On 2016 Dec 12, Lydia Maniatis commented:

It’s amazing how shy Agaoglu and Chung (2017) are in making their legitimate point, and how gently and casually they make it when they finally do get there (even giving credit where credit is not due). We can’t be sure, until we get to the very end (of the article not the abstract), what the authors think about the question posed, somewhat awkwardly, in their title: “Can (should) theories of crowding be unified?” The title begs the question, if they can be unified, then why shouldn’t they be? The explanation for the awkwardness, also saved for the end, is that the term “crowding” is conceptually vague, and has been used as a catch-all term for all manner of demonstrations of the fact that peripheral vision is not as good as central vision. “All in all, although we applaud the attempts at unifying various types of response errors in crowding studies, we think that without a better taxonomy of crowding—instead of calling everything crowding, perhaps introducing types of crowding (as in the masking literature)—unifying attempts will remain unsuccessful.” In other words, the issue of a unifying “theory of crowding” is moot given that we’re talking about a hodge-podge of poorly-understood phenomena.

What is also moot are the experiments reported in the article. While I think they have a lot of problems common to the field (layers of untested/untestable or even false, arbitrarily chosen assumptions), that doesn’t matter. As should be clear from the authors’ own discussion, no new experiments were needed to make the necessary theoretical point, which is that a clear conceptual understanding of phenomena needs to precede any attempt at a technical, causal explanation. Clearly, the experiments added by the authors do not make more acute the stated need for “a better taxonomy of crowding.” The redundancy of the study is reflected in the statement (caps mine) that “Our empirical data and modeling results ALSO [i.e. in addition to previous existing evidence] suggest that crowded percepts cannot be fully accounted for by a single mechanism (or model).” They continue to say that “The part of the problem is that many seemingly similar but mechanistically different phenomena tend to be categorized under the same umbrella in an effort to organize the knowledge in the field. Therefore, constraints for theoretical models become inflated.”

Agaoglu and Chung’s (2016) message, in short, is that the heterogeneous class of phenomena/conditions referred to as “crowding” are clearly not candidates for a common explanation or “model” as they routinely produce mutually conflicting experimental outcomes (one model works for this one but not that one, etc) and that there is a need for investigators to clarify more precisely what they are talking about when they use the term crowding. While, as mentioned, they could surely have made their argument without new experiments, they couldn’t have published it, as the rational argument in science has unfortunately been demoted and degraded in favor of uninterpretable, un-unifyable, premature p-values.

On 2016 Dec 12, Lydia Maniatis commented:

It’s amazing how shy Agaoglu and Chung (2017) are in making their legitimate point, and how gently and casually they make it when they finally do get there (even giving credit where credit is not due). We can’t be sure, until we get to the very end (of the article not the abstract), what the authors think about the question posed, somewhat awkwardly, in their title: “Can (should) theories of crowding be unified?” The title begs the question, if they can be unified, then why shouldn’t they be? The explanation for the awkwardness, also saved for the end, is that the term “crowding” is conceptually vague, and has been used as a catch-all term for all manner of demonstrations of the fact that peripheral vision is not as good as central vision. “All in all, although we applaud the attempts at unifying various types of response errors in crowding studies, we think that without a better taxonomy of crowding—instead of calling everything crowding, perhaps introducing types of crowding (as in the masking literature)—unifying attempts will remain unsuccessful.” In other words, the issue of a unifying “theory of crowding” is moot given that we’re talking about a hodge-podge of poorly-understood phenomena.

What is also moot are the experiments reported in the article. While I think they have a lot of problems common to the field (layers of untested/untestable or even false, arbitrarily chosen assumptions), that doesn’t matter. As should be clear from the authors’ own discussion, no new experiments were needed to make the necessary theoretical point, which is that a clear conceptual understanding of phenomena needs to precede any attempt at a technical, causal explanation. Clearly, the experiments added by the authors do not make more acute the stated need for “a better taxonomy of crowding.” The redundancy of the study is reflected in the statement (caps mine) that “Our empirical data and modeling results ALSO [i.e. in addition to previous existing evidence] suggest that crowded percepts cannot be fully accounted for by a single mechanism (or model).” They continue to say that “The part of the problem is that many seemingly similar but mechanistically different phenomena tend to be categorized under the same umbrella in an effort to organize the knowledge in the field. Therefore, constraints for theoretical models become inflated.”

Agaoglu and Chung’s (2016) message, in short, is that the heterogeneous class of phenomena/conditions referred to as “crowding” are clearly not candidates for a common explanation or “model” as they routinely produce mutually conflicting experimental outcomes (one model works for this one but not that one, etc) and that there is a need for investigators to clarify more precisely what they are talking about when they use the term crowding. While, as mentioned, they could surely have made their argument without new experiments, they couldn’t have published it, as the rational argument in science has unfortunately been demoted and degraded in favor of uninterpretable, un-unifyable, premature p-values.