we learn that Christou and Koenderink (1997) have speculated that the presence of smooth occlusions is a key factor in seeing “shape-from-shading.” Amazingly, in the nearly 20 years since, no one – including its originators - has actually tested this claim. After explaining what a “reasonable” suggestion this is, Todd and Egan note that “virtually all past studies have included visible smooth contours.” So they propose to put this suggestion to a real test, but aren't allowed to state it this strongly so near the beginning of the article; they are merely going to “further examine” the issue.

The discussion also drags its feet in getting to the point, since the first thing we learn is that “One of the basic findings of the present investigation is that observers’ perceptions of 3-D shape from shading are sheared slightly toward the direction of illumination. This phenomenon was first discovered by Koenderink et al. (1996a, 1996b) and Christou and Koenderink (1997).” (The “phenomenon” is context-specific, since perceptions of 3D shape will certainly not always be sheared in the direction of illumination – it will depend on the shapes. All that can be said is that this sometimes happens).

Finally, we get to the point: “[Christou & Koenderink, 1997] speculated that the [high degree of] constancy was most likely due to information provided by visible smooth occlusion contours. The present investigation WAS DESIGNED SPECIFICALLY TO TEST THAT SUGGESTION, and the RESULTS SHOW CLEARLY that the absence of smooth occlusions in the masked conditions had a NEGLIGIBLE IMPACT [all caps mine] on the overall pattern of performance.”

So Christou and Koenderink's purportedly “reasonable” suggestion is apparently false. That's the point, and the authors shouldn't have been forced (or internalized a reequirement) to pussyfoot around it. They should have stated it as clearly up front as they do (eventually) in their discussion.

Falsification of FALSE theories is how science proceeds, and you can't show they're false unless you actually target specific assumptions for testing. Of course, proponents of the idea may, if they'd like, and can, make counterarguments.

So kudos to the authors for directly testing the claim, as they did in a previous article on an another ad hoc suggestion. Next time, I hope they adopt a more direct reporting style as well.

But on theory, there's more to be said, starting with https://pubpeer.com/publications/1220D804D501BB15474B6D0F33FC7D

On 2016 Jun 03, Lydia Maniatis commented:

The authors of this article perform a useful service: They experimentally falsify a claim that currently enjoys some favor. Falsifying (false) assumptions is a fundamental activity in science; a science incapable of rejecting false views is incapable of selecting true(er) ones, and thus incapable of progress.

This simple fact is so antithetical to the current “don't criticize” ethos in vision science that the authors have (either by choice or editorial demand) been forced to bury the lead and accentuate the mildly positive and/or vague. Thus the title of the article is vague, implying nothing and the abstract reports a mixed salad of findings. For a statement of any conclusions we're directed to the article itself.

From the intro, we learn that Christou and Koenderink (1997) have speculated that the presence of smooth occlusions is a key factor in seeing “shape-from-shading.” Amazingly, in the nearly 20 years since, no one – including its originators - has actually tested this claim. After explaining what a “reasonable” suggestion this is, Todd and Egan note that “virtually all past studies have included visible smooth contours.” So they propose to put this suggestion to a real test, but aren't allowed to state it this strongly so near the beginning of the article; they are merely going to “further examine” the issue.

The discussion also drags its feet in getting to the point, since the first thing we learn is that “One of the basic findings of the present investigation is that observers’ perceptions of 3-D shape from shading are sheared slightly toward the direction of illumination. This phenomenon was first discovered by Koenderink et al. (1996a, 1996b) and Christou and Koenderink (1997).” (The “phenomenon” is context-specific, since perceptions of 3D shape will certainly not always be sheared in the direction of illumination – it will depend on the shapes. All that can be said is that this sometimes happens).

Finally, we get to the point: “[Christou & Koenderink, 1997] speculated that the [high degree of] constancy was most likely due to information provided by visible smooth occlusion contours. The present investigation WAS DESIGNED SPECIFICALLY TO TEST THAT SUGGESTION, and the RESULTS SHOW CLEARLY that the absence of smooth occlusions in the masked conditions had a NEGLIGIBLE IMPACT [all caps mine] on the overall pattern of performance.”

So Christou and Koenderink's purportedly “reasonable” suggestion is apparently false. That's the point, and the authors shouldn't have been forced (or internalized a reequirement) to pussyfoot around it. They should have stated it as clearly up front as they do (eventually) in their discussion.

Falsification of FALSE theories is how science proceeds, and you can't show they're false unless you actually target specific assumptions for testing. Of course, proponents of the idea may, if they'd like, and can, make counterarguments.

So kudos to the authors for directly testing the claim, as they did in a previous article on an another ad hoc suggestion. Next time, I hope they adopt a more direct reporting style as well.

But on theory, there's more to be said, starting with https://pubpeer.com/publications/1220D804D501BB15474B6D0F33FC7D