What of studies that measure whether people see a stimulus at all?

The inattentional blindness and emotion-induced blindness literatures respectively suggest that categorization and emotion mold our ability to see things in the first place. Although the authors appear to dismiss such phenomena as “peripheral” effects of attention (sect. 4.5), attention guides perception at multiple points in visual processing, and it is likely incorrect to relegate its role only to the gating of input into early vision. (The authors vaguely concede that attention is not always so peripheral, but ultimately, they gloss over this point.) Evidence indeed suggests that these effects cannot be waved away as merely a result of peripheral selection (nor to memory, which the authors identify as another potential pitfall).

Inattentional blindness

The ability to see that something is present at all depends on more than where we direct our eyes. Nowhere is this more apparent than in the phenomenon of inattentional blindness: the failure to see obvious and salient stimuli that people look directly at while their attention is preoccupied (Mack & Rock Reference Mack and Rock1998; Most Reference Most2010; Most et al. Reference Most, Simons, Scholl, Jimenez, Clifford and Chabris2001; Reference Most, Chun, Widders and Zald2005a; Reference Most, Scholl, Clifford and Simons2005b; Simons & Chabris Reference Simons and Chabris1999).

Inattentional blindness experiments suggest that categorization shapes perception in ways that cannot be attributed to selection for visual input. In one study, participants either tracked four moving digits and ignored four moving letters or tracked the letters while ignoring the digits (Most Reference Most2013). When an unexpected E traveled across the screen, those tracking the letters were more likely to notice it than those tracking the digits, and this pattern reversed when the unexpected object was its mirror image, a block-letter 3. Because this effect was driven by the categorization of the unexpected object, it must have stemmed from selection after a degree of visual processing had already occurred. Evidence further suggests that the role of categorization in inattentional blindness is unlikely caused by confusions between the unexpected object and other members of the nontarget set (Koivisto & Revonsuo Reference Koivisto and Revonsuo2007). Nor do inattentional blindness effects appear to be wholly attributable to failures of memory (Ward & Scholl Reference Ward and Scholl2015).

Emotion-induced blindness

In emotion-induced blindness (EIB), people view rapid serial visual presentations of items and search for a single target within each stream. When the target is preceded by an emotionally powerful picture, people are unable to report the target (Most et al. Reference Most, Chun, Widders and Zald2005a).

At first glance, EIB seems like something that could fall within one of two categories of F&S's pitfalls. First, because emotional stimuli capture attention (MacLeod et al. Reference MacLeod, Mathews and Tata1986), attention could simply be too preoccupied to guide input of the target into the visual system (sect. 4.5). Second, it could be that EIB is more a phenomenon of memory than of perception (sect. 4.6), either because the measure is typically retrospective (people are usually asked to report the target at least half a second after its appearance; see Wolfe Reference Wolfe and Coltheart1999) or because it seems phenomenally related to the attentional blink, which itself has been attributed to failures to consolidate information into visual working memory (e.g., Chun & Potter Reference Chun and Potter1995). However, there are reasons to suspect that neither peripheral aspects of attention nor memory are accountable.

Specifically, in contrast to studies demonstrating that emotional stimuli capture attention to their location, in EIB target perception is worse at the location of the emotional distractor (e.g., Most & Wang Reference Most and Wang2011). This pattern has led to suggestions that emotional distractors compete for neural representation with targets that appear in the same receptive field (Wang et al. Reference Wang, Kennedy and Most2012; also see Keysers & Perrett Reference Keysers and Perrett2002), a suggestion consistent with findings that neural responses to targets and emotional distractors exhibit a trading relationship (Kennedy et al. Reference Kennedy, Rawding, Most and Hoffman2014). Because people tend to prioritize emotional information, it is the emotional distractors that win, an effect that seems to be modulated by mood (Most et al. Reference Most, Laurenceau, Graber, Belcher and Smith2010).

As with inattentional blindness, EIB appears to be a perceptual phenomenon rather than a memorial one. When participants were instructed to respond to targets immediately upon seeing them rather than at the end of each trial, the effect was undiminished (Kennedy & Most Reference Kennedy and Most2012). In addition, the spatially localized nature of the effect suggests that it arises from competition at a stage prior to consolidation into working memory.

In sum, F&S provide an incisive critique of the literature on top-down effects on perception. But they overstate their case when claiming that evidence for such effects is nonexistent. Surveying the literature on what people do or don't see rather than on their perceptual judgments reveals instances that don't easily fall within the categories of pitfall F&S outline. Of course, the authors may argue that such effects still fall outside the bounds of “perception,” strictly defined, and that they fall prey to other misconceptions about the relationship between perception and cognition. So far, however, while stating that perception extends beyond the computations involved in “early vision,” they have left the placement of the line between perception and cognition ambiguous. Good fences make good neighbors; as the target article's authors no doubt consider perception and cognition to be neighboring – if not encroaching – domains of the mind, it would be neighborly to provide a better map of where they think that fence stands.