Despite William James’s famous assertion that “everyone knows what attention is” (James Reference James1890), no one can refute the diversity of phenomena embedded within the word “attention” – it includes spatial attention, temporal attention, feature-based attention, object-based attention, transient attention, sustained attention, and more. This proliferation goes even beyond that laid out in the target article because the very functions and targets of attention are myriad: attention selects, modulates, and maintains (as in working memory). You can attend outwardly to the patterns in a leaf or to your internal thoughts (e.g., Chun, Golomb, & Turk-Browne, Reference Chun, Golomb and Turk-Browne2011). As Rosenholtz suggests, the over-application of the notion of attention renders it nearly impossible to develop a comprehensive theory of what attention is.

It is particularly important that this message reaches beyond attention researchers, who are already well-acquainted with this state of affairs. In clinical psychology, for example, a thriving literature aims to understand not only how disorders are characterized by biases to attend to disorder-relevant stimuli (e.g., threat-related cues in anxiety or food-related cues in binge eating; MacLeod, Grafton, & Notebaert, Reference MacLeod, Grafton and Notebaert2019; Schmitz et al., Reference Schmitz, Naumann, Trentowska and Svaldi2014), but also how retraining of such attentional biases might serve as a form of intervention (Bar-Haim, Reference Bar-Haim2010). Within this literature, discrepant results might be resolved through greater recognition that different attentional bias tasks (e.g., dot probe, visual search, emotion-induced blindness) do not tap into the same mechanism. As noted by one research team working in this area, “[a] notable psychometric weakness of the attentional bias measures is their low degree of convergent validity. Although they are all assumed to measure the same theoretical construct, different measures of attentional bias often do not correlate with one another” (Van Bockstaele et al., Reference Van Bockstaele, Verschuere, Tibboel, De Houwer, Crombez and Koster2014). This assumption – that they measure the same theoretical construct – underscores the importance of Rosenholtz’s point. As a corrective, work in our own lab has aimed to shine a spotlight on distinctions between mechanisms of emotion-driven attentional bias (e.g., Onie & Most, Reference Onie and Most2017, Reference Onie and Most2022), but Rosenholtz’s exercise of banishing the word “attention” in order to bring, well, attention to actual mechanistic distinctions may be a salve that salvages this often inconsistent literature.

That said, inattentional blindness – which features heavily in the target article – is a phenomenon that is difficult to describe without reference to attention. One of Rosenholtz’s alternative suggestions is that inattentional blindness reflects task limits.

However, omitted from the target article’s review of inattentional blindness is the well-established role of “attentional set”. This refers to the fact that people can tune their attention to prioritize some visual features over others and that this tuning accounts for tremendous variation in inattentional blindness. For example, when people tracked white shapes and ignored black ones, almost everyone noticed an unexpected novel white shape traversing the display and almost everyone failed to notice an unexpected novel black shape, with the rates of noticing a light or dark gray shape falling in between. This pattern was reversed when people tracked black shapes and ignored white ones (Most et al., Reference Most, Simons, Scholl, Jimenez, Clifford and Chabris2001; also see Simons & Chabris, Reference Simons and Chabris1999). Rates of noticing followed a non-binary linear pattern: the more similar the unexpected object was to the tracked objects, and the less similar it was to the non-tracked objects, the more it was noticed. Subsequent work showed the top-down flexibility of such prioritization, with similar effects of attentional set when people prioritized information on the basis of shape (Most et al., Reference Most, Scholl, Clifford and Simons2005), category membership (Koivisto & Revonsuo, Reference Koivisto and Revonsuo2007; Most, Reference Most2013), and number (White & Davies, Reference White and Davies2008).

The profound degree to which people’s prioritization of information determines inattentional blindness suggests factors that extend well beyond task limits. Driving this point home, in one related study, participants “drove” through a virtual cityscape and encountered a sign with blue and gold arrows at each intersection. Some participants were instructed to always turn in the direction of a gold arrow among blue arrows, whereas others always turned in the direction of a blue arrow among gold arrows. At a critical intersection, a blue or gold motorcycle veered into their path. When the motorcycle matched the color of the arrows people were following, fewer people crashed into it than when it mismatched (Most & Astur, Reference Most and Astur2007). In this study, the motorcycle was equally distinct from the task of following the arrows regardless of the color-match condition. How can we best account for such patterns without reference to the selective prioritization of some features over others – in other words, to the overarching concept of feature-based attention?

Finally, Rosenholtz notes that any viable explanation for inattentional blindness needs to account for findings that observers sometimes move to avoid collisions with unexpected stimuli they claim not to have seen. Just as we must contend with diversity in the meaning of “attention”, it may be fruitful here to highlight diversity in the meaning of “perception”. Famously, Goodale and Milner (Reference Goodale and Milner1992) differentiated between visual processing pathways in the brain that subserve conscious perception and those that subserve action. A particularly striking illustration of this distinction comes from a patient with blindsight, who was able to avoid several obstacles while walking down a hallway despite not consciously seeing them (de Gelder et al., Reference de Gelder, Tamietto, van Boxtel, Goebel, Sahraie, van den Stock, Stienen, Weiskrantz and Pegna2008). The distinction between “perception” and “action” pathways for visual processing may render the fact that people can avoid collisions with things they did not see unproblematic for accounts of inattentional blindness.

As researchers have learned more about mechanisms that might fall within the scope of the big tent we call “attention”, the proliferation of such mechanisms pushes the dream of a unifying theory further out of reach. But this makes this an exciting time to be an attention researcher. The dream of a unifying theory of attention may be in crisis, but the science of attention – in all its diversity – is thriving.