Featured Video Play Icon

Attention as Selection for Action: A Challenge

 Aaron Henry (University of Toronto)

[Jump to comments]

 

Abstract: This paper raises a challenge for Wayne Wu’s account of attention as selection for action. According to Wu’s account, action poses a selection problem which only attention can solve. The need to solve this problem (and hence to attend) is what, according to Wu, distinguishes action from reflex. My challenge to Wu begins with a dilemma concerning the agential status of attention. Either attention is an action or a reflex. If attention is an action, then a vicious regress results. If it is a reflex, then there is no role for the agent to play in action. In either case, action is revealed to be impossible. While Wu’s account can be developed in a way that avoids this dilemma, the view that results has trouble explaining how attending can itself be an action and also conflicts with a promising view of the neural basis of visual attention called ‘biased competition’ theory.

1.    Introduction

This paper raises a challenge for Wu’s (2011a; 2014) theory of attention as selection for action. According to Wu’s theory, action poses a selection problem that only attention can solve. The need to solve this problem (and hence to attend) is what, on Wu’s account, distinguishes action from mere reflex.

My challenge to Wu begins with a dilemma concerning the agential status of attention. Although Wu’s view can be developed in a way that avoids this dilemma, I argue that the resulting view has trouble accommodating the possibility of attention as action and conflicts with a promising theory of the neural basis of attention called ‘biased competition’ theory. So, my argument doesn’t refute Wu’s claim that attention is selection for action. Rather, it provides a constraint on how to develop that claim. The account must explain how attention can be both a necessary component of actions and a potential action in its own right, and it must do this while respecting the neuroscience of attention.

The plan for the paper is as follows: §2 summarizes Wu’s account; §3 states the dilemma; §4 explains how Wu is likely to try to avoid the dilemma; §§5-6 argue that this way of avoiding the dilemma leaves Wu without a satisfactory account of active attention and also conflicts with biased competition theory; §7 concludes with some suggestions for how to develop the selection for action account in a way that would avoid these difficulties.

2.    Attention as Selection for Action

According to Wu, action poses a selection problem. In intuitive terms, there are many things that you can do to each of the many things that you perceive. In Wu’s terms, you face many perceptual ‘inputs’ and a choice between many potential motor ‘outputs’. This yields a many-many mapping between available inputs and potential outputs or what Wu calls a ‘behavioural space’. In order to act, the agent must select one of the input-output mappings (i.e., a ‘path’ through behavioural space). The problem of how the agent does this is the ‘many-many problem’.

According to Wu, attention solves the many-many problem. Attention is the agent’s selection of a path through behavioural space. Since action requires a solution to the many-many problem and only attention can solve this problem, action requires attention. In particular, a behavioural output is an action only if it is guided by an input that the agent selects for action (i.e., to which she attends).

Wu thinks that both the many-many problem and its solution exist at the personal level. The agent’s behavioural space is defined over personal level states, such as visual experiences (2014: 87).[1] And in order to act the agent must select a path through this space. Selection by sub-personal mechanisms would, Wu argues, abolish agency (2011a: 102).

These points connect with how Wu distinguishes action from mere reflex. Whereas the agent selects which action to perform on which object, she doesn’t select her reflexes.[2] This is because the many-many problem doesn’t arise for reflexes. If reflexes are selected at all, it is purely sub-personally and without the agent’s involvement. What distinguishes action and reflex, then, is that in action, but not in reflex, the agent is presented with options from which to select (Wu 2014: 82, 90). The distinction between action and reflex is, as I read Wu, meant to provide a mutually exclusive and jointly exhaustive classification of behaviour.

With perceptually guided intentional bodily action as his home case, Wu argues that attention is also required for cognitively guided bodily actions, mental actions, and automatic actions. Attention is necessary for action of any sort.

3.    The Dilemma

A challenge for Wu’s account arises when we inquire about the agential status of attention itself. Specifically, Wu seems to face a dilemma: either selection for action is an action or it is a reflex. If it is an action, then we face a regress. For, on Wu’s view, every action requires selection for action. If selection for action is itself an action and every action requires selection for action, then selection for action itself requires selection for action. Given that selection for selection for action is an action as well, it also requires selection for action, and so on ad infinitum. It seems that selection for action cannot be both an action and a precondition for action of any sort.

If, however, selection for action is a reflex, then there is no role for the agent in action. On Wu’s view, action (as opposed to reflex) requires that the agent select one option from many. But if selection for action is a reflex, then it isn’t the agent that selects for action, since reflexes do not implicate the agent. And if the agent doesn’t select, then the agent doesn’t solve the many-many problem and action doesn’t occur. Whether attention is an action or a reflex, then, action is revealed to be impossible.

4.    The Act-Component Proposal

Wu is likely to respond to the above dilemma by adopting what I shall call theact-component proposal’ (henceforth, ‘acp’). ACP avoids the dilemma by denying that attention is either an action or a reflex. According to ACP, attention is not a type of behaviour at all, but a component of behaviour. Specifically, attention is a necessary component of actions. And act-components, unlike actions, don’t require selection for action. Consequently, if selection for action is an act-component, then selection for action doesn’t require prior selection, and we avoid a regress. Neither is attention a reflex, since a reflex is a behaviour in its own right, whereas attention is strictly an aspect of behaviour. The core idea behind ACP is that selection for action is not an additional thing you do in order to act, but part of what it is to act.[3]

There are different ways one can develop ACP (cf. Wu 2014: 97). One can take attention to be a state of the subject—e.g., the input that plays the causal role of guiding behaviour—or a psychological process—e.g., the coupling of a certain input to a certain output. My concern is not with the details of each version of ACP, but with a specific consequence of the view however it is elaborated: the consequence that attention is never itself a behavioural output. Attention is either the input that guides an output or the coupling of input and output, but never an output. For, it is by denying that attention is an output—whether an action or reflex—that ACP avoids our dilemma.

5.    Active Attention

Intuitively, it is possible to form an intention simply to attend to something. And when you attend in this way—i.e., as a result of intending to do so—your attending is itself an action. Wu agrees and calls this ‘controlled’ attention (2014: 33), ‘voluntary’ attention (2011c: 98, 108), ‘active’ attention, and ‘attention as action’ (2011a: 105)

On the face of it, the claim that attention is sometimes an action conflicts with the claim that attention is strictly an act-component. Recall that if attention is an act-component, then it is never an output. But if attending can be an action, then attention can be an output, because actions are a type of output.[4]

According to Wu, this conflict is only apparent. To see why, consider the action of intentionally shifting and maintaining visual attention to an object. This has been demonstrated to alter the appearance of the attended object (e.g., by increasing apparent contrast) (Carrasco, Ling, and Read 2004). With these results in mind, Wu writes:

Shifting and maintaining attention in this active way is just a type of mental action, altering the appearances of a consciously perceived input or maintaining that alteration. Here, input and output are perceptual states. (2011a: 105)

Wu’s proposal is that active (covert, visual) attention consists in selecting a visual state for alteration or maintenance of appearance.[5]

My first objection to Wu’s proposal is that it depicts the agent’s selection as related instrumentally to an end other than selection. You select an object in order to do something else to it (namely, to modify your perception of it). The possibility that active attention seems to raise, however, is that agents can also attend non-instrumentally, i.e., for its own sake. By attempting to explain active attention within the parameters of ACP, Wu seems to lose what’s intuitively distinctive about the phenomenon.

One way for Wu to respond is to deny the existence of active attention in any sense that entails a pure act of selection. This response can in turn be developed in two ways. First, one might draw a distinction between the state of attention and the process of attending. The state of attention is the input that occupies the role of guiding an agent’s action.[6] The process of attending is the action of making an object perceptually salient by altering its appearance. Attending, because an action, has attention as a component.[7]

Alternatively, one might simply abandon the notion of active attention and replace it with another—call the action ‘highlighting’’. This proposal is merely a terminological variant on the previous one: in highlighting, as in ‘attending’, the action is the enhancement of perception. Whatever terminology one chooses to adopt, the point is that ACP remains tenable once we specify the specific action that attention is being used to serve (in the present case, perceptual enhancement). There cannot, however, be an action which consists simply in the agent’s selection of an object.

In my view, this line of response—in either of its terminological variants—fails adequately to accommodate the intuition that agents can attend non-instrumentally. To accommodate that intuition adequately, one would have to allow that you can attend to a thing—i.e., select it—without having the further aim of acting on it (beyond, that is, what’s involved in its selection). Otherwise, the action that you select will be distinct from your selection of it and so not a case of attention as action (which would entail a coincidence of selection and action).

Rather than pressing this intuition farther, however, I take a different tack for the remainder of the paper. I shall argue on empirical grounds that Wu’s proposal about active attention mischaracterizes the relationship between attention and perception: attention doesn’t guide changes of perception, but partly consists in such changes.

6.    The Neural Basis of Selection

This section argues that Wu’s proposal about active attention sits uneasily with a promising account of the neural basis of perceptual attention called ‘biased competition theory’—an account for which Wu himself has recently expressed sympathy (forthcoming).

Before reviewing the neuroscience, let’s note a consequence of Wu’s proposal about active attention. It follows from Wu’s proposal that attention can alter which inputs an agent receives. When you select an input for alteration or maintenance of visual appearance, the output of your selection is the altered or maintained visual state. That visual state then becomes as input to the next round of action selection. It follows that how you visually attend can partly determine which inputs you receive.

Significantly, though, the sense in which attention can ‘determine’ input on Wu’s view is causal rather than constitutive. To see why it cannot be constitutive, it is sufficient to reflect on the general form of the many-many problem. The many-many problem is defined over a pre-specified set of inputs, and the alleged solution is to select, from this given set, which input to inform a response. This solution is intelligible only if which inputs you receive at a time is independent of how you select at that time (since how you select at a time itself depends on which inputs are available to be selected). This suggests that, on Wu’s view, inputs are constitutively independent of attention. We must therefore interpret Wu’s claim about active attention to be that attention can causally determine input.[8],[9]

The neuroscientific account of visual attention I present below disagrees with Wu on these points. This account implies that visual attention doesn’t consist in selection from a pre-attentively constituted class of inputs, but in the generation of the subject’s input state. Let’s turn, now, to review the neuroscientific results relevant to understanding the biased competition theory of visual attention.[10]

One mark of attention at the neural level is receptive field remapping. To understand this, it is necessary to explain a few principles of neural activity. First, visual neurons possess ‘receptive fields’. This is the region of the retina stimulation of which causes a neuron to fire above baseline. In addition to having receptive fields, neurons possess stimulus preferences. A neuron’s preferred stimulus is the type of stimulus in response to which it fires most intensely (e.g., a certain line orientation). The less similar a stimulus is to the neuron’s preferred type, the less intensely that neuron fires in response to it.

What happens if two stimuli—one preferred, another non-preferred—stimulate a neuron’s receptive field at the same time? In one class of visual neuron, the response will be intermediate between what it would be had only the preferred stimulus been present and what it would be had only the non-preferred stimulus been present.

This changes with attention, however (Moran and Desimone 1985; Chelazzi et al. 2001). If, while both the preferred and non-preferred stimuli are stimulating the neuron’s receptive field, the subject shifts attention to one of these stimuli in order to act on it in some way (e.g., to redirect one’s eyes toward it), the neuron behaves as if only the attended stimulus were present. This is receptive field remapping: it is as if the neuron’s receptive field shrinks around the task-relevant stimulus, excluding the irrelevant stimulus from determining the neuron’s response.

Receptive field remapping is an example of biased competition. When the agent has the task of acting on one of two stimuli hitting the same neuron’s receptive field, priority is given to the task-relevant stimulus over the task-irrelevant stimulus in determining the outcome of the competition for the neuron’s response. A remapped receptive field is the result of this biased competition.

What, according to the biased competition theory, is the relationship between visual attention and biased competition? According to the theory’s original proponents, attention is the ‘emergent’ result of biased competition (Desimone and Duncan 1995). In our example, the neural realizer of visual attention (in a single neuron) is the state of having a remapped receptive field (rather than, e.g., the bias signal which causes the remapping).[11]

Now, it is known that activation of visual ventral neurons partly underlie a subject’s visual state and that changes in activation of these neurons partly underlie changes of visual state (Milner and Goodale 2006).[12] If the neural modulations that realize a subject’s visual attention—e.g., the receptive field remapping of ventral neurons—also realize a change of visual state, then visual attention entails a change of visual state. The claim here is not that attention causes this visual change, but that attention partly consists in this change. For, the change of visual state that receptive field remapping realizes is the result of task-biased competition, and according to biased competition theory visual attention is this resultant state.

Here, then, is the conflict. On the biased competition theory, visual attention consists in a certain change of the subject’s visual state, which is realized by a modified neural representation of the object on which it is one’s task to act. In contrast, on Wu’s account of active attention, task-biased changes of visual state are effects or outputs of attention (namely, the agent’s selection of an input for alteration). Whereas the first view takes attention to consist in a change of visual state, the second view takes attention to be the cause of this change.

7.    Conclusion

I’ve been arguing that Wu’s account of attention as selection for action faces difficulties. To avoid the dilemma posed in §3, Wu must adopt ACP. But ACP has trouble accommodating the claim that attention is sometimes an action. Moreover, Wu’s specific attempt to accommodate this claim runs into conflict with biased competition theory. I close with a few remarks about how one might develop the selection for action account to avoid these difficulties.

First, to reconcile the selection for action account with biased competition theory, we must reconsider certain assumptions behind the many-many problem. In particular, we must reject the assumption that attention consists in an agent’s selection, from a class of pre-attentively constituted inputs, which input to guide action (including, as a special case, the action of highlighting). This is because there is no role in biased competition theory for attention understood as selection from a pre-attentively constituted domain of personal level inputs. The theory instead identifies visual attention with the task-biased generation of a subject’s input state. To this extent, biased competition theory seems inconsistent with Wu’s characterization of the many-many problem and its solution.

Rejection of this feature of the many-many problem is, however, consistent with the broader claim that attention is selection for action. We might, e.g., understand the agent’s selection as the result or outcome of a motivationally guided process: the process of bringing perception into a state which is suitable to guide one’s action and thereby fulfil the guiding motivational state.[13]

Notice the comparative ease with which this version of the theory can accommodate the intuition that, in active attention, agents attend non-instrumentally. Following biased competition theory, suppose that visual attention consists in the visual state that results from task-biased competition. There’s no barrier to attention being an action on this view, because the task that biases the outcome of neural competition might itself be directed at the achievement of a certain visual state. When you intentionally highlight an object (and do nothing else), the task that biases your resulting visual state is given by your intention to visually attend to a certain object, and visual attention to an object is the (motivationally guided) achievement of a modified perception of it. Selection and action here coincide.

Modifying the view in this way requires rejecting ACP. For, we are rejecting the claim that active attention (highlighting) is guided by attention and has attention as a component, and ACP holds that every action possesses attention as a guiding component (including active attention). If we reject ACP, then we confront the dilemma posed in §3. The best solution may be to deny that every action is guided by attention. Attention, when an action, is not thus guided.

 

 

Bibliography

Allport, Alan. 2011. “Attention and Integration.” In Attention: Philosophical and Psychological Essays, edited by Christopher Mole, Declan Smithies and Wayne Wu, 24-59. New York: Oxford University Press.

Carrasco, Marisa, Sam Ling, and Sarah Read. 2004. “Attention Alters Appearance.” Nature Neuroscience 7 (3): 308-313.

Chelazzi, Leonardo, Earl K. Miller, John Duncan, and Robert Desimone. 2001. “Responses of Neurons in Macaque Area V4 during Memory-Guided Visual Search.” Cerebral Cortex 11 (8): 761-772.

Desimone, Robert and John Duncan. 1995. “Neural Mechanisms of Visual Selective Attention.” Annual Review of Neuroscience 18: 193-222.

Johnston, William A. and Veronica J. Dark. 1986. “Selective Attention.” Annual Review of Psychology 37: 43-75.

Milner, David A. and Melvyn A. Goodale. 2006. The Visual Brain in Action. Second ed. New York: Oxford University Press.

Mole, Christopher. 2011. Attention is Cognitive Unison: An Essay in Philosophical Psychology. New York: Oxford University Press.

Moran, Jeffrey and Robert Desimone. 1985. “Selective Attention Gates Visual Processing in the Extrastriate Cortex.” Science 229 (4715): 782-784.

Stinson, Catherine. 2009. “Searching for the Source of Executive Attention.” Psyche 15 (1): 137-154.

Watzl, Sebastian. 2011. “Attention as Structuring of the Stream of Consciousness.” In Attention: Philosophical and Psychological Essays, edited by Christopher Mole, Declan Smithies and Wayne Wu, 145-173. New York: Oxford University Press.

Wu, Wayne. 2011a. “Attention as Selection for Action.” In Attention: Philosophical and Psychological Essays, edited by Christopher Mole, Declan Smithies and Wayne Wu, 97-116. New York: Oxford University Press.

———. 2011b. “Confronting Many‐Many Problems: Attention and Agentive Control.” Noûs 45 (1): 50-76.

———. 2011c. “What is Conscious Attention?” Philosophy and Phenomenological Research 82 (1): 93-120.

———. 2013. “Mental Action and the Threat of Automaticity.” In Decomposing the Will, edited by Andy Clark, Julian Kiverstein and Tillman Vierkant, 244-261. New York: Oxford University Press.

———. 2014. Attention. New York: Routledge.

———. forthcoming. “Shaking Up the Mind’s Ground Floor: TheCognitive Penetration of Visual Attention.”

 

 

Notes

[1] In this paper, I focus exclusively on the visual inputs.

[2] At least, she doesn’t do so directly. One could choose to bring about the conditions that cause a reflex, thereby selecting a reflex indirectly. But I set this possibility aside.

[3] Cf. Wu (2013):

In mental action, solving the Many-Many Problem by making appropriate selection imputes a certain form of activity to the agent. This activity, however, is not an additional thing that the agent does so as to act in the intended way … Rather, these are just aspects of or part of the mental action itself … The same point applies to any selection relevant to solving the Many-Many Problem (251-2, italics added).

It’s reasonable to read Wu as endorsing ACP in this passage. Moreover, Wu (2011b) argues that what guides an action is a component of the action. Specifically, he argues that an action’s guidance mechanism is part of the action’s ‘internal causal structure’ (i.e., part of the causal process of acting), rather than something that controls the action from outside (ibid. 63-5). If, as Wu thinks, attention is necessarily what guides action, then it is, by these assumptions, a necessary component of action.

[4] More precisely, actions are outputs, together with whatever it is that guides them (see n. 3).

[5] Elsewhere Wu offers a different account of active attention (Wu 2011c). On this alternative conception, active attention consists in a type of thought about the attended object. In particular, when you intentionally attend to a thing, you are aware that you are attending to it, where this consists in thinking demonstratively about the attended object as being the target of your attention (ibid. 110). I cannot engage with this proposal here, though it is vulnerable to the same sort of objection that I am about to raise against the account of active attention in terms of perceptual modification.

[6] This entails the state conception of attention discussed at the end of §4.

[7] This response was suggested to me by Wu in correspondence.

[8] Is the claim that attention can causally determine an agent’s inputs consistent with the claim that every input is constitutively pre-attentive? Yes. To see why, take a case in which you select an input in order to highlight it. Wu’s proposal is that attention to the un-highlighted input causes your highlighted visual state, and that the highlighted visual state then becomes an input to the next cycle of action selection. From the perspective of the second cycle of action selection—the same perspective from which the highlighted visual state qualifies as input as opposed to output—the highlighted state is no different than other inputs in being a member of the set of inputs over which the agent’s current many-many problem is defined. Insofar as it is an input, then, the highlighted state is constitutively pre-attentive, even if it is caused by earlier attention to a different input state.

[9] Further evidence that Wu regards the relationship, here, as causal rather than constitutive comes from his discussion of attention’s contribution to phenomenal character (2014: ch.4). Wu argues that attention-induced changes in visual appearance (e.g., as documented by Carrasco and colleagues) fail to characterize the phenomenology of attention and instead characterizes the effects of attention on visual phenomenology (2014: 115, 140). That is, although attention makes a difference to phenomenal character, its contribution is causal. Changes in how you attend cause changes in how things visually appear to you, but such changes in visual appearance aren’t constitutive of attention or its phenomenal character.

[10] My presentation of this material closely follows Wu’s own (2014: 57-60; forthcoming).

[11] Cf. Allport (2011: 32-33), Mole (2011: 134), Stinson (2009), Wu (2014: 60; forthcoming), each of whom argue for an ‘effect’ construal of biased competition theory. For the distinction between cause and effect theories, see Johnston and Dark (1986).

[12] Indeed, Wu (forthcoming) argues that intention-guided visual attention is an instance of cognitive penetration. His thesis is that, by modulating visual attention, the content of an agent’s intention to act influences computations performed in the ventral visual system and that this influence fulfills reasonable criteria for the cognitive penetration of vision. Crucial in the present context is the assumption that in influencing a subject’s visual attention intention influences the subject’s visual state.

[13] It is tempting, in this context, to use the language of ‘structuring’ (Watzl 2011): the agent’s selection for action is the result of structuring the perceptual field so that it is apt to guide action. Here, the agent’s state of attention is the constitutive result of the motivationally guided process of attending (or structuring).

Published by

Aaron

I am a PhD student at the University of Toronto writing my dissertation on perceptual attention at the interface between the philosophy of perception and the philosophy of action.

45 thoughts on “Attention as Selection for Action: A Challenge”

  1. It is a pleasure to have a chance to engage with the work of Wayne Wu and Aaron Henry at more length, and I thank the organizers of this conference for giving me the opportunity to do so. I first came across Wu’s work in 2009 and while we share much in terms of approach, we differ in our published views (see, e.g., my review of his recent book in NDPR). I met Henry at a workshop on attention in Antwerp in 2012, and we have had a few opportunities to discuss our differing perspectives since then. In this paper he challenges Wu, but his challenges are much different than my own, raising some important new issues. In this commentary I will argue that Wu can meet Henry’s challenges without capitulation.

    So, first, here is a very short summary of Wu’s extensive work on attention and action: For Wu, attention is necessary for action. It plays the role of connecting incoming sensory information to potential responses to that information, guided by the intentions and motivations of an agent (see Figure 1). Without attention, behavior would not be guided by the intentions of the agent, and so would not qualify as action. Moreover, without action there is no attention—attention just is selection-for-action, according to Wu. That attention can guide action in this way is uncontroversial, and was first suggested by psychologists Allport and Neumann in the late 80s. That attention is necessary for action is more controversial, and neither Allport nor Neumann argued for this view. I have argued against it myself in an article with Bence Nanay (Analysis, 2016). However, the focus of this discussion is an entirely different set of concerns about Wu’s selection-for-action view.

    Figure 1

    Figure 1. In the selection-for-action view, incoming sensory input is subjected to selection for action, guided by the agent’s intentions. Those intentions make some of the incoming sensory input more salient, which itself leads to a specific behavioral output (mental or physical).

    According to Henry, Wu’s selection-for-action view faces a serious problem: Wu argues that attention is necessary for action but also calls attention an action. Thus, the selection-for-action view is open to a vicious regress. Henry thinks that one way Wu’s account can accommodate this problem is by making attention a component of action (ACP), rather than a stand-alone action. (And Wu sometimes seems to espouse this view.) Yet, this view meets its own difficulties, such as not allowing that an agent might attend for its own sake and not as a component of some other action. Henry thus favors a different approach, which he takes to be informed by the Biased Competition Theory of attention.

    A first note about this debate: it can get confusing. I think that it would be less confusing if we made (you guessed it!) some distinctions. One such distinction is between the terms “input” and “output,” which are sometimes used to describe the input of sensory information and the output of behavioral response, but are at other times used to describe the input to the process of attention and the resulting output of that process. I think clarifying these terms allows us to respond to some of Henry’s arguments against Wu (see the cartoon below).

    Figure 2

    Figure 2. In order to clarify the debate, the input to the process of attention is labeled “Input 1,” whereas the output of the process of attention is labeled “Input 2/Output 2.” The overall behavioral output is labeled “Output 1,” which, in the selection-for-action view, relies on Input 2, the output of the process of attention.

    A second distinction I think is useful is between two types of action, which I call “Pure Action” and “Component Action.” The concept of Pure Action is actually suggested by Henry, but I take the term in a direction that I think is more promising for Wu’s approach. Essentially, Pure Action is an action that could not be mistaken for arbitrary or reflexive behavior, whereas Component Action has pure action as a component. It is the other components of Component Action that allow it to be mistaken for arbitrary or reflexive behavior.

    With those distinctions in mind, I will respond to what I take to be Henry’s most challenging argument. This is Henry’s argument against ACP, or the view that attention is not an action but a component of action. Henry argues that a consequence of this view is that attention is “never an output.” This view is problematic, he thinks, because we sometimes attend simply for the sake of attending, and not as part of another action. Yet, a slight modification of ACP makes this possible. In the modified version, attention is not a Component Action but a Pure Action that is itself a component of Component Action. Attention does not result in outputs of the Output 1 type, which is the type of output yielded by Component Action, but it does result in outputs of the Output 2 type. Thus, in ACP, one can attend for its own sake, or for the sake of Output 2. Importantly, attention is only necessary for Component Action. In my view, these modifications are enough to rescue ACP.

    Henry puts forward a second argument against the selection-for-action view that I take to be less promising. Henry argues that Wu is forced into the position of claiming that attention causes a change in inputs (from Input 1 to Input 2 in the language above), rather than constituting this change in inputs, and that this flies in the face of a promising model of attention, the Biased Competition Theory. Unlike Henry, I do not think that Wu is forced into this position. Henry’s reasoning is as follows:

    It follows from Wu’s proposal that attention can alter which inputs an agent receives. When you select an input for alteration or maintenance of visual appearance, the output of your selection is the altered or maintained visual state. That visual state then becomes as input to the next round of action selection. It follows that how you visually attend can partly determine which inputs you receive.
    Significantly, though, the sense in which attention can ‘determine’ input on Wu’s view is causal rather than constitutive. To see why it cannot be constitutive, it is sufficient to reflect on the general form of the many-many problem. The many-many problem is defined over a pre-specified set of inputs, and the alleged solution is to select, from this given set, which input to inform a response. This solution is intelligible only if which inputs you receive at a time is independent of how you select at that time (since how you select at a time itself depends on which inputs are available to be selected). This suggests that, on Wu’s view, inputs are constitutively independent of attention. We must therefore interpret Wu’s claim about active attention to be that attention can causally determine input.

    In other words, the role of attention is to match inputs to outputs and there is no need for this role if attention is already constitutively determining the inputs. Thus, attention cannot constitutively determine inputs, and so must only causally determine inputs. Yet, I think this argument fails when we distinguish between different senses of “input”: the role of attention is to match Input 1 to Output 1 and there is no need for this role if attention is already constitutively determining Input 1. Thus, attention cannot constitutively determine Input 2, and so must only causally determine Input 2. Clearly, the premise and conclusion are unrelated in this version of the argument, rendering it invalid. Thus, Wu can argue that attention constitutes the change from Input 1 to Input 2, aligning the selection-for-action view with Henry’s reading of the Biased Competition Theory.

    In sum, while Henry and I both take a critical stance with respect to Wu’s selection-for-action view, we differ in which aspects of Wu’s theory we find problematic. Henry has raised important issues for Wu, but I think that Wu has the resources to answer them. As a final note: Wu has provided commentary on this paper in the past (at the 2016 Meeting of the Central APA, which I did not attend) and it may be interesting to hear more about his take on these matters, perhaps in the comments. It is my hope that this will lead to further discussion on how Henry’s views differ from those of Wu and myself, so that we can better assess these views in light of currently available and future evidence.

    1. Thanks Carolyn for the defense! I think on the first point, in terms of understanding “action”, I would take a different route which I detail below in my response to Aaron (not sure where I’m supposed to enter responses). So, I don’t think I would opt for that way out, only because it’s not the way that seems as consistent with what I say about action elsewhere.

      Your distinction between two sorts of input is closer to what I have in mind. In your scheme, I would say that attention should be identified with input 2 and not input 1 as its field of operation, or even the transition from input 1 to 2. That’s too much the spotlight model, attention as a selector. This is where intention does the work that attention has been said to do, at least for voluntary attention which is what is at issue here. Understanding the role of intention in attention is, partly, where I think we are divided regarding attention and action, thinking about your Analysis paper with Bence criticizing the MMP as I conceive of it. But perhaps this isn’t the place to have that discussion (I wish we could have it in Analysis, and I might try to submit the paper again there so that we can, once the editorship changes).

      1. Thanks for this, Wayne! I thought that was how you saw attention, but wasn’t sure, so this is a helpful clarification. I am not sure about the intention point, but I look forward to more discussion on the Analysis paper, in Analysis or wherever.

  2. In his ambitious paper, Henry issues a well-articulated, cutting challenge to Wu’s account of perceptual attention as selection for action and argues that even given the most charitably developed version of the account, the proposal is at odds with key work in neuroscience of visual attention. First I’ll provide some further background on Wu’s account of attention. I will then briefly outline Henry’s challenge, highlighting what I take to be merits and shortcomings or omissions of that challenge. I will end with some questions for Wu (and other interested commentators).

    Wu argues that agents face what he calls the ‘Many-Many Problem’ (see, e.g., 2011; 2014) and that action requires a solution to this problem. Attention, inclusive of perceptual (e.g., visual) attention, is required to solve the Many-Many Problem, and so attention is necessary for action. Specifically, attention is the selection of perceptual inputs, including proprioceptive states,that guide action; to map relevant perceptual inputs to the appropriate behavioral output. This selection process is guided by intentions (and other motivating states). Wu is keen to maintain that attention is not just necessary for action but indeed a “necessary part” of action (in other places discussed as a constituent or part of the “internal causal structure” of action (2011, 51)). Bodily movement itself is another proper part of intentional bodily action. This account is motivated, to some extent, by concerns that prominent causal theories of action are at best incomplete – leave out how intentions guide action — or at worst “abolish agency” (2011, 51). Hence, Wu departs from those causal theorists who hold that what makes a bodily movement an action, and so under control of the agent, is numerically distinct from the action itself: say an intention to so act that initiates, guides, and sustains bodily movement. Rather, Wu argues that “control is internal to the action” and that intentions (and other suitable motivating states) do guide action but as structural causes (2011, 51). The control-grounding aspects of action are part of it. I take it that the role of intentions and the contrast with causal theorists, especially in response to the Causal Deviance problem, is an important aspect of Wu’s account — one that Henry does not put much weight on — and hence I will return to this in my objections and questions below.

    Henry presents a dilemma for Wu’s account of attention: selection for action (i.e., attention) is either an action or a reflex. Henry argues that, on first glance, if selection for action is an action, then we’re off on a vicious regress. Henry further claims that if selection for action is a reflex, the agent is left out. Why would selection as a reflex have this implication? Reflexes are subpersonal and this conflicts with Wu’s commitment to “personal-level, fine-grained selectivity” (Wu 2011, 60). More carefully, Wu argues that subpersonal items cannot exclusively do the work of selection for action because in the subpersonal we lose the agent. Instead, attention is the “subject’s own selectivity” (Wu 2011, 60).

    Regarding the first prong of Henry’s dilemma for Wu (at least on this first pass assessment of Henry’s), I find the problem to be well motivated. Wu himself states that mental actions, like bodily actions, are subject to a Many-Many Problem (2011, 56) and hence if attention is both a mental action and is necessary to solve the Many-Many Problem, we are indeed off to the races (although Wu’s remarks on the Deliberative Many-Many Problem – versus the Non-Deliberative Many-Many Problem at issue in this set of papers — may affect this). I’ll return briefly to the more nuanced attention as an action and action-component potential reply.

    The second prong of the dilemma is weaker. Henry sets it up such that selection for action can only be an action or reflex, but this leaves out many plausible alternatives. If the constraint, set by Wu, is that attention must be personal-level, why not leave open that selection for action is a non-reflex personal-level state, process, or perhaps (non-actional) event? Intentions are often taken to be personal-level states, even if the formation of that intention under uncertainty is taken to be an action. Actions don’t appear to exhaust the available candidates for attention as personal-level. All Wu needs is for attention to be personal-level and guided by intention (or other relevant motivational states).

    Henry then argues that the best response for Wu is to embrace the ‘act-component proposal,’ to argue that “attention is a necessary component of actions” (Henry, 3). The thought is that on this approach attention needn’t require selection for attention because it is merely an act-component and not an action. This potential response does seem the best route for Wu and is consistent with, if not just the same as, his outlined view (see, e.g., Wu 2011; 2013). In fact, as mentioned earlier, Wu is constrained to an answer of this stripe already given his dissatisfaction with the causal theorists reply to Causal Deviance: Attention is a necessary part of (partly constitutive of) action. Here Henry takes up some options that he doesn’t mention in his discussion of the second prong, attention as a state or attention as a psychological process. This kind of take on selection for action is consistent with what Wu expresses in places (e.g., that bodily action is a process with parts). Henry notes that a merit of this tack is that attention wouldn’t then be a behavioral output and so wouldn’t fall prey to the dilemma (he has other worries, of course). This reading of Wu conflicts with some claims in his work, though: In spots Wu is at least open to attention and action as “events (or event-like)” (2011, 70) and yet in the same breath endorses that “events can compose other events” and that attention may be a proper part of action. Likewise, as Henry aptly points out, Wu accepts that in some cases your attending can be an action. Henry thus argues that there is a conflict that, although addressed by Wu, is not adequately resolved: attention as an act-component cannot be an output, but if some attending is action then attention can be an output (if any action is an output). I haven’t the space to engage properly with Henry’s treatment of Wu’s answer, but I will note briefly that there seems no obstacle to actions being components of larger actions, including necessary parts of actions, given mundane examples (e.g., action of running a mile as an act component of running two miles). So, Wu’s response (which Henry cites as in correspondence) of “attending, because an action, has attention as a component” (Henry, 5) has intuitive appeal – if the regress can be stopped. Perhaps the Many-Many Problem, if it arises here for a mental action of attention, draws on distinct capacities for a solution. I’ll leave this aside though.

    Finally, some questions for Henry regarding his objection from the mismatch of Wu’s fleshed out account of attention with the theoretical implications of biased competition theory: Henry notes that the original theorists of biased attention theory take “attention to be the ‘emergent’ result of biased competition” (8) and that this does not square with Wu’s claims that active attention is a cause of changes in visual state. However, competing ‘effect’ construals of biased competition theory are mentioned in a footnote in Henry’s paper but not explored. Hence, it’d be interesting to hear more about reasons for rejecting these competing understandings of biased competition theory, a theory to which Wu himself, as Henry acknowledges, is sympathetic. Further, in past discussions of both active attention and biased competition theory, Wu has leaned on motivational states, especial intentions, in his explanation of how selection for action fits with biasing. Henry mentions Wu’s proposal for the role of intention in action at times, but it isn’t a central player in Henry’s summary of Wu’s account of attention. Intention seems to play a significant role in causing and guiding attention and bodily action on Wu’s watch (see, e.g., 2011). Hence I’ll end with a question: Might Wu’s proposed role for intention in action, both for intentional bodily action and active attention, aid him in answering some of Henry’s well thought-out challenges?

    References
    Henry, A. “Attention as Selection for Action: A Challenge.”
    Wu, W. 2011. “Confronting Many-Many Problems: Attention and Agentive Control.” Noûs 45(1): 50-76.
    Wu, W. 2013. “Mental Action and the Threat of Automaticity.” In A. Clark, J. Kiverstein, & T. Vierkant (eds.) Decomposing the Will. New York: Oxford University Press, pp. 244-261.
    Wu, W. 2014. Attention. New York: Routledge.

    1. Thanks Robyn for your thoughts on the work and responses to Aaron. I find them quite sympathetic though on the first part of the dilemma, you can tell me if the distinctions I draw make a difference to your reading of Aaron’s paper and on how strong his objection is (you are sympathetic to his take there). The economical view is to take attention as the input state that is for guiding response, and it is that role which is critical to addressing broader issues in the philosophy of action. But attention, as selection, is not an action. The action is a special case of looking at attention as guiding a specific type of output, perhaps eye movement as when you are following someone with your eyes, or in just changing your perceptual state (which Aaron discusses below, so I will pick up there since it’s partly in addressing your last point).

      I am also grateful that you highlight the fact that the theory is meant to engage with traditional questions about guidance and control in the philosophy of action. The theory is a theory of attention, and that is what Aaron is pressing on, but it is also a theory of agency and indeed, on traditional questions about it (Carolyn’s work has pressed on this point as well). So, that broader connection is worth highlighting and speaks to, what I think Carolyn, Aaron and I agree on (and perhaps you too), the central role attention plays in different aspects of mind and thus, why it is central to understand if we are to have an accurate view of those aspects of mind. I’m grateful for your mentioning it.

      On the biased competition stuff, I’ll pick that up in comments to Aaron’s reply to you.

  3. I’d like to thank Carolyn Jennings and Robyn Waller for their excellent commentaries and the organizers of Minds Online for their great work putting this conference together. I’m grateful for the opportunity to present my paper here and to benefit from the careful attention of my commentators. I’ll reply to my commentators in sequence.

    1. JENNINGS
    Jennings argues that my attempt to refute the selection for action account of attention (henceforth, ‘SfA’) is ultimately unsuccessful. It’s important to note, however, that I do not attempt to refute SfA in my paper. SfA is a broad thesis that can be elaborated in more than one way. With this in mind, I raise a challenge for one version of SfA (namely, Wu’s), and note how this challenge generates a constraint on how adequately to develop that thesis. To attack my argument successfully, then, one would need to show that Wu’s version of SfA is defensible as it stands. Rather than trying to show this, Jennings proposes to amend Wu’s view in ways that (I will suggest) are inconsistent with some of his own assumptions and explanatory aims. So, by choosing to respond in the way that she does, I think that Jennings implicitly acknowledges the force of my objection against Wu. Her comments reinforce my point that while the overall SfA view remains viable (and, I think, attractive), Wu’s own development of it faces a problem regarding the agential status of attention.

    Let’s consider Jennings’ proposed distinction between ‘Pure Action’ and ‘Component Action’. As I understand the distinction, Pure Actions and Component Actions differ in whether or not they share parts in common with reflexes. Whereas Component Actions share parts with reflexes, Pure Actions do not. In particular, Jennings’ idea seems to be that the output of a Component Action is, in principle, the same type of behaviour that occurs in a reflex. For example, intentionally raising your arm is a Component Action, because it has a bodily movement (arm rising) as a component—a bodily movement that can be duplicated passively, as when a gust of wind throws up your arm.

    With this distinction in mind, Jennings proposes that selection for action is a Pure Action, whereas most (all?) other actions are Component Actions that have selection for action as a part. Significantly, Jennings adds that only Component Actions require selection for action. She adds this qualification presumably to avoid a regress like the one I discuss in §2. On the view Jennings puts forward on Wu’s behalf, then, every Component action requires selection for action, but selection for action doesn’t itself require selection for action, because selection for action is Pure.

    While I think Jennings’ proposal may have some merit, it goes against one of Wu’s core claims. As I read him, Wu advances a fully general thesis about the difference between action and reflex: action, unlike reflex, requires alternative behavioural possibilities, and therefore action, unlike reflex, requires the agent to select one option over others. These ideas are central to Wu’s views of action and of attention’s role in enabling action. Now, if Pure actions do not require selection for action (presumably since they arise from the need to solve a many-many problem), then Wu’s account of the difference between action and reflex is not adequate as it stands. We are owed an explanation of what makes Pure Action an action if not that that the agent selects it in response to a many-many problem. Wu’s account seems unable to answer this obligation given the fundamental role he gives to the many-many problem in his account of action.

    Given Wu’s commitments, then, it makes more sense for him to pursue the view that selection for action is not an action in any sense (Pure or not), but strictly a component of action. This ‘act-component proposal’ is the view I spend most of the paper considering and arguing against (and which I’ll return to in my reply to Waller). Nevertheless, Jennings’ alternative proposal may be viable if we allow ourselves to depart from the details of Wu’s own account. Indeed, her proposal is similar to the one I begin to sketch in the paper’s conclusion, when I conclude that attention, when an action, isn’t guided by selection for action. To get to that point, though, we must have already rejected Wu’s assumptions that every action (including active attention) requires selection for action and that attention is constitutively post-perceptual. The second assumption requires a bit more explanation, and I will come back to it at the end of my response to Waller.

    2. WALLER
    First I’ll address Waller’s claim that the second prong of my dilemma—the one that begins by assuming that selection for action is a reflex and ends by concluding that action is impossible—is less convincing than the first prong—the one that begins by assuming that selection for action is an action and ends by concluding that action is impossible. This is an interesting suggestion, but it’s worth noting that it doesn’t match the objection that Waller proceeds to develop when elaborating on that suggestion. In elaborating the point, the worry seems instead to be that the dilemma rests on an overly restricted set of options. So, there are two potential objections here: first, that the second prong of the dilemma is less convincing than the first; second, that the terms of the dilemma can be resisted (because of neglected alternatives). I agree with Waller about the second point, so let me discuss that point first.

    As Waller points out, to avoid a vicious regress Wu simply needs selection for action to be a personal level phenomenon other than action. There are many possibilities for what type of entity it might be, and Waller’s concern is that my dilemma falsely assumes that if attention is not an action then it must be a reflex. At that stage in the paper, though, I’m thinking about the options open to Wu on the assumption that attention is a type of behaviour. And on Wu’s view, the genus of behaviour is exhausted by two species: behaviours that presuppose a solution to a many-many problem (actions) and behaviours that do not (reflexes). I show that thinking about attention in either way generates the bad result that action is impossible, and I suggest that the most promising way out of the dilemma is to deny that attention a type of behaviour at all. Then I lift the restriction I had previously been assuming and consider the suggestion that attention is a personal level state or process that is not itself an action—a suggestion I consider under the banner of the ‘act-component proposal’. So, I think Waller and I might agree about the range of options open to Wu for avoiding the dilemma I pose. (It’s my fault, though, for not being clear about which options are under consideration at which points in the paper—I’ll have to fix this). In particular, I agree that the dilemma can be avoided by holding that attention is a personal level state or process other than an action. I’ll come back to that suggestion soon, since it’s a core aim of the paper to explain why I nonetheless think that view turns out to be unsatisfactory.

    Before considering that view, though, let’s consider Waller’s claim that the second horn of my dilemma is weaker than the first. By this, I interpret her to mean that one could rationally accept that selection for action is a reflex and yet resist the conclusion that action is impossible. My discussion of this half of the dilemma is too quick, so let me try to go into more detail now. With it, I intend only to be drawing out a line of thought that I find in Wu (2014: 90; 2011: 102). Wu claims that action would be impossible if attention were a type of ‘sub-personal’ selection. He reasons that if selection were sub-personal, then it wouldn’t be the agent that selects (but a mechanism inside of her), which would be sufficient to disqualify the selected behaviour from being an action. Now, in general, if ϕ is a reflex, then it isn’t the agent that ϕ’s. For example, if an arm-rising is a reflex, then it isn’t the agent that raises the arm. Similarly, if selection is a reflex, then it isn’t the agent that selects. But, as we just saw, on Wu’s view this is enough to disqualify the selected behaviour from being an action. So, the assumption that selection for action is a reflex (together with other assumptions) implies that action is impossible. I’m curious if Waller or others find any fault in this line of thought.

    Let’s turn to the issues Waller raises in the 6th paragraph regarding the act-component proposal. There, she raises the possibility that attention is both an action and a component of other actions (e.g. reaching-to-grasp). Analogy: running one mile is a component of running two miles. In my view, this proposal raises the same question that Jennings’ ‘Pure Action’ proposal raises: what explains why attention gets to be an action? Taking Wu’s account at face value, if attention is an action, then it requires selection for action. So, if the attention that is an act-component of reaching-to-grasp is itself an action, then that occurrence of attention will require a distinct occurrence of attention as act-component. Now, if attention is fundamentally an action, then an infinite regress will ensue. To avoid that result, Wu can say that attention is fundamentally an act-component. This seems to be his considered view. He now thinks attention is fundamentally a psychological state that guides (and partly constitutes) actions, including (as a special case) the action of attending (personal correspondence). In other words, the action of attending (qua psychological process) is simply one of the many types of action that attention (qua psychological state) enables an agent to perform. Like any other action, Wu thinks, attending is guided by attention, because attending requires a solution to the many-many problem and attention is the solution.
    What is attending on Wu’s view? It is the selection of a certain perceptual input for a change of perceptual appearance (2011: 105). In other words, attending is ‘highlighting’, ‘enhancing’, or ‘sharpening’ one’s perception of something. (If I understand it correctly, Jennings’ second amendment to Wu’s view (in terms of a pair of distinctions between ‘input 1’ and ‘input 2’, on the one hand, and between ‘output 1’ and ‘output 2’, on the other hand) is a variant on Wu’s proposal here. The key idea in each cases seems to be to take the ‘output’ of attending to be a perceptual state. If Jennings disagrees with this, perhaps she can tell me what other difference I’ve missed.)

    My reasons for resisting this proposal are: (i) it sacrifices the intuitive thought that attention can be deployed non-instrumentally; and (ii) it conflicts with what I understands to be the theoretical implications of the biased competition theory for the relationship between perception and attention. Regarding (i), I claim in the paper that when you attend for its own sake, you select something without the further aim of doing anything to the thing you’ve selected. By contrast, on Wu’s proposal about attending, you select in order to do something to the thing that you’ve selected (namely, to modify your perception of it). I confess, though, that the intuitions in support of (i) may carry little weight in the end. Wu, at least, is prepared to deny that attention is ever deployed non-instrumentally in the way I suggest it is. Fearing an impasse, I devote more time to defending (ii) in the paper.
    I’ll now discuss my second line of argument against Wu’s proposal about attending: that his proposal conflicts with the theoretical implications of the biased competition theory concerning the relationship between attention and perception. This allows me to address some of the points Waller raises in her final paragraph and to expand on the differences that I see between these views. As I express the point in my paper, Wu’s proposal about attending (as action) entails that attention (as state) guides the process of perceptual enhancement and causes the enhanced perceptual state that is the outcome of that perceptual process. By contrast, I take the biased competition theory to suggest that attention is the outcome of that perceptual process (i.e., the enhanced perceptual state). More specifically, visual attention consists, at the neural level, in the resolution of competitions occurring throughout visual cortex in favour of a bias that is introduced by a top-down signal likely emanating from a distributed network of areas in frontal and parietal cortex. No role is given, on this view, to attention as a causal mechanism for bringing about that outcome. Characterized now at the personal level, attention consists in the achievement of a motivationally structured perceptual state by a guiding motive (e.g., an intention). (If you object that there are cases of ‘bottom-up’ attention to which motivation doesn’t contribute, then let’s limit the discussion to cases of so-called ‘top-down’ attention. In other work I address the ‘top-down’/‘bottom-up’ distinction, and argue that the relationship between motivation and attentional capture is more complex than is usually acknowledged. But this is a long story which isn’t immediately relevant here.)

    The above difference between Wu’s proposal about attending and biased competition theory concerns the causal relationship between attention and a certain perceptual process (specifically, whether attention is cause or the effect of that process). Another difference between the two views (which I don’t go into in the paper) is in the relationship each identifies between a state of attention and a process of attending. Wu’s view is that states of attention enable acts of attending. Specifically, a state of attention is the component of the process of attending that manifests the agent’s guidance over the process. By contrast, on the view I find support for in the biased competition theory, visual attention does not enable the process of visually attending. The relation of dependence seems instead the reverse: a state of attention is what the agent achieves as a result of attending, where attending is the causal process consisting in the guidance of perceptual processing by the frontal-parietal network. Attention is not the guiding component of that process, but its end state. The guiding component of the process is instead the agent’s current intention (or other motivational state).

    Noting that there are other ways to interpret biased competition theory, Waller asks whether Wu might not prefer to adopt a different interpretation of that theory to the one I give in the paper, and Jennings raises a similar question in e-mail correspondence. The first thing to note is that Wu and I are broadly in agreement about the right way to interpret the biased competition theory (see Wu 2014: 59-60; forthcoming). Indeed, my interpretation of the theory has been influenced by reading Wu on this topic (as well as others who have pursued this interpretive question, including Allport (2011), Mole (2011), and Stinson (2009)). To adopt some terminology that originates from William James, we are both sympathetic to an ‘effect’ theoretic interpretation (according to which visual attention is correctly identified with the outcome of biased competition), not to a ‘cause’ theoretic interpretation (on which attention is identified with that which causes biased competition). On the effect-theoretic interpretation, the distributed network of areas in the frontal and parietal cortices that bias competition for processing resources throughout visual cortex does not realize attention (as it does on the cause-theoretic interpretation), but is instead among the control structures responsible for directing attention, where attention itself is realized in visual areas. One of my aims is thus to show that if one accepts the effect theoretic construal of the biased competition theory (as Wu himself seems to do), then one must view the relationship between attention and perception (and, I would now add, the relationship between ‘attention’ and ‘attending’) otherwise than how Wu’s account officially describes it, and, moreover, in a way that makes the intuitive notion of attending ‘for its own sake’ totally unproblematic.

    I must acknowledge that I don’t have a knock-down argument for the effect-theoretic construal of biased competition theory that I favour (at present, anyway). Two potential advantages of the effect theory over the cause theory, though, are theoretical economy and demystification. The effect interpretation is economical, because once you know about the subject’s control states (e.g., motivational states like intention) and how those states interact with the visual system, you have all you need to understand visual attention. Visually attending, at the neural level, may be nothing more than the way control structures in the frontal-parietal network (and potentially other motivational structures, such as the amygdala, see Pessoa (2013)) come to guide perceptual processing by modulating the distribution of processing resource in favour of motivationally relevant sources of information. Attention is not a further psychological mechanism or capacity apart from motivation and perception, as it is on the cause theory, but simply the outcome of their interaction. Moreover, whereas the effect theoretic interpretation has the potential to explain attentional selectivity, cause theories have tended to conflate questions about attentional control with questions about attention, thereby giving rise to the appearance of a mysterious inner agent that doesn’t add to our understanding of the mind (cf. Allport 2011; Stinson 2009). I acknowledge that each of these points requires further elaboration and defence, but hopefully it helps to show why I find the effect-theoretic interpretation more attractive than the cause interpretation.

    Bibliography
    Allport, Alan. 2011. “Attention and Integration.” In Attention: Philosophical and Psychological Essays, edited by Christopher Mole, Declan Smithies and Wayne Wu, 24-59. New York: Oxford University Press.
    Mole, Christopher. 2011. Attention is Cognitive Unison: An Essay in Philosophical Psychology. New York: Oxford University Press.
    Pessoa, Luiz. 2013. The Cognitive-Emotional Brain: From Interactions to Integration. Cambridge, MA: MIT press.
    Stinson, Catherine. 2009. “Searching for the Source of Executive Attention.” Psyche 15 (1): 137-154.
    Wu, Wayne. 2011. “Attention as Selection for Action.” In Attention: Philosophical and Psychological Essays, edited by Christopher Mole, Declan Smithies and Wayne Wu, 97-116. New York: Oxford University Press.
    ———. 2014. Attention. New York: Routledge.
    ———. forthcoming. “Shaking Up the Mind’s Ground Floor: The Cognitive Penetration of Visual Attention.”

    1. Hi Aaron: Thanks again for your continued engagement with the issues. Just two points (see comment below for the first substantive reply). On attending non-instrumentally, as you note, I am just inclined to deny that. You want such attending to be an action but I just don’t see that there are such actions that aren’t themselves coupled to a clear output, say moving one’s eyes to track something (or generally, one’s body), committing things to memory, or sharpening perceptual representations (or otherwise altering them). If attending isn’t that, then I’m not sure it’s anything action wise. It’s not even clear how you would learn to do such a thing without a clear output that one could track. But as you say, we might just differ on intuitions.

      On the biased competition point, I think I see what your worry is. Given the distinctions I draw below, you would say that on my picture, attention is still a cause since it would be an input that guides the response. And if the response here, let’s say, is sharpening representations, then attention is causing a specific change in perception where, you also say, it’s the change that we would call attention and not the input that leads to the change.

      I would say that we can all (and should) agree with the biased competition account and deny that attention is a spotlight, the thing that makes the transition from input 1 to 2 in Carolyn’s diagram (actually, I’m not sure what Carolyn would say on that point from her subject centered theory of attention, but she can weigh in). I think that’s all we should read the biased competition account as committed to: attention is tied to resolution of competition, but not as its cause. As you know, I’ve argued that the signal that resolves competition should be from intention, and that’s why I say that the spotlight role is taken not by attention but intention. But after that, I think everything is up for grabs, so I don’t see that what I say is particularly at odds with Desimone and Duncan.

      That said, as people point out, lots of things get called “attention”. You want to identify the end state as attention, the perceptual state that is altered by resolution of competition. I think it’s the noticeable part of attending as action, but the attentional state in once sense remains the input. Perhaps we are inclined to speak of the output as attentional. At that point, I think it’s more nomenclature than a substantive issue, and the SfA view can allow for that nomenclature even if it sticks to its core point.

      1. Wayne, thanks so much for your comments throughout. It’s helpful for making sure I don’t inadvertently mischaracterize your views, and it’s great to have you contributing to the discussion. I can only scratch the surface of your comments at the moment, so let me jump to the place where I feel some confusion and I’m tempted to resist.

        While you’ve clarified many things for me, I’m now wondering if I understand what you take active attention (or attending) to be. I’ve been relying here on your remarks in the 2011 paper, where you invoke Carrasco’s results as part of a proposal about what it is you’re doing when you intentionally shift and maintain visual attention to something. Going by Carolyn’s image, I took your view to be that attending consists in selecting an input 1 for the purpose making it an input 2/output 2, where input 2/output 2 is some enhanced/sharpened visual representation. As you’ve made clear in one of your other comments, you identify attention (/selection for action) with the guiding perceptual input, which in this case is the subject’s attention is input 1. The action of attending, then, consists in the process of turning it into input 2, where attention (input 1) is what guides that perceptual process.

        Am I right to interpret you this way? Your remarks in the above comment make me second guess that interpretation, because you say “I would say that we can all (and should) agree with the biased competition account and deny that attention is … the thing that makes the transition from input 1 to 2 in Carolyn’s diagram … I think that’s all we should read the biased competition account as committed to: attention is tied to resolution of competition, but not as its cause. As you know, I’ve argued that the signal that resolves competition should be from intention, and that’s why I say that the spotlight role is taken not by attention but intention.”

        But if the perceptual process that consists in the transition from input 1 to input 2/output 2 (which we earlier agreed to call ‘attending’) is achieved simply by the guidance of an intention and not by attention, then we have a case of action (attending) achieved without guidance by attention (/the agent’s selection). You say this is possibly a matter of nomenclature, but I’m not yet convinced that it is. One of my main aims in section 6 was to show that the transition from input 1 to input 2 is a motivationally guided process, which given that the cause theory is false, isn’t achieved by attention. And I though we also agreed that it was an action. If so, then what’s your account of this action, and is it an exception to the universal claim that every action is guided by attention as a component?

        1. One more quote (this time from your reply to Carolyn’s commentary) that gives me pause: “In your [Carolyn’s] scheme, I would say that attention should be identified with input 2 and not input 1 as its field of operation, or even the transition from input 1 to 2. That’s too much the spotlight model, attention as a selector.”

          Again, this is not what I took your view to be. I think the view you express sympathy for in the above quote suggests a rather different view of the relationship between attention and perception than the one arising from your discussion of the many-many problem, and the one that is made explicit in your (2011) account of active attention as a process of perceptual enhancement. It is more like the view I sketch in the conclusion of my paper.

          1. Hi Aaron: First, it is almost certain that I have not been not consistent across the papers/books. I would take what I say in my 2014, chapter 3 to supercede anything from the 2011 paper, so where they conflict, I would side with the 2014 paper. The 2011 paper was written well before the publication date so would be a very early view of the account. It’s perfectly reasonable to quote and rely on it, but if there are inconsistencies between the 2011 and 2014, they should be acknowledged as changes of view or infelicities in the earlier paper (assuming I do better in the later book!).

            Let me pull away from my attempts to read my view onto Carolyn’s diagram. It was my attempt to map the view on her representation in this context, but it might be best to rely on my own depictions.

            In the diagrams to illustrate the Many-Many Problem, there isn’t a “filter” per se, as is implied by Carolyn’s diagram (from input 1 to 2). Each input corresponds to a sensory state, it’s just that only one or a subset of those states get connected to action (Carolyn might not have intended to suggest a filter). So, the transition, if you like, is between being in various sensory states, that we individuate by their intentional properties such that we pick one out as the state that fulfills a further functional role of informing the action. My view isn’t that attention does any of that selection (I guess it would seem like a shift from input 1 to 2 on Carolyn’s diagram which is what I was riffing on). Given an intention, one of those states comes to play a functional role, namely guiding action. When that sensory state does so, then that is when we have attention (selection for response) in that modality. So if it’s the visual state directed at a soccer ball that guides my kicking it (the kicking movement), then that visual state constitutes one’s visual attention to the soccer ball. In that sense, it’s a transition from input 1 to 2, but the other inputs (sensory states) might still be there. I don’t know if that helps. I’m sorry the writings haven’t been clearer on this.

            You write: “But if the perceptual process that consists in the transition from input 1 to input 2/output 2 (which we earlier agreed to call ‘attending’) is achieved simply by the guidance of an intention and not by attention, then we have a case of action (attending) achieved without guidance by attention (/the agent’s selection).” I guess I would disagree. The intention promotes one sensory state as functionally critical, the guidance of the response which we are imagining amounts to the sharpening of the corresponding sensory representation. So, the sensory state input informs the output, namely the sharpening of that state’s representation of its target. So, attention still informs the generation of a response, as part of the act of attending for its own sake (admittedly, you think this is still instrumental, but I don’t see what non-instrumental attention could be).

            I don’t know if that helps. My words might not be adequate, though I have the same diagram in my head that captures the view.

            1. Hi Wayne,

              Thanks, and my apologies if I’ve been holding you to views that you no longer endorse. I’ll happily let go of the 2011 discussion and regard the 2014 discussion as superceding it. (Note, though, that as recently as your 2014 manuscript (chapter 4), you say things that are reminiscent of the 2011 proposal about active attention. You say that changes in perceptual appearance of the sort that Carrasco and document aren’t constitutive of attention’s phenomenology. Rather, those changes in the subject’s visual state are attention’s effects on visual phenomenology. This is rather like saying that those changes in perceptual appearance are outputs/responses of which attention is the cause. So, I think I find the same assumption in your 2014 discussion that I am drawing from your 2011 discussion of active attention. What they share is the claim that attention brings about an altered visual appearance, and hence causally affects one’s perceptual state. The relevant contrast is with a view that regards the altered appearance as partly constitutive of a state of attention. I understand that your 2014 book is a bit of a moving target, in that you are also constantly presenting other peoples’ views about attention. But I think you were endorsing this claim as part of your overall case against there being a phenomenology of attention).

              I’m also sorry if I’m being a bit thick. I’m trying to wrap my head around your view of active attention. This is where I’ve got to so far (perhaps you can correct me when I go wrong). On your account, attention is the input state that plays a certain functional role, namely that of informing a response. Attention is not the mechanism that selects an input to play that functional role (ala spotlight/cause theory). Action is the process that consists in the guidance of a response by a certain input state (i.e., by attention). A special case of action, you allow, is when the subject selects a certain visual input for sharpening. This serves as your account of active attention or attending. This action consists in the transition from the original unsharpened input to a sharpened input. (This corresponds to the case in which you alter its appearance, rather than maintain that alteration). It seems to follow that you identify attention with the input state that guides the process of sharpening.

              Now, following BCT, you hold that attention is the visual state that results from an intention-biased competition (at least in the case of top-down attention). I then add that there is good reason to think that this visual state will also be a ‘sharpened’ representation (or one that is otherwise enhanced in task-relevant ways). That is, I think there is good reason to think that the neural basis of the sort of perceptual changes that Carrasco and others document are realized by the sorts of neural changes that occur a consequence of biased competition (e.g., receptive field remapping, gain amplification, noise reduction, and other types of signal enhancement). You seem prepared to grant me this (though I acknowledge that it’s an empirical question, which I would need to argue for). Note that all of this is consistent with saying, as you wish to say, that the resolution of biased competition coincides with a change in the functional profile of one’s input state, such as whether it informs a response and, if so, which response it informs. I’m adding to this that it will also coincide with other sorts of changes, such as changes in how things perceptually appear to you (along the lines of ‘sharpening’). Bringing things back to the input 1/input 2 distinction, I can accept that the transition from input 1 to input 2 partly corresponds to a change in functional role, such as whether the input is informing a response, but I’m adding that it also corresponds to a change in perceptual appearance.

              Here’s where I start having trouble. If you accept the above characterization of BCT, then I don’t see how you can give an account of active attention as consisting in the process of sharpening, i.e., as the process that consists in the transition from an unsharpened state to a sharpened state. This is because that account of active attention entails that attention is the state that temporally precedes, and informs, the generation of the sharpened representation. But if things are as BCT says they are, then the state of attention (the resolution of biased competition) is already sharpened. So, sharpening it isn’t a behavioural possibility available to the agent. It has already been structured by the agent’s intention.

              Even setting aside our opposing intuitions about the idea of non-instrumental attention, then, I think this is a good reason to think that the action of attending cannot be as you describe it to be. It must be some other response that attention informs or guides. The question that will arise for any candidate response is the same: in virtue of what similarity between the response that you select and your selection of it warrants classifying both as an instance of attention? I think it will be hard to find an adequate answer to this given your purely functional characterization of attention as simply that input state that informs a response.

              That’s my attempt both to clarify what I understand your view to be and where I see the problem for your account as lying. (And my apologies for being repetitive!)

              1. Hi Aaron:

                I haven’t read the 2011 piece in a while, perhaps 6 years, so I’m not sure if I would agree with everything I wrote there. Where there’s consistency across time, then for sure, I must mean it 🙂 Otherwise, if there’s inconsistency, I suspect I would opt for the 2014 version. But if I am consistent about conscious attention, then good (thank goodness for some continuity!). By the way, I think your summary of the view is very accurate (in the above comment).

                I think our disagreement here depends on reading a bit into the biased competition model which works best, details wise, as an explanation of neural data, e.g. single neuron responses or population responses, at least as I understand it. It’s a further leap to jump to claims about phenomenology or even mental states specified by a functional role. So, in the 2014 book, I try to make the argument about parallels to functional role by taking a Marrian line with respect to selection for response (action). The Many-Many Problem (MMP) illustrates, at least in typical action, that some selection is entailed by action. This idea of subject selection for task is part of psychological/neurobiological experimental practice, so the issue then is about the realization of the subjective/mental selective states assumed by cog scientists to be present during task performance. So, the biased competition account serves that purpose. That’s the best argument I can think of linking that model (and its implementation in divisive normalization accounts) of neural computation to mentality. Marisa Carrasco has connected this to the sharpened representations she argues occurs with attention.

                Your argument is that the intention based influence leads to the sharpening explained by biased competition and that this means that the input state is already sharpened. I think this is not clear. Selection of an object relative to other objects, whose neural correlate is receptive field remapping, can yield a state that informs then sharpening of contrast or spatial resolution, whose correlate would be sharpening of tuning curves. Those would be different effects, right?

                But let’s assume you’re right, that the intention shifts processing so that we get sharpening of contrast and not just selection of one object rather than another. Call the sharpened experience “phenomenal attention”. Now there’s a difference between maintaining phenomena attention and coming to attend in that way. On my view, that’s the difference between the input already being sharpened versus the input not being sharpened but coming to be that way. The case you are imagining where intention sharpens a visual state then has the following structure: a visual state at coarser resolution coming to be of finer resolution and not others. Otherwise, there’s no act of coming to attend as opposed to maintaining attention. This coming to attend is precisely diagramed in the way I diagram every action, many inputs with one coupling to a specific output. When I attend in the sense at issue here, a specific visual state was shifted from among many others. To deny that would be to deny the distinction between maintaining phenomenal attention and coming to phenomenally attend.

                Maybe what drives the worry is that intention here is still being treated too much like a spotlight, that it shines its beam down on the input to change it, i.e. make it functionally linked to informing response. But I would think of intention’s role first as just “setting the weights” so that a coupling between input and output occurs, and in that coupling there is the relevant sharpening of representations, the coming to phenomenally attend. When the coupling occurs, then we can speak of a state’s having been selected. But I don’t see anything in biased competition that is inconsistent with this picture.

                I have no idea if this helps or just clouds the issues. I hope the former!

                1. Thanks, Wayne. This is becoming more challenging! I’ll just try to make two points here. First I want to clarify where I’ve been trying to be careful and precise, and where I haven’t been. Then I’ll try to respond to a neat challenge you’re raising in the second part of your comment, which I’m still thinking through.

                  First off, you write:
                  “Your argument is that the intention based influence leads to the sharpening explained by biased competition and that this means that the input state is already sharpened. I think this is not clear. Selection of an object relative to other objects, whose neural correlate is receptive field remapping, can yield a state that informs then sharpening of contrast or spatial resolution, whose correlate would be sharpening of tuning curves. Those would be different effects, right?”

                  Right, but I need to acknowledge here that I’m not being that precise in how I use ‘sharpening’. I certainly didn’t mean to commit myself to a specific effect, such as tuning curve sharpening. I’m using the words ‘sharpening’, ‘highlight’, and ‘enhancing’ as placeholders for the contribution that attention (variously) makes to one’s perceptual state—or, to avoid sounding to spotlight theoretic, the attentional structure of perception (but with no commitment to Sebastian’s account of that structure). Admittedly, I specifically discuss receptive field remapping in the paper, but this was meant to be only an example of the kind of transformation that can occur through biased competition. My understanding (correct me if I’m mistaken) is that many such effects can be achieved through biased competition. I’m being fairly rough here insofar as I’m ignoring potentially differences among them and the exact contribution that each of them make to the subject’s perceptual state. My main point is that there are a host of neural effects—e.g., receptive field remapping, increases in contrast and response gain, etc.—which bear a suggestive relationship to the changes in perceptual experience that psychologists are now researching, and researchers have made specific proposals about the specific neural changes that likely underlie and explain the changes in perceptual experience that accompany attention. E.g., if biased competition generates an increase in contrast gain in a visual neuron then, from the visual neuron’s point of view, it will be as if the stimulus itself increased its contrast. This effect bears an intelligible relationship to boost in apparent contrast that purportedly accompanies some shifts of attention. (So, researchers have defended ‘contrast gain models’ to explain that particular perceptual effect). That’s the kind of explanation, in very broad strokes, I’m gesturing toward, together with the claim that such neural effects are the outcome of a process of biased competition. If you’re worried I’m being too rough and ready, though, I can work on being more precise about specific mechanisms and specific places where I’ve seen these arguments made (Carrasco 2011 cites a number of relevant studies and computational models). At some point I know that the devil is in the details, and it’s important to be very careful in how we draw on empirical literature.

                  Next, you raise an interesting worry that, if things are I describe them as being, I won’t be able to explain shifts of phenomenal attention, but only the maintenance of phenomenal attention. This is an interesting challenge I’ll need to consider in more detail. My response for now is that the shift of attention is, like any episode of attention, the result of top-down feedback, likely from a frontal-parietal network, which biases processing for a certain part of the visual field (say, a certain location) and inhibits the processing of others. Rather than denying the distinction between shifting and maintaining phenomenal attention, what I’m denying is the assumption that shifting phenomenal attention requires selecting the location for a shift of phenomenal attention. While it might be true, in general, that actions require selecting a location or an object to be the target for some intervention, I think attending is unlike other actions in this regard. You don’t have to target something to shift attention to it, because to target it would have already be to have attended to it (and hence to have already influenced the outcome of competition). So, I’m resisting the move you want to make which assimilates this particular action to other cases of action in which attention is needed to do the target-setting. There is no target-setting for acts of attending. That’s the view.

                  I’ll certainly give the question more thought.

                  As for your closing suggestion about where I might be worried, I don’t think that’s quite the source of the worry. I’m happy with the overall move of replacing the attention spotlight with the intention spotlight, though intentions raise many questions of their own! (As far as the project of demystifying the mind goes, shifting the burden onto intention can seem a bit like pushing the bump to another corner of the carpet. But we need intentions to explain other things, so it still seems like explanatory progress is made by invoking intention as the top-down component of attention and I’m friendly to the proposal).

      2. “I would say that we can all (and should) agree with the biased competition account and deny that attention is a spotlight, the thing that makes the transition from input 1 to 2 in Carolyn’s diagram (actually, I’m not sure what Carolyn would say on that point from her subject centered theory of attention, but she can weigh in). I think that’s all we should read the biased competition account as committed to: attention is tied to resolution of competition, but not as its cause. As you know, I’ve argued that the signal that resolves competition should be from intention, and that’s why I say that the spotlight role is taken not by attention but intention. But after that, I think everything is up for grabs, so I don’t see that what I say is particularly at odds with Desimone and Duncan.”

        On my subject-centered account, attention is a process of prioritizing mental/neural events that is directed by a subject. In the sense that attention is a process that brings about a change to prioritization, it can be thought of as a cause, but not in the spotlight sense. The only place I disagree with Desimone and Duncan is in the insistence that the data resolve the question of whether attention simply emerges bottom-up or is subject to top-down control. I don’t think they do.

        On the diagram: this is intended to help make the point about inputs as clear as possible but, no, I didn’t mean to commit anyone to the idea that attention is a filter (I know that isn’t Wayne’s view!). That is just the drawback of diagrams (or my diagrams!). If I were creative enough, I would have thought of a better way of representing the matching relation between intention and sensory input, but…

        1. Hi Carolyn!

          I was curious to hear more about your take on the Desimone & Duncan (1995) position. In your comment above, it sounds a bit like you regard the top-down/bottom-up distinction as mapping onto the spotlight/no-spotlight distinction (or what I’ve been calling the ’cause theory’/’effect theory’ distinction). Do you see these distinctions as being related?

          I ask because my understanding was that Desimone & Duncan (as well as other proponents of BCT) allow that attention can be controlled both top-down and bottom-up, but that in either case attention is an ’emergent’ result or outcome of biased competition as opposed to its cause. The top-down/bottom-up distinction concerns the source of the bias signal (whether something like one’s executive control system or something like stimulus salience).

          1. Well, “top-down” is relative. Their model clearly allows for top-down influences. What I object to is this, in the conclusion: “attention is an emergent property of slow, competitive interactions that work in parallel across the visual field.” So, there may be some top-down influences and some bottom-up influences, but at the most general, “top” level it is bottom-up influences that determine the result (e.g. competition is resolved bottom-up). I don’t think they need to claim this, and they do allow that it is controversial. I think it is needlessly so. To me the spotlight issue is separate, as it is mixed up with issues of spatial localization.

            1. Interesting. The ‘spotlight’ label might be less than ideal, because you’re right that traditionally spotlights have to do with spatial attention. But I think that what Wayne intends with the expression isn’t specifically about spatial attention. (Correct me if I’m wrong, Wayne). It has rather to do with a conception of attention as a supra-modal mechanism that is separate from sensory-motor circuits and that influences the processing that goes on in those systems (cf. Posner 1980). That, anyway, is what I mean to refer to when I speak of a ’cause’ theory in this context. The alternative is a view like the one D&D seem to endorse, on which attention is realized in sensory systems. The latter type of view is consistent with the claim that non-sensory systems act on the sensory system (which is what happens whenever there is a top-down bias signal that influences sensory processing), but not with the claim that attention is one of those non-sensory systems. So: I take the effect theory (anti-spotlight) version of BCT to be the view that the resolution of competitive interactions occurring throughout visual areas realizes or is constitutive of visual attention. I think that’s what the authors are saying in the passage you site. This is consistent with saying that the bias influencing the outcome of that competition can be either bottom-up (paradigm: stimulus salience) or top-down (paradigm: the agent’s current goal). I don’t take them to be in any sense privileging bottom-up control.

  4. Thanks Aaron! I’m wondering if you can help clarify something for me (and apologies if they is a naive question). To avoid a vicious regress, what’s needed is that not all attention is an action. To account for the apparently non-instrumental character of active attention, what’s needed is that some attention is an action.

    But those aren’t inconsistent. What’s to stop Wu from simply saying that, while all action requires prior attention, some but not all attention is itself an action? When attention is an action it would then require prior attention, but that prior attention need not itself be an action.

    I had a related question about the biased competition issue. The key idea there seems to be that attention is sometimes a constitutive determinant of ‘how things seem’, i.e. of what will be the input to the next round of selection. It couldn’t tell if you thought this was incompatible with Wu’s theory from the get-go, or if you thought it was only incompatible with the ACP reply to your dilemma.

    (You seem to suggest the first when you say “the sense in which attention can ‘determine’ input on Wu’s view is causal rather than constitutive. To see why it cannot be constitutive, it is sufficient to reflect on [the fact that]…This solution is intelligible only if which inputs you receive at a time is independent of how you select at that time. This suggests that, on Wu’s view, inputs are constitutively independent of attention.”)

    If the problem is that attention can only select things which are not themselves products of attention, and so if attention is constitutive of inputs then it can’t also be selection for action, then the ACP seems unnecessary to pose the problem. The ACP is important, though, if the point is that Wu posits attention as merely causally affecting inputs in order to account for active attention in a way compatible with the ACP. But it wasn’t always clear to me which you had in mind.

    Whichever of these two problems was intended, I was wondering if you could clarify why (as with my first point, above), it’s not acceptable for Wu to simply say that attention can work differently on different occasions. Sometimes it causally affects inputs, sometimes it constitutively affects them, sometimes it operates on previously given inputs – there’s just lots of different sorts of attending going on. I feel perhaps I’m missing something about why that sort of view isn’t tenable?

    1. Thanks Luke, those are great (and not so naïve) questions.

      You’re right that to avoid a regress what’s needed is that not all attention is an action, and you’re right that one could therefore say that, while not all attention is an action, some attention is an action and that when attention is an action it requires prior attention. This allows one to preserve the claim that every action requires selection for action, including active attention. This seems to be Wu’s considered view, and it’s the position I discuss in section 5 when I consider his proposal about active attention. The claim there is that active attention (attending) is an action that, like any action, requires selection for action and is guided by that selection.

      My reasons in the paper for resisting that view are twofold: (i) it sacrifices the intuitive thought that attention can be deployed non-instrumentally; and (ii) it conflicts with what I take to be the theoretical implications of the biased competition theory for understanding the relationship between perception and attention. Regarding (i), my thought is that if attention is sometimes an action, then it ought to be a situation in which you select something without any further aim of acting on it (beyond having selected it). That’s what it would be (I suggest) for attention to be an action in its own right. But the above proposal cannot accommodate this possibility.

      To see this, begin by considering how Wu develops the proposal. He suggests that when you actively attend, what you do is select an input for some perceptual modification (‘highlighting’). My concern here is that this still depicts attention (selection for action) as instrumentally related to an end other than itself: the agent selects a certain input in order to do something else to it (namely, to highlight it). We might call the action “attending”, but selection for action is still being used to serve an action that is distinct from one’s selection of it. So, it doesn’t fulfil the criteria that I advance for an adequate account of attention as action.

      Moreover, while I don’t go into this point in the paper, we can also ask: what property do attention or selection for action (qua act-component) and attending (qua action) share, such that it is correct to classify each as species of attention? Why should we single out that action in particular as constitutive of attending? At least on Wu’s view, there is no essential or necessary connection between selection for action and the sort of perceptual changes that Carrasco and colleagues document. To see this, observe that, for Wu, those changes in perceptual appearance are among attention’s effects on visual experience and hence not features of attention itself (Wu 2014, 115, 140). Wu uses this point to argue that attention-induced alterations in perceptual appearances fail to characterize the phenomenology of attention, and are instead among attention’s effects on visual phenomenology. Though our topic here isn’t the phenomenology of attention specifically, Wu’s remarks about the phenomenology of attention are still relevant. If alterations in perceptual appearance are merely attention’s effects on visual experience and so unsuited to characterize attentional phenomenology, then they appear equally unsuited to figure in an account of what it is to actively attend. The reason, in both cases, is that the perceptual modifications are inessential and extrinsic to the agent’s selection for action. It thus seems arbitrary to single out such changes as somehow constitutive of attending.

      Before turning to your second question (about ACP and BCT), let me finish answering this question by jumping to your concluding thought about there being different sorts of attention or attending. This seems like a fair suggestion, but it’s worth bearing in mind the explanatory ambitions of the selection for action account. It attempts to articulate the essence or nature of attention, and sets itself the aim of accounting for every well-documented case of attention. The account is motivated to a large extent by the underlying unity it purports to uncover across a seemingly disparate range of cases. Moreover, the view seems sufficiently abstract or general to do quite well on this score. It can explain, for example, perceptually guided bodily action (e.g., reaching to grasp something), perceptually guided mental action (e.g., perceptually based demonstrative thought), memory guided bodily action (e.g., reaching for a light switch in the dark), memory guided mental action (e.g., reasoning on the basis of background knowledge). I think it can explain the differences among feature-based, object-based, and spatial attention (in the domain of perceptual attention), and it has something to say about the different ways that attention gets deployed (whether in an intention-guided or automatic fashion). It does all this by appealing to a single, underlying concept of attention: selection for action. Given this theoretical background, it seems inadvisable to allow too much diversity in the kinds of attention we allow. And, at least for Wu’s purposes, it arguably matters that every episode of attention be explainable by reference to the many-many problem, which clearly invokes a domain of already constituted inputs from which the agent then selects (including as a special case, the action of attending). If you depart too far from that structure, you’ve lost one of the central pieces of the view.

      I’m less sure what I want to say in response to your second question. I suspect that you may be right that there are actually two objections in my paper, only one of which relies on the act-component proposal. And I do think that what you call the ‘key idea’ behind (my reading of) BCT is inconsistent with the basic structure of the many-many problem (as you say, from the get-go), because BCT takes attention to be constitutive determinant of one’s input state, whereas the many-many problem takes attention to operate over a pre-attentively constituted input state. So, I’ll need to think more about this for the moment. But this particular reply is getting long, so I’ll leave it at that for now! Thanks again for your insightful questions.

        1. No problem. I was going to apologize for the same thing! (I think you posted your comment before I posted my response to Luke). I’m thinking about your question now and will reply shortly.

      1. OK, I’m back with more, Luke. You’ve raised an important question about how to interpret my argument. Here’s where I’ve gotten so far to with clarifying it, though I still need to work on it. First, I think you’re right that there’s a conflict right from the start between the basic assumptions behind Wu’s many-many problem—according to which the agent is presented with a set of pre-attentively constituted inputs from which she must then select—and the view that I find support for in BCT—according to which attention is a constitutive determinant of the subject’s input state (attention being the achievement of a motivationally structured perceptual state). From that perspective, the argument could bypass the act-component proposal. But I also think that the act-component proposal is: (a) Wu’s considered view, and (b) the only coherent way to develop his view while also maintaining what Sebastian below calls the universal claim (that every action has attention as a component part), given the dilemma that I raise early in the paper. I also think that the basic assumptions behind the many-many problem (that I think conflict with BCT) become especially obvious only when we examine Wu’s attempt, within the context of ACP, to accommodate the possibility of active attention, since that proposal makes it clear that although attention can causally affect which inputs one later receives, it is never a constitutive determinant of an input state (qua input). It also helps to bring out the essentially instrumental nature of selection for action on his view, which precludes the possibility of attention ‘for its own sake’. So, by discussing the dilemma and the ACP, I’m trying to isolate the specific commitments of Wu’s view that I take to conflict with BCT, and I’m trying to show how it is those commitments that prevent one from acknowledging the possibility of attending as an action in its own right.

        I hope that helps!

        1. Thanks Aaron, that does help a lot! And thanks Sebastian for developing some of my ideas better than I could have!

          I can see that the worry about unity is one reason to resist a proliferation of different sorts of attention: what do they have in common? I was wondering in my original comment if that was the main issue for you, or if the sort of thing that Sebastian brings up below (that it’s just implausible that we attend ‘twice’ whenever we actively attend) was more at work. Since you expressed a lack of sympathy with that argument, I suppose I have my answer!

          I wonder though if there’s a way to answer these two challenges – though I’m here freely departing from Wu’s view and simply looking at a different way to develop ‘attention-as-selection-for-action’. I certainly did not mean, in my original comment, to be talking about Wu’s specific idea that active attention is an instrumental attempt to sharpen perceptions or anything – I agree with you that attending can sometimes be non-instrumental, at least on the face of it. But thinking that active attention is sometimes non-instrumental seems consistent with also thinking that it requires prior attention.

          You might answer Sebastian’s worry by noting that for someone to voluntarily shift their attention to something, they need to already be aware of it as a potential target of attention, and this isn’t guaranteed just by it impinging on the sense organs. There might be situations where someone’s failure to actively attend to X is explained by X failing to be sufficiently salient that they noticed it as a potential target for attention, even though it was being processed by the brain. So you think that the prior, non-active attention is just the ‘making-a-little-bit-salient’ of the object which allows it to be selected for active attention. And since that making-a-little-bit-salient involves selecting particular things from amongst many inputs, it is itself a way of solving a version of the many-many problem and thus deserves to be called ‘attention’ (addressing your worry).

          Perhaps that’s an overly-broad reading of what counts as a many-many problem, or perhaps it’s not agent-y enough in some sense. But that’s roughly what I had in mind when I first read the paper, and while it’s not Wu’s view it’s at least somewhere in the vicinity perhaps?

          1. Hi Luke,

            Regarding what the main issue is for me, I would say that it’s partly the unity issue. but partly also the suggestion that, as you put it, when we actively attend, there are two instances of attention (one an action and one an act-component of that action). What I was unsympathetic to in Sebastian’s proposal was simply the assumption that the relevant way to think about the act-component of active attention is in terms of passive or involuntary attention. I hadn’t considered that possibility until he raised it, but the active/passive distinction doesn’t seem to map onto the action/act-component distinction. But since you’re pursuing the idea here, I’ll explore the idea further.

            First off, to clarify: by ‘Sebastian’s worry’, do you mean that if active attention has non-active attention as a component, then the non-active attention is presumably involuntary, and, as Sebastian suggests, there is no evidence to suggest that episodes of active attention are enabled by involuntary attention?

            Let’s look at your proposal more closely: active attention (i.e., the action of attending) presupposes attention in a non-active sense, the latter being the act-component of active attention. Now, your proposal about non-active attention invokes the notion of salience. Salience is a troublesome notion, though, so we need to be careful with how we use it: people think about the relationship between attention and salience in multiple, sometimes conflicting ways. On one usage, salience is the result of attention, such that, once you attend to something, it becomes salient for you. It stands out, is foregrounded, is central, etc. (I usually prefer to call this this ‘prominence’.) On another usage, salience is what automatically drives or captures attention. When you automatically or involuntarily orient your attention to something, on this view, you are reacting to its salience.

            So, one question for your proposal is which notion of salience you’re working with. On the one hand, you seem to assume that if something is salient for a subject, that’s because she has made it salient by giving her attention to it. This figures in your account of non-active attention. It seems to me, though, that a more natural account of non-active attention would be framed in terms of the second notion of salience, on which you are drawn to attend to something because it pulls on or calls out for attention. It’s possible, though, that anything which pulls or calls out for attention to some extent succeeds in receiving it. So, that could support your idea that the first non-active stage of attention makes things a little bit salient in the first sense of the term (i.e., prominent).

            You also suggest that the experience of salience enables a more active form of attention. One possibility here is that the experience of salience enables active attention by motivating it. As a result of feeling the pull to attend, you respond by actively reorienting attention to the salient thing. Maybe this is still too passive, though, as it suggests that active attention is always an impulsive reaction to some urge to shift, and intuitively not all instances of attention are like that. Another possibility, which is probably closer to what you have in mind, is that the little bit of prominence things have in virtue of the non-active stage of attention enables active attention by making something a potential target for attention, but without necessarily motivating the attention shift in the way I just described. It simply suffices to make it “noticed” and hence something that you could deliberately shift your attention toward. That’s an interesting suggestion.

            Anyway, I’m happy to continue talking about this, both because I’m curious to hear more about how you think this helps with the many-many problem, and because I got a bit caught up here in the details of your proposal and may have lost the forest for the trees. The idea that salience could play a role in solving the many-many problem is interesting, and it relates to ideas that were actually explored in the Gestalt psychological tradition (that in turn influenced J.J. Gibson’s thinking about affordances). Some of the authors in that tradition were explicitly interested in why only some of the seemingly countless action opportunities in the environment present themselves to the subject for action, and what it is that guides the particular course of action an agent ends up taking. Ideas like salience (and ‘valence’) seem to have been in the air at the time as a way of explaining why the agent ends up selecting one course of action rather than any of the countless others the environment presents her with). In this sense, I think the many-many problem and its connections with salience and attention have been around for some time.

            1. Thanks Aaron, you’ve developed my idea further than I did! You’re right that I was operating with an idea of ‘salience’ as something that both results from (in some sense) attention conferred by attention and then serves to attract or ‘pull’ further attention, and I don’t have an especially fleshed-out idea of what that is. But on the face of it it doesn’t seem wrong to me to think of salience (or ‘prominence’, or perhaps what I think Sebastian calls ‘centrality’) as having this dual character. Certainly, if one were going with the idea of multiple stages of attention (non-active and then active, or whatever) building on one another, then it would seem natural to think this way: there’s a sort of evolving assignment of relative salience to different bits of the world, which each new act of attention both builds on and changes.

              1. Yeah, it’s interesting. A bit of empirical evidence for that broad idea: selection history is known to influences salience. So, if you have attended to something in the past, it is apt to capture attention in the future. However, a couple of caveats. Most of the evidence for that concerns fairly recent selection history (e.g., inter-trial priming). Second, some authors have suggested (I think very intriguingly) that the effects of selection history may actually be the consequence of an internal reward received upon attending to the thing. So, the effect of selection history on salience might reduce to the effect of reward history. In this sense, attention might simply behave like other habits: you repeated the same behaviour that has been associated with reward (via positive learning mechanisms).

                Nevertheless, there’s something to the idea that what captures attention is in part a function of what we have given our attention to in the past.

                1. also historical tidbit: the idea that there are passive and active “stages” to attending is in Husserl’s discussion of passive and active synthesis.Active attention, he seems to have thought, depends on prior passive attention.

    2. Hi Luke: Your second paragraph seems to me to be exactly right, but it requires some distinctions I have not clearly made in print though I think with a lot of effort, it’s there (see below somewhere). I would add that what I do want to deny, and this is in agreement with Aaron who thinks I have not gone far enough, is to deny a traditional spotlight mechanism of attention where attention is like mechanism for selection. The better idea is that attention is the result of neural mechanisms of selection, that are often constrained by the subject’s task at a time and hence, by the subject’s intention or plan.

  5. Hi Aaron, great paper and presentation! I would like to follow up on some of the things Carolyn and Luke raised above. The first problem Luke raised for Aaron above (some but not all cases of attention are actions) also occurred to me when I read Aaron’s paper. But it seems to me that the problem of active attention still stands.

    Wu can either accept or reject the following claim:

    (Universal claim) Every action has attention as a component part.

    Consider, first, what happens if Wu rejects that claim. One option, for example, would be to accept – as Carolyn suggests – that attention is not a component part of some special class of “pure actions”. I don’t think that rejecting the universal claim is very friendly to Wu’s project, though. Wu’s treatment of the many-many problem suggests that every action raises the many-many problem, because that’s part of what makes particular piece of behavior an action. But then for every successfully performed action, the many-many problem needs to be solved. So, if there are some actions that do not have attention as a component part, then we would like to hear how the many-many problem is solved in those cases, and we would like to know why attention is necessary in some cases but not all. So, this route really calls into question a core of Wu’s project, i.e. that we need attention to solve the many-many problem.

    Now consider, second, what happens if Wu accepts the universal claim. Given the fact that some cases of attention are actions, Wu – on the pain of regress – will have to claim that every active attention has non-active attention as a component part. That, as Luke rightly points out, is a logical possibility. The problem is that there doesn’t seem any evidence for the claim or indeed much plausibility. Consider actively watching/tracking a (possibly changing, e.g. moving) object. That’s a paradigmatic case of active attention. What the subject is doing is to intentionally and voluntarily visually paying attention to the object. The claim must then be that this active visual attention has as a component part a passive form of attention. The same would be true for an active shift of attention. Suppose the subject through an active decision of hers shifts her visual attention to an object on the left. If the universal claim is accepted, then that paradigm of an endogenous, voluntary attention shift, visual search or tracking has as a component a form of involuntary (exogenous?) attention. That, it seems to me, would at least need a lot of defense (which Wu hasn’t given)– as it seems to go against the standard way of talking about attention in the cognitive sciences.

    Now, Wu might – as Aaron points out toward the end of his paper – say that what has non-active attention as a component part isn’t really attention but, say, “enhancing perceptual input”. I agree with both of Aaron’s point about why that’s implausible.

    So, either way, active attention raises an important challenge to Wu’s work. Either attention is just one of many ways of solving the many-many problem, or we get some weird nesting of active/passive attention. For, what it is worth, I think the right way to go is the first: attention helps what I like to called decoupled agency, but it isn’t necessary for it. So, overall I guess, this is a friendly comment for Aaron, suggesting ways of elaborating his points so as to take into account what Carolyn and Luke said above. Anyway, cool paper!

    1. Thanks, Sebastian!

      I fully agree that Wu seems to be committed to the universal claim, and that this is why Carolyn’s Pure Action proposal is a significant departure from his account. (I say all of this in my response to Carolyn). If Pure Actions don’t require selection for action, it casts doubt on Wu’s claim that attention is necessary for action because we need attention to solve the many-many problem. It may still be true that actions in general require attention, but it’s questionable whether Wu has correctly uncovered the underlying connection between attention and action. (My own view is that Wu’s account is too wedded to a broadly libertarian/‘could have done otherwise’ conception of agency, and that we should instead think about attention’s role in action as enabling the motivational guidance of behaviour by the agent, none of which requires the presence of alternate behavioural possibilities (cf. Frankfurt)).

      As for your next point, I’m not sure I agree with this particular argument you’re exploring. I think we should distinguish the active/passive distinction (and relatedly, the distinction between endogenous and exogenous control, goal-directed and stimulus-driven, etc.) from the action/act-component distinction. As I envision the act-component proposal, the act-component isn’t meant to be a ‘passive’ or exogenous element of the action. If anything it is meant to be the opposite: it constitutes the agent’s participation in the action, such that the ensuing behaviour is not merely a reflex. To avoid a regress, though, the agent’s participation in the action (i.e., her selection) cannot itself be an action. How can this be? One option is that it is a psychological state of the agent, whose role in the guidance of behaviour is sufficient to make the episode an action. (Here’s an analogy from a more traditional causal theory of action: an agent’s intention to act is not itself an action, but a psychological state of the agent. Nonetheless, many have thought that its guidance of behaviour is sufficient to make that behaviour an action.). It seems to be a leap, though, to infer anything about whether this psychological state is determined exogenously or in a stimulus-driven manner.

      Wu can, within the parameters of the act-component proposal, distinguish active and passive (or endogenous vs. exogenous) attention by whether or not the agent’s motivational states play any role in biasing her selection. That is, Wu thinks that in attention capture, the agent selects something for action without her motivational states playing any role. I think this account of attention capture may be problematic for different reasons. (In what sense is it psychologically possible, as it must be given Wu’s conception of agency, for the agent to select otherwise than she actually does in such a case? If there’s agency at all here, it must depend on the agent’s motivational states being somehow implicated.) But the problem his view faces from attention capture is arguably orthogonal to the issue of act-components.

      I could be missing your proposal though, so please let me know if you think I’m not doing it justice here. I’m glad that you like the paper and are sympathetic!

      1. Hi Aaron: Can you elaborate on your worry here in the penultimate paragraph? I am realizing that I am using “action” perhaps a bit idiosyncratically, but that’s only because the theory dictates a specific use. Intentional action has the typical structure of intentions constraining solution to the many-many problem. As you note, remove the intention, and the same behavior (input-output) link would be a passive action. But depending on my prior states or perhaps on genuine indeterminacy in the system, I might have done something differently (in a passive way), i.e. have a different input (a subject-level psychological state) couple to a different response. This can be independent of a motivational state, and certainly independent of intention. So, to the extent that it could have gone differently, this can be determined by non-motivational factors. There were different passive actions that could have occurred. At that point, the theory is taking over, dictating how we should deal with the case of non-intentional actions, but the theory begins, as most do, with the intentional case as first in the order of explanation.

        1. Hi Wayne,

          One way to see the problem is to consider how attention capture is modelled by people like Itti & Koch, namely by appeal to a salience map. The process by which attention gets captured on such views seems to be deterministic: it’s a function of which location in a salience map is computed to be of highest salience on the basis of image-based algorithms. When attention evolves automatically, as in attention capture, it is the salience map, together inhibition of return, that causally determines what gets selected. Your view, by contrast, requires that the agent always have alternative options open to her. It must be possible for her to select otherwise than she does, where presumably the grade of possibility is possibility relative to psychological law. (I grant that it’s possible in various other senses to attend otherwise than one actually does, but we want it to be psychologically possible for the agent to do so). If the salience map is a good approximation of how automatic attention works, then there will be cases of exogenous shifts of attention that are a deterministic consequence of stimulus salience.

          Now, I personally think we should reject the notion of a salience map in favour of what I take to be its theoretical competitor, the priority map, and I think the differences between these two constructs are philosophically significant for our assessment of the relationship between attention and agency. But as far as I can tell, the differences between them don’t alter the issue I discuss above. There are probably cases of capture in which the agent could not have done otherwise than to shift attention to the attention-grabbing stimulus, which by your criteria would make it a pure reflex. What do you think of that line of thought?

          1. If it really is driven so that the agent can’t do otherwise, yes, I would take it to be a reflex, but how to understand necessity here is crucial. Let’s make it nomological necessity for present purposes. Perhaps there are what I call pure reflexes in real life. That said, I think the importance of noise both externally and internally makes a difference to how we understand how the brain works and thus, ultimately, how we as psychological subjects work. There are also interactions between top-down signals that influence what can capture attention (lots of debate about pure attentional capture), so I suspect that the cases you have in mind aren’t that common, though certainly nomologically possible.

            That said, I like the line you’re taking here and it seems like there are lots of issues in the background of your paper that are motivating it. That would be a way of developing the challenge in interesting ways that I think would push the debate forward, in the direction of linking the cognitive science with the philosophical issues about agency.

    2. Hi Sebastian!

      I agree with what Aaron says below about your objection to the universal claim, but to put it my own way, the force of your objection seems to require that the active/passive distinction map on to the action/act-component distinction which it doesn’t. So, the view isn’t committed to the claim that in every act of voluntary attending, there is an act of passive attending as part. That would not be a good consequence, on that I agree, but it isn’t a consequence. Does that shift your sympathies? 🙂

  6. Hi everyone! Thanks for engaging the work! I hope to add some thoughts regarding the contributions and comments asap. I’m just back from traveling and am teaching this morning, so more later this afternoon.

  7. Thanks everyone for engaging with the work. Sorry I couldn’t be part of this more substantively from the beginning. Not surprisingly, I think the basic ideas of the selection for action view, as I have developed it, must be right especially if we are to have a conception of attention that links philosophical and empirical concerns. Let me clarify a few things with respect to Aaron’s piece. I think the clarifications should resolve the worries about the regress. I’ll come to the commentators and comments etc. as I work through them today. Now, on Aaron’s paper:

    “Selection for action” is not my terminology but, as Carolyn points out, Allport and Neuman’s. Allport and Neuman’s (A&N’s) 1984 papers are rich and worth reading if you haven’t already. I’m just a latecomer to their view, though I’ve elaborated it on more philosophical grounds and with my own spin on it. But I don’t take myself to have departed in important ways.

    That said, “SfA” is not the best description since “action” probably means response in A&N’s usage and not “action” in our philosophical sense which is what gets Aaron’s regress going. I use the description because it marks a well-known view in psychology, but it’s a terrible description for philosophers. I took up the preexisting banner, but didn’t realize that I should have altered the colors. So, these days I regret using it since to get right, the slogan, “SfA” requires qualifications, if “action” is used in the philosophical sense of something one does. Let me know qualify.

    Worrying more about precision and not a pithy expression: attention, a component of action, guides the production of a response which is itself a component of action. In the primary sense, attention isn’t selection for action but for response (so secondarily, it is selection for action, but in a qualified sense). You can see this in my diagrams for the Many-Many Problem, for the state that gets identified as attention is always guiding the response, the output.

    Now, the question Aaron raises is: Can attention as action involve attention as a selection for response component? If you think of attention as any input state that guides response (subject to conditions I spell out in my work), then the question can be restated: Can attention as action involve an input state that guides a response? Yes, and in saying that, there is no threat of a regress. The component, not being an action, does not invite the question of what the relevant selection for action is.

    That said, I strongly suspect I have not been clear about this. In the 2014 book, I’m clearer that “attention” should mark the input when it fulfills a functional role of guiding production of the response (see diagram 3.2 in the 2014 book). But I also think that “attending” is a process and there, we are speaking of the action, the input’s coupling to guide the response (“coupling” is Allport’s term), one’s taking, by the mind, of a specific X to deal effectively with it. On my view, all action involves attention as the selective state (input deployed for guidance), but not all actions involve attending (though many do).
    In the 2014 book, I think (I think!) I’m clearer that I endorse the biased competition view of the state of attention as what emerges from the resolution of competition, where there is competition. Attention is not a further process beyond an input itself coupling to an output. So, attention is not a cause in the sense that Aaron is considering regarding input selection, a view that he attributes to me. The causal role for attention that invites this spotlight view is, for me, largely taken over by intention. There isn’t an attentional spotlight in the traditional sense. Still these are complicated matters that requires some nuance to get right, for all cases. It would not surprise me, however, that I’ve seemingly endorsed the spotlight view in earlier work.

    1. Hi Wayne! The last thing you wrote was super helpful. I am getting it now, I think: you use “attention” attention both to pick out the state where a set of inputs is marked for priority for response, and for the process where that priority marked input gets coupled to a response. So, for you there is state-attention and process-attention, where state-attention is constitutive part of process-attention. Some (but not all) instances of process-attention are actions. But no instances of state-attention are actions. And when process-attention is an action then it has state-attention as a component part (in that it’s like all actions). Does that correctly describe your view?

      I think I got confused, because I thought that attention is always a process or event, and not a state (also on your view). Specifically, I thought that you thought that it was always an event, the event of a subject’s selecting X for some action. Selecting, I thought, was always a specific temporally localized event. And then I thought that that event can either be an action or not an action, when it is an action then it’s an active form of attention/selection and when it is not an action it is a passive form of attention/selection. So, I thought that every subject involving event of the form S is phi-ing at t, is either active or passive (where every non-active phi-ing is a passive phi-ing) – guess I still think that – and then it seemed to me that active attention if it has attention as a component part must have passive attention as a component part. I now see that there is a form of attention for you (state-attention) where the active/passive distinction just doesn’t arise, and that may get you out of the problem.

      [sidenote: I find it a bit confusing to think of the priority marked input space (i.e. your state attention) as a subject’s selecting something given that the subject isn’t doing anything (either passively or actively)]

      Anyway, curious to hear Aaron’s thoughts on your response as well!

      1. Good to hear from you Sebastian. It’s been a while since we’ve crossed paths (Sicily, I think, two years ago). Yes, your summary is right.

        The sidenote is worth expanding on a bit. Absolutely, I’ve used “subject’s selecting X” to talk about attention and the gerund there certainly makes one think process and not state. It will be important to distinguish between attention as an activity and attention as a state. This is one way to understand Aaron’s critique, how does one fit different activity/action conceptions of attention into the view. Your sidenote echoes that, as did your original reply. Whether the theory will perfectly match our pretheoretical ways of talking about attention is not fully clear. But here are some distinctions

        Your attending as an action for its own sake (Aaron and I disagree on how to understand that, and you will have your own views).
        Your attending as part of doing something else: this need not be an action but perhaps merely an activity, something that is part of your doing something.
        States of attention: one’s selective orientation towards the world.

        Take A–>B, where A is the input perceptual state and B is the output response. This link (coupling) is the solution to a Many-Many Problem.

        Actions are A–>B, and where intentional, the intention structures that solution. So, attending as an action has the full structure of coupling input to output (Aaron and I disagree here on how to understand this)
        You can speak, if you like, about the activity of attending where this is A–> (if you like) the process of guiding. It’s not a full blown action, but there’s a sense to the agent’s selecting for
        You can speak, as I like, of states of attention which is A in the context of A–>, the agent’s selective orientation towards what A represents or is of.

        So, action comes first in the order of explanation, but from that, different ways of thinking about attention emerge. This is rough and ready, but meant to map our attention talk to the structure of action. It becomes complicated, but important, to introduce and automaticity/control distinction too (this is the passive/active distinction, perhaps). But I’ll leave that aside.

        Does the structure above perfectly map onto pretheoretical talk? Probably not, but at some point, the theory has to take over. As I suggested in comments to Aaron on biased competition, the structure above provides what Marr might have called the computational theory of attention, but embedded in a theory of action. Biased computation provides the algorithmic level and actual neural implementations, the hardware.

        1. This is great, Wayne! On the sidenote: I am realizing that the way I’ve been thinking about attention and the way you’ve been thinking are actually even a bit more similar than I thought. What you call a selective orientation towards the world (the state) is pretty much what I call a priority structure. Attention, for me, is a subject’s regulating (changing or maintaining) those structures (either for its own sake or in the course of another activity). That regulating is what can be automatic or controlled etc. I pretty much accept what you call the computational theory of attention (what I’d say is that something like this is the function of attention). Guess, I don’t think that this role is essential or necessary (attention isn’t functionally defined). It’s metaphysically possible to have pointless selective orientation, though in us it does play that role (I think attention is a bit like knowledge here: knowledge is super important for action, and serves various roles in action, but you can’t define knowledge via it’s role in action).

          Anyway: thanks for all the clarification. I have a much better grip on your view now!

          1. Glad to hear that there’s some convergence (or has been). On the necessity of the computational story, how to proceed? I suspect that much of what we grasp pretheoretically regarding attention is tied to things that we do (James’ definition). So at some point, it’s stipulative that attention is X. My own interest is to find a notion that is theoretically fruitful in linking the cognitive science to genuine philosophical concerns in action theory, epistemology, and the theory of perception/consciousness. I think the selection for action (response) account does just that. So sure, we can disagree on whether it is necessary or not. I tend to take that line and don’t see any downsides to it but many theoretical upsides.

            Attention and consciousness is much more fraught, and again, this is partly what Aaron and I are disagreeing about in talk of sharpened experiences. But you have a book on that general topic (attention/consciousness) and I’ve made some remarks in various places. We should save that conversation for when your book comes out!

          2. Sebastian, as I see it, there are important similarities as well as important differences between your views. If you get a chance to read the conclusion of my paper (and the last few paragraphs of my response to Carolyn and Robyn, where I discuss the relationship between states of attention and processes of attending), I think you’ll see that I’m trying to bring Wayne’s view more into line with yours. As I see it, you two officially endorse diametrically opposed views about the state/process relationship, and you two also hold very different views about the relationship between attention and perceptual experience. While I’m quite sympathetic to the overarching claim that attention is selection for action—I think Wayne is right about attention’s explanatory role in mental life—I’m also quite sympathetic to the idea that attention is structuring or prioritizing, and that this has a phenomenal character. So, my closing revisions to Wayne’s SfA account are meant to sketch how those different ideas can be reconciled, while respecting BCT. But changes are required on both sides to make that work. I think I abandon Wayne’s purely functional conception of attention in favour of a partly phenomenal conception of attention, but I also don’t endorse your specific characterization of structuring/prioritizing as involving a primitive/sui generis phenomenal character. I think the duplication arguments you give fail to establish that conclusion, though they do succeed at showing something! Obviously I don’t go into those arguments here, though.

            I also echo Wayne’s point that part of what we’re debating here is the relationship between attention and perceptual experience/consciousness. Talk of “sharpening” “highlighting” and “enhancing” is serving as a bit of a placeholder for the phenomenal contribution attention makes to perception.

Comments are closed.