Hall-of-Fame baseball pitcher Pedro Martinez's experience can be explained by the action-specific account of perception. According to this account, people see the distance to or size of objects relative to their ability to act on these objects. At issue is whether supporting empirical findings reflect genuine effects on perception, or instead are a result of one of the six pitfalls Firestone & Scholl (F&S) outline. Fortunately, their claim that these issues have been “largely neglected” (sect. 4.4, para. 2) does not account for much empirical evidence directly addressing the issue with respect to action.

Their claim that no top-down effects on perception exist can be felled with the demonstration that one effect survives all pitfalls. We count four effects that meet this criterion. The first three are treadmill manipulations on perceived distance, reach-extending tools on perceived distance, and body-based manipulations in virtual reality on perceived size (see Philbeck & Witt Reference Philbeck and Witt2015). We describe the fourth in detail.

In a paradigm known as Pong, participants attempted to catch a moving ball with a paddle that varied in size from trial to trial, and then estimated the speed of the ball. Previous research demonstrates that when participants play with a small paddle, the ball is harder to catch and is therefore subsequently judged to be moving faster than when they play with a big paddle (Witt & Sugovic Reference Witt and Sugovic2010). Notably, paddle size influences perceptual judgments only when paddle size also impacts performance. When the ball is similarly easy to catch regardless of paddle size, the paddle has no effect on apparent speed (Witt & Sugovic Reference Witt and Sugovic2012; Witt et al. Reference Witt, Sugovic and Taylor2012). These findings offer both disconfirmatory findings (Pitfall 1) and rule out low-level differences (Pitfall 4).

F&S criticized the term “perceptual judgments” as being vague and ambiguous. However, its use is frequently the researchers' acknowledgment that differentiating perception from judgment is nuanced and difficult. Indeed, F&S were unable to provide a scientific definition, instead relying too heavily on their own intuitions to distinguish perception and judgment (Pitfall 2). For example, comfort could very well be an affordance of an object that can be perceived directly (Gibson Reference Gibson1979). Nevertheless, the issue of distinguishing perception from judgment has been previously addressed. One strategy has been to use action-based measures for which no judgment is required. We modified the ball-catching task so that instead of continuously controlling the paddle, participants had only one opportunity per trial to move the paddle. Successful catches required precisely timing the action, and we analyzed this timing as an action-based measure of perceived speed. If the ball genuinely appears faster when the paddle is small, participants should act earlier than when the paddle is big. As predicted, participants acted earlier with the small paddle, indicating that the ball appeared faster, than with the big paddle (Witt & Sugovic Reference Witt and Sugovic2013a). Because this measure is of action, and not an explicit judgment, the measure eliminates the concern of judgment-based effects (Pitfall 2). This measure also avoids the pitfall of relying on memory (Pitfall 6) because the action was performed while the ball was visibly moving.

Effects with action-based measures can also be taken as evidence against task demands (Pitfall 3). Additionally, we have directly measured participants' willingness to comply with task demands by purposefully inserting task demands into the design of the experiment. Participants were instructed on how to respond (e.g., to make sure to classify all fast speeds correctly), and we grouped participants based on their willingness to conform to these instructions. Importantly, both conforming and nonconforming participants showed identical action-specific effects of paddle size on apparent ball speed (Witt & Sugovic Reference Witt and Sugovic2013b). The finding that nonconforming participants still show the same action-specific effect is evidence against a task demand explanation (Pitfall 3).

A final set of experiments explored the role of attention (Pitfall 5) in the Pong task by adding a secondary, attentionally demanding task (Witt et al. Reference Witt, Sugovic and Dodd2016). In one experiment, the secondary task was to count the number of flashes that occurred at the center of the screen. In another, the secondary task was to fixate on the ball and count the number of flashes that occurred on the ball as it moved across the screen. Regardless of attentional load location, paddle size continued to influence both perceptual judgments and action-based measures of ball speed. In other words, attention-based manipulations did nothing to diminish the action-specific effect; the effect of paddle size on apparent speed persisted in both cases. These studies rule out the final pitfall by showing that attention does not account for this particular action-specific effect.

We commend F&S for raising concrete concerns and future-oriented suggestions. We applied their checklist to one action-specific effect and found that none of the pitfalls could satisfactorily explain the effect of paddle size on apparent ball speed. We therefore conclude this effect is perceptual and demonstrates a genuine top-down influence on perception. Balls that are easier to catch are perceived to be moving slower than balls that are more difficult to catch. Going forward, researchers should apply this checklist to their own work to differentiate between effects that fall into the category of genuine perceptual effects and those that do not. However, the debate about whether there are any top-down effects on perception is decidedly in favor of a nonmodular view of vision.