“Finding out” about the visual world as approached from the organismic level may well include the “filling-in” type of perceptual completion if considered in terms of underlying neurophysiological mechanisms. But “filling in” can be interpreted not only as a result of within-level propagating of neural activity, but as a byproduct of the process that is necessary for modulating preconscious information about physically present objects or events so as to generate conscious quality in attending to them.

Schyns et al. argue that flexibility in categorisation implies “feature creation.” We argue that this notion is flawed, that flexibility can be explained by combinations over fixed feature sets, and that feature creation would in any case fail to explain categorisation. We suggest that flexibility in categorisation is due to pragmatic factors influencing feature combination, rendering feature creation unnecessary.

Chow sets his version of statistical significance testing in an impoverished context of “theory corroboration” that explicitly excludes well-posed theories admitting of strong support by precise empirical evidence. He demonstrates no scientific usefulness for the problematic procedure he recommends instead. The important role played by significance testing in today's behavioral and brain sciences is wholly inconsistent with the rhetoric he would enforce.

The precise relationship between callosal morphology and specific connectivity is not yet known. Callosal axons are often presumed to be arranged according to their origin. In humans, this is true for the genu and the splenium, which convey axons from the prefrontal and occipital cortices, respectively, but not for the body, where axons from wide parts of the cortex are intermingled.

Van Gelder's example of a dynamical model is a Perceptron. The similarity of dynamical models and Perceptrons in turn exemplifies the close relationship between dynamical and algorithmic models. Both are models, not literal descriptions of brains. The brain states of standard modeling are better conceived as processes in the dynamical sense, but algorithmic models remain useful.

Prediction, like filling-in, is an example of pattern completion and both are likely to involve processes of statistical inference. Furthermore, there is no incompatibility between inference and neural filling-in, for the neural processes may be mediating the inferential processes. The usefulness of the “bridge locus” is defended, and it is also suggested that the interpersonal level needs to be included when considering subjective experience.

Under shift, caused for example by eye movement, or by relative movement of the subject or object of perception, the cortical representation undergoes very large changes in “size” and “shape.” Space-variance of cortical representation rules out models that fundamentally require linear interpolation between shifted patterns (e.g., Edelman's model) or rigid shift of an invariant retinal stimulus corresponding to shift at the cortex (e.g., the shifter theory of van Essen). Recently, a computational solution of “quasi-shift” invariance for space-variant mappings has been constructed (Bonmassar & Schwartz 1997a; 1997b).

Evidence accumulates that pre- and postnatally circulating estrogens play an active role in the differentiation of the female brain: the susceptible period for feminization of the brain seems to extend far beyond the period during which masculinization of the brain occurs. Thus, there is a need to reevaluate the widely accepted “concept of basic femaleness” in sexual brain differentiation.

Chow's defense of NHSTP ignores the fact that in psychology it is used to test substantive hypotheses in theory-corroborating research. In this role, NHSTP is not only inadequate, but damaging to the progress of psychology as a science. NHSTP does not fulfill the Popperian requirement that theories be tested severely. It also encourages nonspecific predictions and feeble theoretical formulations.

Medical education provides many examples of the development of functional features, but as a response to deliberate instruction. These features require so much specificity and context sensitivity that they seem likely to require the development of new categories of appearances rather than just reweighting old features. A suggested implication is that feature development may help to explain the problematic noticing of features in diagnosis.

New knowledge about the sexual differentiation of the brain profoundly changes our understanding of basic topics in brain development such as the false dichotomy between long-lasting and transient effects of hormones on neural activity, the importance of ovarian hormones in brain development, the plasticity of neural structures throughout the life span, and the way measurement issues affect research conclusions.

Although we find the idea of representation by similarities attractive as such, we have two main objections to the specific proposal of Edelman. First, he does not consider complexity issues in terms of storage and speed of recall for recognition. Related to this, the appearance of objects depends on far more factors than just shape, illumination, and pose. This requires an intermediate shape abstraction process that extracts category-specific shape properties from the mixed appearance of images.

Pessoa et al. misinterpret some of Dennett's discussion of filling-in. Their argument against the representational conception of vision and for a Gibsonian alternative is also flawed.

The concept of virtual machine allows us to combine the dynamical and computational hypotheses in an investigation of cognition. Van Gelder explicitly rejects this approach, but not only does it allow us to use the modelling technique most appropriate to the task, it also opens up a new range of phenomena where these techniques interact.

Implementational-level evidence of dissociable neural subsystems is a critical element that is missing from the analysis in the Schyns et al. target article. The question of whether fixed or flexible features are used in visual form recognition may have different answers for different subsystems; the evidence that features typically are created during category learning may apply most directly to a specific visual form subsystem.

In this commentary, I agree with Chow's treatment of null hypothesis significance testing as a noninferential procedure. However, I dispute his reconstruction of the logic of theory corroboration. I also challenge recent criticisms of NHSTP based on power analysis and meta-analysis.

Representing objects and concepts as points in low-dimensional shape space defined by distances to other complete object exemplars or prototypes, expressed as single numbers, misses the key advantages of representation in terms of hierarchically constructed, meaningful features of the environment. Generalisation along statistically significant, near-independent, sparse, cooperative features that stand directly for various aspects of a concept is essential.

Fitch & Denenberg have shown that manipulations of ovarian and testicular hormones early in development can influence the adult size of the corpus callosum in the rat. The human corpus callosum is highly variable in size and shape, but data are only now beginning to emerge on whether sex steroids influence callosal differentiation in humans. I describe recent data from our own laboratory and suggest avenues for future research.

Van Gelder's specification of the dynamical hypothesis does not improve on previous notions. All three key attributes of dynamical systems apply to Turing machines and are hence too general. However, when a more restricted definition of a dynamical system is adopted, it becomes clear that the dynamical hypothesis is too underspecified to constitute an interesting cognitive claim.

I disagree with several of Chow's traditional descriptions and justifications of null hypothesis testing: (1) accepting the null hypothesis whenever p > .05; (2) random sampling from a population; (3) the frequentist interpretation of probability; (4) having the null hypothesis generate both a probability distribution and a complement of the desired conclusion; (5) assuming that researchers must fix their sample size before performing their study.