As an outsider to this field of study, I am struck by the preponderance of data from experimental studies in which subjects are in one way or another instructed on what they must attend to. This is even true to a large extent in studies with nonlinguistic animals in which there is training about which stimulus will be rewarded and then assessment of how much and how selectively they attend to exemplars (Zentall, Reference Zentall2005). But how does attention function in naturally behaving organisms?
In Tomasello (Reference Tomasello2022), I had occasion to think about attention from an evolutionary perspective. My primary question was what it is for, why attention as distinct from perception, evolve in the first place? Simple organisms like amoebae very likely just perceive the world directly, without attention as a distinct process. One hypothesis is that this is because they are not making agentive decisions. Generally supportive of this view, Gibson (Reference Gibson1979) proposes that attention is best thought of as intentional perception, that is, perceptual processes that the agentive organism actively controls in pursuit of its goals. This implies that only organisms that act as goal-directed agents operate with attention as distinct from perception (very likely excluding organisms without a nervous system and perhaps others). One could even hypothesize that only in organisms that operate with an executive tier of operation – or at least with internally represented goals – is attention distinct from perception in general. In this view, attention evolved as one aspect of a control system or executive process to facilitate behavioral decision making, enabling goal-directed organisms to selectively engage with those aspects of its perceptual environment that are relevant to its decision making. The proposal that organisms naturally attend to situations relevant to their behavioral decision making is generally consistent with Rosenholtz’s suggestion that what organisms are attending to is not sensory stimuli but “tasks” and with Krauzlis et al.’s (Reference Krauzlis, Bollimunta, Arcizet and Wang2014) focus on the role of attention in value-based decision making.
Attention may operate either top-down or bottom-up. Top-down attention evolved to facilitate the organism’s perceptual engagement with situations in the environment that are relevant to its goals and values, that is, as either opportunities or obstacles. A foraging primate will naturally attend to the situation where there is edible food in this tree. It will also attend to the situations: the food is accessible, it is ripe, and there is a conspecific competitor nearby. It can attend to all of these various situations in a single visual percept of a banana in a tree – although perhaps not all at the same time. The key construct here is the relevance of situations to potential goal-directed (adaptive) decision making and action. Thus, when squirrels are making a decision between two paths to a goal, they often show “head switch” behavior, attending first to one and then the other alternative before choosing (Chow et al., Reference Chow, Leaver, Wang and Lea2015). Bottom-up attention operates in situations that are potentially more urgent, so that natural selection cannot leave it to the individual organism’s judgment of whether a situation is or is not relevant. For example, a loud noise in a certain direction demands attention from most organisms, as does the smell of certain foods in a given direction. It does not matter what the organism’s immediate goals are; natural selection has determined that it should pay attention to some situations no matter what: natural selection has predetermined their relevance for adaptive action. Thus, in addition to attending to the fact that food is accessible, the foraging primate may have its attention drawn, bottom-up, to the loud screech of a predator as indicating that it is nearby.
It is important that what the organism attends to in natural, action-based situations is not an object or stimulus property (a prejudice that is likely a holdover from behaviorism’s focus on stimulus and response). When the issue is behavioral decision making, the target of attention is a whole, fact-like situation. The primate attends to the fact-like situations that there is a banana in this tree, that it is ripe, that it is accessible, and, bottom-up, to the fact that a predator is nearby. It attends to the world in this way because its goals – as the telos of its actions – are formatted in this way, that is, as situations. The goal pursued by the primate is not a banana but rather the situation of having the banana in its possession. And so it must attend to relevant environmental situations as opportunities and obstacles for goal attainment. Of course, an organism may attend closely to an object to determine if it is, for example, a banana. But that act of attention is only relevant because it identifies possible food in this tree in front of me, which constitutes a situation of opportunity for foraging. Attention to objects and properties is always in the service of recognizing potentially relevant situations. The proposal that naturally behaving organisms naturally attend to whole situations or states of affairs is generally consistent with Rosenholtz’s observation that humans attend to and comprehend especially quickly and easily “the gist of a scene.”
Researchers may still want to focus on subjects’ attention to prespecified objects or properties in controlled experiments to answer important questions of mechanism. But it might also be the case that considering why attention evolved in the first place and how it operates in its “natural habitat” of value-based decision making – its function of securing perceptual engagement with relevant situations – may open up new avenues of research for escaping from what Rosenholtz calls the “crisis” in the field.
You have
Access
As an outsider to this field of study, I am struck by the preponderance of data from experimental studies in which subjects are in one way or another instructed on what they must attend to. This is even true to a large extent in studies with nonlinguistic animals in which there is training about which stimulus will be rewarded and then assessment of how much and how selectively they attend to exemplars (Zentall, Reference Zentall2005). But how does attention function in naturally behaving organisms?
In Tomasello (Reference Tomasello2022), I had occasion to think about attention from an evolutionary perspective. My primary question was what it is for, why attention as distinct from perception, evolve in the first place? Simple organisms like amoebae very likely just perceive the world directly, without attention as a distinct process. One hypothesis is that this is because they are not making agentive decisions. Generally supportive of this view, Gibson (Reference Gibson1979) proposes that attention is best thought of as intentional perception, that is, perceptual processes that the agentive organism actively controls in pursuit of its goals. This implies that only organisms that act as goal-directed agents operate with attention as distinct from perception (very likely excluding organisms without a nervous system and perhaps others). One could even hypothesize that only in organisms that operate with an executive tier of operation – or at least with internally represented goals – is attention distinct from perception in general. In this view, attention evolved as one aspect of a control system or executive process to facilitate behavioral decision making, enabling goal-directed organisms to selectively engage with those aspects of its perceptual environment that are relevant to its decision making. The proposal that organisms naturally attend to situations relevant to their behavioral decision making is generally consistent with Rosenholtz’s suggestion that what organisms are attending to is not sensory stimuli but “tasks” and with Krauzlis et al.’s (Reference Krauzlis, Bollimunta, Arcizet and Wang2014) focus on the role of attention in value-based decision making.
Attention may operate either top-down or bottom-up. Top-down attention evolved to facilitate the organism’s perceptual engagement with situations in the environment that are relevant to its goals and values, that is, as either opportunities or obstacles. A foraging primate will naturally attend to the situation where there is edible food in this tree. It will also attend to the situations: the food is accessible, it is ripe, and there is a conspecific competitor nearby. It can attend to all of these various situations in a single visual percept of a banana in a tree – although perhaps not all at the same time. The key construct here is the relevance of situations to potential goal-directed (adaptive) decision making and action. Thus, when squirrels are making a decision between two paths to a goal, they often show “head switch” behavior, attending first to one and then the other alternative before choosing (Chow et al., Reference Chow, Leaver, Wang and Lea2015). Bottom-up attention operates in situations that are potentially more urgent, so that natural selection cannot leave it to the individual organism’s judgment of whether a situation is or is not relevant. For example, a loud noise in a certain direction demands attention from most organisms, as does the smell of certain foods in a given direction. It does not matter what the organism’s immediate goals are; natural selection has determined that it should pay attention to some situations no matter what: natural selection has predetermined their relevance for adaptive action. Thus, in addition to attending to the fact that food is accessible, the foraging primate may have its attention drawn, bottom-up, to the loud screech of a predator as indicating that it is nearby.
It is important that what the organism attends to in natural, action-based situations is not an object or stimulus property (a prejudice that is likely a holdover from behaviorism’s focus on stimulus and response). When the issue is behavioral decision making, the target of attention is a whole, fact-like situation. The primate attends to the fact-like situations that there is a banana in this tree, that it is ripe, that it is accessible, and, bottom-up, to the fact that a predator is nearby. It attends to the world in this way because its goals – as the telos of its actions – are formatted in this way, that is, as situations. The goal pursued by the primate is not a banana but rather the situation of having the banana in its possession. And so it must attend to relevant environmental situations as opportunities and obstacles for goal attainment. Of course, an organism may attend closely to an object to determine if it is, for example, a banana. But that act of attention is only relevant because it identifies possible food in this tree in front of me, which constitutes a situation of opportunity for foraging. Attention to objects and properties is always in the service of recognizing potentially relevant situations. The proposal that naturally behaving organisms naturally attend to whole situations or states of affairs is generally consistent with Rosenholtz’s observation that humans attend to and comprehend especially quickly and easily “the gist of a scene.”
Researchers may still want to focus on subjects’ attention to prespecified objects or properties in controlled experiments to answer important questions of mechanism. But it might also be the case that considering why attention evolved in the first place and how it operates in its “natural habitat” of value-based decision making – its function of securing perceptual engagement with relevant situations – may open up new avenues of research for escaping from what Rosenholtz calls the “crisis” in the field.
Financial support
None.
Competing interests
None.