Denis Buehler (Universidad Nacional Autónoma de México)[Jump to comments]
Abstract: I offer a novel way of addressing Harry Frankfurt’s ‘problem of action’: From reflection on empirical psychological research, I develop an empirical explication of the notion guidance by the individual. I speculate on how this explication may contribute to understanding the nature of agency. 
1 Understanding Agency
The deepest challenge in action theory, says Harry Frankfurt, is to explain the distinction between processes that an individual guides and processes that occur without an individual’s guidance. We have to understand the nature of guidance in order to understand the nature of agency.
A macaque searches some bushes for raspberries. He actively guides his visual attention from one location to the next until he finds the berries. When a bright object abruptly appears in the periphery of the macaque’s field of view, the object captures the macaque’s attention. His attention then shifts passively. A jumping spider walks across the jungle floor toward her prey. The spider actively guides the movements of her legs in so walking. When the spider’s legs contract in a shreck-reaction upon getting in contact with a toxic substance, the spider does not guide her legs’ movements. Often when you solve a chess problem you systematically consider different possible solutions. You are then actively guiding your reasoning toward a solution of the problem. You do not guide your reasoning when the solution ‘pops’ into your mind.
How can we understand this distinction? As I will explain later, all extant accounts of agency fail to answer this question in a satisfying way. This fact motivates the search for a novel account. It also motivates the application of a new methodology in action theory. In this paper I attempt both.
Much of this paper is devoted to developing an empirical explication of individuals’ guidance by investigating actual primates’ guidance of visual attention. I will argue for this explication from a case study of such attention shifts’ psychology. The explication’s basic idea is that some specific type of psychological process realizes these individuals’ guidance. I show that
Guidance of Visual Attention
Central executive control realizes primates’ guidance of their visual attention shifts.
Empirical psychology provides extensive evidence that actual primates have a central executive system that coordinates and integrates psychological states and processes from different modalities and other psychological subsystems. Central executive control is the control that the central executive system exercises over a process. My appeal to central executive system meets two requirements on an explication of individuals’ guidance. First, central executive control correlates with those individuals’ guidance. But second, appeals to central executive control help deepen our understanding of individuals’ guidance.
An explication of a concept must not be conflated with a reductive account, a definition, or the giving of necessary and sufficient conditions on a nature. An explication of a concept need not even apply to all instances of that concept in the actual world. Many explications identify uses of a concept that centrally guide its applications. Many explications identify salient marks of such uses of a concept. In either way explications can deepen understanding of the concept. Explications often deepen our understanding of a notion by deepening our understanding of the kind that the notion denotes. Such explications often depend on empirical information. I do not claim that my initial explication offers a necessary condition on guidance. I do not claim that it offers a condition on all animals, even those in the actual world. I do not even claim to have shown that all exercises of agency in actual primates are constituted by central executive control. My more modest, initial aim is to offer an empirical explication of individuals’ guidance.
Nevertheless I will suggest that my empirical explication may bear on the constitution of guidance, and hence of agency. A constitutive explication of a concept provides a necessary condition on the kind that the concept denotes, appeals to which condition deepen our understanding of the concept. A constitutive explication contributes to an account of the kind. A cursory overview of guided processes in actual primates makes it plausible that all these processes are subject to central executive control. I suggest that therefore
Guidance in Primates
Central executive control (partly) constitutes primates’ guidance.
Next, I conjecture that this result may point us in the direction of an account of guidance for creatures beyond primates.
Constitution of Guidance
The control of a process by an individual’s central structures (partly) constitutes that individual’s guidance.
Central structures function to integrate and coordinate activity in an individual’s sub-systems. These structures need not involve central executive control. They need not be psychological structures. I conjecture that for all agents, when they guide their act, some central structure unifies their different capacities so as to constitute the individual’s guidance of that act. Showing that either proposal specifies a constitutive condition on agency, even just in primates, would not only require investigating other kinds of agency in primates and other animals. It would also require arguing that my explication provides a necessary condition on agency’s nature. Both tasks are beyond the scope of the present paper. Nevertheless I conclude by explaining why I think that the two constitutive explications hold some promise of being true.
My first aim in this paper is to contribute to our understanding of individuals’ guidance. My second aim is to advertise a new methodology for action theory. I develop my explications of guidance from reflection on empirical psychological research. Frankfurt, like many other action theorists, sought a constitutive explication of guidance exclusively by way of armchair reflection. Here I depart from Frankfurt. I doubt that armchair reflection alone will yield a sufficiently full explication of guidance. But, to be sure, I also do not want to suggest that we should merely discard our concepts in favor of scientific notions. I rather hope to integrate philosophical armchair reflection methods with reflection on scientific knowledge. My method will rely on intuitively clear cases of individuals’ guidance, psychology’s description of these processes, and armchair reflection on marks of our concept guidance. I argue that reflection on all three items supports my explication Guidance of Visual Attention. I will have achieved my methodological aim if my reader acknowledges that a serious engagement with empirical research can yield interesting, new philosophical insights.
I begin, in section 2, by briefly explaining how recent accounts of the nature of agency fail to meet plausible constraints on any such account. The paper’s main argument will be presented in section 3. It establishes that central executive control realizes primates’ guidance of visual attention. In section 4 I conclude with reflections on the constitution of agency.
2 Recent Action Theory
When Frankfurt introduces the challenge of explaining guidance, he in effect formulates what I will call the comprehensiveness-constraint and the attributability-constraint on an account of agency. The six most influential accounts of agency fail to meet these constraints.
The contrast between guided and non-guided episodes can be found across species. This observation motivates the
An account of agency should explain all instances of agency (across kinds of agents and species).
An account of agency should not narrowly focus on sophisticated human forms of acting. Much recent action theory fails to meet this constraint. Reflection-based accounts of agency take reflective thought to be constitutive of agency. For an event to be an action, the event must be caused by an intention (or a desire, or a reason) and a higher-order, reflective state that endorses acting from that intention. Thus says Velleman: “action is activity regulated by … reflective understanding.” Responsibility-based accounts contend that an event is an action just in case the individual is answerable for the event (in some specific way). Anscombe, for example, suggests that intentional actions are “actions to which a certain sense of the question ‘Why?’ is given application, [namely the sense] in which the answer, if positive, gives a reason for acting.” Intention-based accounts have it that an event is an action just in case it is caused by individuals’ intention to act. Intentions are propositional states (or thoughts) – possible conclusions of practical reasoning. This approach stems from Donald Davidson, who claims that someone “is the agent of an act if what he does can be described under an aspect that makes it intentional.” For Davidson, an event is intentional just in case a belief and a pro-attitude of the agent caused the event.
These proposals hold no promise to account for agency across species. A lioness acts when she stalks her prey. Birds act when they build their nest. A shark acts when mating. An ant acts when it walks back to its hive. These animals surely do not engage in reflection on their own mental lives. These animals also do not ask why-questions in the way humans do. These animals do not engage in any form of reason-giving. And while some mammals may be capable of forms of practical reasoning and hence may have propositional intentions, it is an empirical question whether birds and fishes have them, and highly unlikely that arthropods do. Even for human agency, the accounts seem to over-intellectualize. It seems implausible that humans guide all their acts by reflecting upon them. Some human agency may not even be guided by propositional intentions. Many human acts are highly over-learned or reflexive. Think of tying your shoelaces, weaving a jab, or walking down the street. Usually individuals do not reflect on their intention to tie their laces. It is unclear whether individuals must propositionally intend to weave the jab, or to walk there in order to perform these actions. Again, these issues are open to empirical inquiry.
These first three accounts of agency promise at best to capture forms of relatively sophisticated human agency. Goal-based accounts hold more promise for meeting the comprehensiveness-constraint. On these accounts, for an event to be an action, it must be caused by a state that represents a goal. This state need not involve reflection. It need not be a propositional intention. Instead, the goal-representing state might be a more primitive, image-like state. Thus Alfred Mele acknowledges that actions “may be partially accounted for by low-level representations of … various trajectories of ‘to-be-performed’ movements.” Mele allows that such representations may be (elements of) action-guiding states. Many animals have perceptual capacities. They should hence be able to guide their acts through an image-like representation of the act’s goal. The same kind of image-like goal-representation may guide non-sophisticated human acts.
All extant such accounts, however, fail to meet the
An account of agency should explain under which conditions the guidance of a process is attributable to the whole individual.
The attempt to find “a place for agents in the explanatory order of the world” motivates this constraint. The constraint has two components. First, accounts of guidance must distinguish, in a principled way, between states and events that occur at the level of the whole individual, and those that occur at the level of an individual’s sub-systems alone. Actions and perceptions are individual-level events and states. Many processes in the circulatory or visual systems, for example, occur at the level of those sub-systems alone. The circulatory system, not the individual, transports nutrients. The visual system, not the individual, computes edge-representations. The same distinction can be made for the case of action. Performing any motor action involves computations over a vast number of perceptual and motor states. Suppose that an individual guides her hand to pick up an apple. Visual information about the apple integrates with proprioceptive information about the position of different parts of arm and hand. The motor system computes a representation of arm and hand’s goal-states. If the individual’s finger was straight, a representation may specify its goal-state as being bent at some angle. If the wrist was relaxed, a representation may specify the degree to which it must be turned. Many of these image-like, goal-representing states occur only at the level of the individual’s motor system. They cannot be attributed to the whole individual. Since they are not the individual’s events or states, they cannot plausibly constitute her guidance of the movement. Goal-based accounts should distinguish sub-individual goal-representing states from individual-level constitutors of guidance.
Second, accounts of guidance must distinguish, in a principled way, between individual-level states and events that merely causally influence a process, and those that constitute the individual’s guidance of an act. An image-like, goal-representing state, a propositional intention, or a reflective, higher-order commitment may each dispose an individual to pick up an apple. An image-like representation of an apple, an intention to eat apples, or a higher-order commitment to eat fruit may each increase the likelihood that the individual should choose an apple as her food. But suppose that causation by her individual-level image-like, goal-representation as of this apple constitutes the individual’s guidance of the action. This latter individual-level state’s causal influence on the action must importantly differ from that of the earlier individual-level states, which provide merely a background condition for the individual’s guidance. In virtue of what does the image-like, goal-representing state constitute the individual’s guidance? Goal-based accounts should distinguish between individual-level states that constitute an individual’s guidance, and those that provide a causal background condition for guidance. An account of agency will be incomplete if it does not meet the attributability-constraint.
Agent-causation accounts postulate that an event is an action just in case the individual herself causes the event to occur. Roderick Chisholm writes: “when a person acts … he causes his own undertaking and … there is no event or set of events constituting a sufficient causal condition for that undertaking.” This kind of account meets the attributability-constraint. The account could plausibly be broadened so as to encompass other animals’ acts, and hence meet the comprehensiveness-constraint. However, agent-causation accounts postulate a primitive form of causation – causation by an agent. Agents are substances. Substances are not events. Natural science recognizes different types of event-causation. Natural science does not acknowledge causation by substances. Agent-causation explications have been widely dismissed for conflicting with a scientific world-view. Moreover, agent-causation accounts do not contribute much to explaining under which conditions an individual guides a process. These accounts rather stipulate that whenever an individual guides an act, that individual primitively causes the action. So the accounts do not address the attributability-constraint in a satisfying way.
Trying-based accounts do not appeal to a primitive form of causation. Tryings are agential events that cause, and are parts of, an action. Trying-based accounts contend that an event is an action just in case the individual’s trying caused the event in the right way. Thus writes Brian O’Shaughnessy: “All physical action involves a willing or bringing about of act-neutral bodily events.” Tryings are plausibly individual-level events. It seems possible to specify tryings in such a way that both humans and other animals undergo such events. So trying-based accounts, too, hold some promise for meeting both the comprehensiveness- and the attributability-constraint. But no trying-theorist has ever specified the nature of a trying. Intuitively, one may think of tryings as conscious efforts. Proponents of trying-based accounts tend to reject this conception of a trying. Indeed, if this were the correct account of a trying, then it would be at best unclear whether all animal agents, such as spiders and fishes, do try to act, whenever they act. And it seems questionable whether all human action, such as effortless active inference, involves a trying. But pending some independent specification, tryings are identified merely as agential events that constitute acts. In this form, trying-based accounts do not satisfyingly explain under which circumstances the guidance of a process can be attributed to an individual.
None of the six accounts of agency meet both the comprehensiveness-constraint and the attributability-constraint in a satisfying way. This fact motivates searching for a new account.
3 Guidance of Visual Attention
In this section, I argue for an empirical explication of individuals’ guidance. Later, I will rely on this explication to motivate a new account of agency. I will then also explain how we can hope to address the comprehensiveness- and the attributability-constraint in a more satisfying manner. But for now, my focus is on establishing that:
Guidance of Visual Attention
Central executive control realizes primates’ guidance of their visual attention shifts.
The central executive system coordinates and integrates psychological states and processes from different modalities and different psychological subsystems. The system’s coordination and integration enables many individual-level processes, such as problem solving, goal-represented bodily action, and attention shifts. The system achieves such coordination and integration by deploying its own central processing and storage resources. Central processes involve the transformation of non-modular and intermodal states. This processing is intermodal insofar as the central executive system serves as an interface for integrating information from different modalities and psychological capacities. The processing is non-modular insofar as it accepts a wide range of inputs.The central executive system controls individual-level processes, most fundamentally, by flexibly allocating processing and storage resources to the respective processes. Three executive functions that characterize the system’s functioning are the competencies to switch mental set, maintain relevant memories, and inhibit prepotent responses and distractor stimuli. Switching consists in the process of abandoning one process or task and initiating another. Switching is an exercise of a competence to organize different, competing activities at a time, and over time. Maintenance is the activation and holding active of relevant memories, and of encoding incoming task-relevant information into working memory. Inhibition of prepotent responses and distractors is the process of suppressing irrelevant processes and stimuli. Inhibition consists not merely in withdrawing central resources from processes and stimuli, but in suppressing their influence on the execution of ongoing tasks. By allocating central processing and storage resources, the central executive system determines which information and psychological processes from different psychological sub-systems contribute to the course and outcome of individual-level psychological processes. The central executive system controls individuals’ processes in this way.
I abstract from specific models of the central executive system. I do not think of it as primarily a central processing unit, some specific computational or neural mechanism, or some specific part of the brain. Psychologists identify central executive functions, as I think of them, on the basis of behavioral experiments involving a variety of tasks. There are many promising attempts to integrate this behavioral research with computational modeling and neuroscience. But it seems too early to me to strongly commit to the success of any of the specific attempts.
As I explained earlier, for central executive control to help explicate individuals’ guidance, it must, first, appropriately correlate with individuals’ guidance. For central executive control to appropriately correlate with individuals’ guidance, such control must be absent from processes that individuals do not guide. And it must correlate with individuals’ exercises of guidance, when a process is guided. But second, appeals to this central executive control must deepen our understanding of individuals’ guidance. The argument in this section has, accordingly two large parts. They correspond to two steps in providing an empirical explication. First, in section 3.1, I survey psychological factors that help shift attention and I suggest that only central executive control appropriately correlates with guided processes. Second, in section 3.2, I reflect on the notion individuals’ guidance, and on how appeals to central executive control promote our understanding of that notion.
3.1 Central Executive Control as Correlating with Individuals’ Guidance of Visual Attention
Remember the macaque. The monkey guides his shifts of attention when he is searching for berries. His attention shifts passively when a salient stimulus like a spider captures his attention. What psychological processes underlie these different ways in which individuals shift visual attention?
3.1.1 Captured Attention
What happens when attention is captured? Psychologists distinguish two psychological systems for orienting attention: the exogenous and the endogenous system. The exogenous system exclusively responds to stimuli that are physically salient. The system orients attention reflexively, rapidly. It is very difficult to voluntarily suppress this system’s activity. Intentions, beliefs, expectations often do not influence its activity. The endogenous system’s activity, on the other hand, is not restricted to any specific type of stimulus. Intentions, beliefs, expectations, and goals typically shape its activity. The exogenous and endogenous systems both generate assignments of priority on a priority map. The priority map is a topographical representation of the visual scene that assigns priority values to objects and locations in the scene. Attention shifts to the location with highest priority.
When attention is captured, it shifts passively. Individuals do not guide such shifts. We can distinguish two ways in which visual attention can be captured. First, visual attention can be captured in absence of individuals’ guidance. Suppose that the macaque passively stares at his surroundings. A suddenly appearing spider will capture his visual attention. But, second, the appearance of the spider may also override individuals’ guidance. The macaque, while guiding his visual attention shifts in his search for berries, may have his attention captured by the spider, in spite of his guidance. The capturing stimulus overrides the monkey’s guidance. Attention shifts during instances of attention capture can be fully explained by appeal to the exogenous system’s activity alone. Endogenous factors do not contribute to these shifts. The exogenous system’s activity can hence not be the primary realizer of individuals’ guidance. Such activity does not appropriately correlate with exercises of individuals’ guidance.
3.1.2 Drawn Attention
Many shifts of attention result from the interaction of the endogenous and exogenous system. Factors that do and factors that do not correlate with individuals’ guidance co-occur in such cases. How can we distinguish these different factors? Shifts during visual search typically involve the interaction of the two systems. In visual search, individuals guide their attention shifts toward the goal of their search. When a stimulus captures attention during visual search, the exogenous system overrides individuals’ guidance. When the exogenous system influences, but does not fully override or interrupt endogenously driven shifts, I will say that attention is drawn by a stimulus.
Suppose that the subject in an experiment searches for a green diamond. It so happens that the green diamond is also the only green item in the display. So the green diamond is the physically salient stimulus in the display. Folk et al. suggest that the individual’s goal to find the green diamond may tune the exogenous system to be more sensitive, and react more rapidly to, green items. The increased sensitivity to the salient item – which is also the target of the search – enhances the individual’s search. Individuals become faster and more accurate at shifting attention to the diamond. This increased sensitivity is due to the individual’s goal to find the diamond. Despite the contribution to these shifts by the exogenous system, we recognize the individual as guiding these shifts. But also, clearly, the contribution of the exogenous system to these shifts is not the factor that appropriately correlates with the individuals’ guidance. The salient stimulus may equally draw attention away from the target’s location. The stimulus may thus interfere with the individual’s search. Suppose that a physically salient green diamond attracts attention away from the search target, a red circle, say. But suppose also that the distractor diamond does not capture attention. The distractor merely deflects the attention shift, bends the trajectory of the shift, without ever fully disrupting the shift. Despite the exogenous system’s interference, the individual still guides her shift toward the search target. Godijin and Theeuwes identify different types of ‘bent’ trajectories for attention shifts that result from the exogenous system’s interference with attention shifts during visual search. Again, the exogenous system’s influence clearly does not appropriately correlate with the factor that realizes guidance.
For both types of drawn attention, we recognize the exogenous system’s influence as not correlating with individuals’ guidance. Cases of interference with individuals’ search are especially clear in this respect. Whatever factor realizes individuals’ guidance does not normally also interfere with it. When we intuitively recognize some psychological factor as interfering with an individual’s guidance of a process, we have reason to think that that factor cannot be the correlate of individuals’ guidance. In the next section I will appeal to drawn attention in order to rule out a range of endogenous factors as realizers of guidance.
3.1.3 Shifts Drawn by Non-Guiding Endogenous Factors
Psychologists consider an extremely wide range of factors to be ‘endogenous.’ Not all these factors plausibly realize the individual’s guidance. Some of these factors, too, draw attention. For example, individuals can be primed to be more sensitive and react more readily to certain stimuli. If an individual repeatedly searches for a green diamond, she will get faster at finding green diamonds. Once primed for green diamonds, the individual will find it more difficult to ignore green diamonds during subsequent searches – even if now she is looking for a red circle. Priming can draw attention to some type of stimulus. Also, individuals store statistical information about configurations in visual scenes. When a search target consistently appears at a certain distance and angle from, say, a heptagon of vertical bars, individuals more readily shift attention to a location at the same distance and angle from such a configuration. The location attracts the individual’s attention even if the configuration is no longer indicative of the search target. The memory for the configuration draws attention to that location. We obtain similar effects due to associations of past reward with types of stimuli, and due to large amounts of information about objects and scenes stored in long-term memory. Even working memory can draw attention. Suppose that an individual maintains a red circle in visual working memory. This memory may then influence subsequent search for a green diamond. Suppose that an individual is told to search a display that contains the target, a green diamond, but also a range of distractors. Suppose that one of the distractors is a red circle, matching the representation maintained in visual working memory. Soto et al. provided experimental evidence that the distractor will draw the individual’s guided shifts during her search. Search times increase and the shifts decrease in accuracy. The influence of the red circle stored in working memory interferes with the individual’s search. Clearly, the memory of the red circle does not realize the individual’s guidance.
All these endogenous factors draw attention. We intuitively recognize them as not realizing the individual’s guidance. We furthermore recognize them as interfering with individuals’ guidance. These factors hence do not plausibly correlate with individuals’ guidance. Which endogenous factor plays this role?
3.1.4 Central Executive Control as the Actual Correlate of Guidance
Psychological research suggests that central executive control appropriately correlates with exercises of individuals’ guidance. A study by Olivers and Eimer illustrates this research. They performed experiments similar to those sketched in the last sub-section. In their studies, too, individuals had to memorize a color, perform a visual search, and a memory test. The memorized color influenced attention shifts during a subsequent search. When their subjects could not predict whether they would first complete the search or the memory test, individuals pursued both goals simultaneously. Olivers and Eimer found that the effect of the memorized color on subsequent search doubled, relative to the condition in which individuals could predict which task they would have to complete next. Olivers and Eimer explained this result as due to the allocation of central resources to both representations in working memory. When individuals guide their search on the basis of a representation of the search-goal in working memory, then their central executive system allocates central resources to that representation. This allocation increases the representation’s influence on the priority map and hence subsequent search. When working memory interferes with search, no central resources activate the memory as a search goal. Experiments of this type support the idea that the central executive system’s allocation of resources to some specific process correlates with individual’s guidance of attention shifts to locations.
There is converging evidence for all three signature executive functions in visual search. Switching of mental set is required to abandon a task and initiate a visual search. Walther and Fei-Fei showed that visual search exhibits typical effects of switching. They asked subjects to switch back and forth between the task of searching for a target in a display, and that of reporting the color of the display’s frame. The central executive system takes between 200 ms and 800 ms to switch tasks. Individuals’ performance on the second task was only impaired when they had less than 200 ms to switch sets. These experiments support the claim that initiating a visual search requires switching. Maintenance of relevant representations in memory is required for the execution of visual search. For example, Oh and Kim showed that when individuals had to memorize locations of four squares on a screen, a subsequent visual search slowed down. The individuals were less effective at finding the search target. Oh and Kim explained their results by pointing out that these individuals’ visuo-spatial working memory was filled to capacity. If the memory storage for locations is filled, the central executive cannot effectively control assignments of priority on the priority map. Inhibition of irrelevant distractors is required to search a cluttered display. Lavie et al. found that individuals shift attention to distractors more often when the individuals have to concurrently generate random numbers or perform calculations on numbers. The amount of time that they needed to find the target increased in proportion with the amount of unrelated central processing individuals had to carry out. Lavie et al. thus showed that the central executive system is needed, in visual search, to inhibit the effects of distractors.
This empirical research supports the claim that whenever primates guide their attention shifts, then these individuals’ central executive systems control the shifts of visual attention. Earlier I argued that the central executive system does not control shifts during attentional capture. Those shifts are entirely controlled by the exogenous system. I have also argued that the endogenous factors that draw attention are not exercises of central executive control. So exercises of central executive control plausibly correlate with primates’ guidance of visual attention.
How does the central executive system realize individuals’ guidance of attention shifts during visual search? Remember, again, the macaque’s visually searching for a berry. The monkey guides his attention during his search. When the monkey sets out to find a berry, the central executive system activates an iconic, picture-like memory of the berry. The central executive system implements the monkey’s goal by prioritizing the memory of the berry as the template for computations of priority on the priority map. This input governs the computations of priority for shifting attention to specific locations. When the monkey begins his search, the central executive system initiates computations that yield a priority assignment for possible target locations on the priority map. The computational mechanisms underlying the priority map prioritize locations on the basis of how visually similar the target, stored in memory, is to items at locations in the visual scene. Suppose that there was no berry where the monkey shifted attention first. When the monkey shifts attention to the next location the priority map represents this new location as the next-most likely location for the target. The process repeats until the search terminates. When the monkey finds the berry, the visual attentional system computes a sufficient match between the item at a searched location and the iconic memory of the berry activated by the central executive system. The central executive system may now switch mental set from the goal of finding the berry to that of reaching for it.
I conclude that there is strong empirical support for the claim that central executive control correlates with primates’ guidance of attention shifts.
3.2 Central Executive Control as Explanatory of Primates’ Guidance of Visual Attention
An explication must help ground explanations of what it explicates. I now want to reflect on a way in which my explication furthers our understanding of the notion of guidance. Appeal to central executive control helps deepen our understanding of individuals’ guidance by making the concept shaper, more precise. Reflection on the notion individuals’ guidance suggests several marks that we associate with this notion. Such marks do not constitute necessary or sufficient conditions on guidance. Rather, these marks represent salient aspects of central instances of individuals’ guidance. We can see how appeal to central executive control helps explain individuals’ guidance, if we can see that central executive control can realize salient aspects of individual’s guidance. Appeal to the scientific explanation of these aspects’ realization sharpens the notion’s pre-theoretical marks.
What is an everyday example of a good guide? Consider the example of a mountain guide. First, the good guide sets the goal for the expedition’s ascent. She picks a peak as her expedition’s destination and orients her expedition toward the peak. Second, the good mountain guide has relevant organizational competencies. She brings the map, rope, pickax, and crampons. She knows how to use them. The good mountain guide is able to negotiate the terrain that the expedition crosses. She knows the route to that peak. Third, a good mountain guide uses these competencies, as needed, to stably steer toward the goal of the expedition. She gets the crampons out when the expedition crosses ice. If necessary, she quells resistance on a stretch that strikes her companions as unduly strenuous. Appeals to central executive control help us understand individuals’ guidance by connecting it with a kind that appears in psychological explanations. Reflection on this psychological kind reveals how central executive control realizes the pre-theoretic marks of the notion guidance laid out in the foregoing paragraph. Such reflection yields empirical conditions that realize aspects of guidance, and hence plausibly guidance itself, in actual psychologies. Central executive control realizes the individuals’ pursuit of a goal. The system organizes competencies and resources for attaining the goal. And it exercises these competencies, as needed, to stably steer toward that goal. Remember the case of visual search. First, the central executive system implements the individual’s goal by activating a representation in working memory. Second, the central executive system realizes organizational competencies for attaining the individual’s goal insofar as it activates or holds available storage and processing resources. Third, the central executive system, when functioning well, exercises its executive functions – that is, allocates resources – in ways that further the individual’s goal-directed activity.
Psychology can help explain guidance’s marks by appeal to central executive functions – switching, maintenance, and inhibition – that characterize the central executive system. Psychology offers fairly sharp criteria that characterize each executive function and confirm its exercise. The scientific notions switching, maintenance, and inhibition are certainly more sharply characterized than our pre-theoretic marks of individuals’ guidance. Appeal to these scientific notions in explanations of the marks of guidance hence helps sharpens the latter notion.
Central executive control furthers our understanding of individuals’ guidance in similar fashion. Appeal to central executive control helps illuminate conditions for attributing guidance to individuals, rather than their subsystems alone. We should first note that the central executive system is a prima facie plausible candidate for realizing guidance by the individual, because the central executive system helps realize instances of processes that are attributable to the whole individual, such as instances of problem solving, inference, and attention shifts. Moreover, the central executive system functions to integrate and coordinate individuals’ states and competencies for the successful execution of individual-level processes. The system organizes and activates resources for a process. The system coordinates individual-level processes both simultaneously and through time. By integrating capacities for an individual-level process, the central executive system unifies the activities and products of the individual’s different psychological sub-systems and enables a process at the level of the whole individual. In particular, the central executive system can perform these functions for attaining individuals’ goals. The central executive system has a function to implement individuals’ goal-represented processes. The integration and coordination achieved by the central executive system serves, in central cases, the attainment of individuals’ goals. In explanations of many other processes, psychology does not appeal to goals of the individual. Computations of edge-representations from representations of illumination gradients, for example, are normally explained in terms of transformations by the visual system, without influence from individuals’ goals. Central executive control thus marks a level of integration and coordination that helps unify psychological processes in the individual’s sub-systems so as to generate the individual’s activities, directed at the individual’s goals. This system contributes to unifying psychological processing in different psychological sub-systems so as to make them an individual’s goal-represented activities. Because central executive control marks such a level of integration and coordination, the central executive system is a plausible realizer of individuals’ guidance.
Psychology’s explanations of processes in terms of central executive functions are more rigorous and precise than our pre-theoretic notion of individuals’ guidance. As before, appeal to these scientific notions in explaining marks of individuals’ guidance hence yields a sharpening of the latter notion. So we can see how appeal to an empirical condition helps deepen our understanding of guidance by the individual by making the concept shaper, more precise.
3.3. Executive Control as Explicating Primates’ Guidance of Visual Attention
I have argued that central executive control correlates with primates’ guidance. I have also argued that appeal to central executive control helps us better understand guidance. I conclude that
Guidance of Visual Attention
Central executive control realizes primates’ guidance of their visual attention shifts.
My argument appealed to three factors. It was driven by our intuitions about cases. I also appealed to pre-theoretic marks of our concept guidance. Finally, I appealed to our scientific knowledge of the processes that occur when individuals guide their acts. I did not merely suggest replacing our pre-theoretic notion with a scientific notion. Nor did lay down pre-theoretic marks of guidance as armchair criteria for any mechanism that is to realize individuals’ guidance. My claim is rather that, taking into account all three factors, our best explanation of primates’ guidance of visual attention appeals to central executive control. We can explicate individuals’ guidance by appeal to this explanation.
4 The Constitution of Agency
Earlier I promised that I would use my empirical explication of guidance to motivate a novel account of agency. I promised that I would sketch how such an account might help us address the comprehensiveness- and the attributability-constraint. I want to return to these promises now. An account of agency provides constitutive conditions on the kind agency. Constitutive conditions on a kind are “conditions that are necessary, sufficient, or necessary and sufficient to be something of [a] kind or with [a] nature, and [they] are in principle potentially relevant to explaining, understanding, illuminating the kind or nature.” In what way does Guidance of Visual Attention help uncover constitutive conditions on agency?
4.1 Guidance in Primates
I begin by motivating an account of guidance in primates. I have argued that appeals to central executive control help us explicate individuals’ guidance. There is ample evidence that the central executive system controls primates’ acts in many other cases, including central instances of mental and bodily action. The evidence makes it plausible that central executive control correlates with all exercises of primates’ guidance in the actual world. The latter observation, in turn, lends plausibility to the claim that
Guidance in Primates
Central executive control (partly) constitutes primates’ guidance.
Suppose that the meaning of concepts such as agency and individuals’ guidance is partly determined by relations to their denotation, just as in the case of water and life. Just as in these latter two cases, providing a constitutive explication for agency and individuals’ guidance would accordingly require investigating the respective kinds in the actual world. Further suppose that central executive control actually correlates with primates’ guidance. Pending considerations to the contrary there would then be some reason to think that central executive control constitutes a necessary condition on primates’ guidance. Each of these claims requires extensive support than I cannot provide here. But they do have some antecedent plausibility. Earlier I explained how appeal to central executive control deepens our understanding of individuals’ guidance. The same considerations apply to Guidance in Primates. If central executive control is a necessary condition on primates’ guidance, and if appeal to central executive control indeed deepens our understanding of primates’ guidance, then central executive control is a constitutive condition on such individuals’ guidance. So there is some reason to think that central executive control partly constitutes primates’ guidance.
4.2 Addressing the Attributability-Constraint
This constitutive explication of primates’ guidance would constitute progress over recent action theories as regards the attributability-constraint. This constraint requires explaining under which conditions the guidance of a process is attributable to the whole individual, as opposed to merely one of her sub-systems. The central executive system functions to integrate and coordinate activity in different psychological sub-systems so as to yield individual-level processes. Appeal to this system plausibly helps us understand why the guidance of a process is attributable to the whole individual, not merely one of her sub-systems.
Guidance in Primates hence does justice to the basic insight of agent-causation accounts of agency: that in order to explain agency we must acknowledge the agent’s role in acting. The explication does so without being committed to an unscientific form of causation. A mark of the agent’s involvement in her act would be the exercise of central executive control. Guidance in Primates might be seen as a way of specifying the nature of tryings. We might consider a trying to be the allocation of central executive resources to some process. If the main problem with trying-based accounts was that they do not specify the nature of tryings, my proposal constitutes progress. The appeal to empirical science constitutes genuine explanatory progress in both cases.
Guidance in Primates further promises a solution for the difficulties facing the reflection-based, intention-based, and goal-based accounts of agency. It seems likely that sometimes, we guide our acts through reflection. At other times, plausibly, a thought or propositional intention guides our doings. And equally, it seems plausible that under certain circumstances an image-like representation alone can implement individuals’ guidance. Under what circumstances do the respective states realize an individuals’ guidance? When do reflection, intentions, or goal-representations merely causally influence an individual’s acts? When do image-like goal-representing states contribute to an action at the sub-individual level alone? Guidance in Primates suggests an answer to these questions. In visual search we have seen that the central executive system controls processes by activating an image-like search goal in working memory. The activation of the search goal consists in the central executive system’s allocation of central resources to that search goal. We might say that the individual then guides his search through an image-like goal-representation. The allocation of central executive control to goal-representation and search process marks that representation’s role as a guider of the search. Absent such allocation of central resources, the goal-representation at best causally influences the individual’s guidance. If a goal-representing state is beyond the central executive system’s reach, that state may occur at the level of a sub-system alone. And a similar explanation is now available for cases in which an individual’s reflection or a propositional intention helps guide her act. We might say that the individual guides his search through reflection or intention whenever her central executive system activates – allocates resources to – a reflective state or an intention in similar ways. What unifies different exercises of agency, on this picture, is the central executive system’s control. What distinguish different exercises of agency are the capacities and states through which an act is guided.
In these ways, Guidance in Primates promises to contribute to understanding responsible agency. As I said earlier, not all primates are answerable for their acts. Responsibility-based accounts hence misidentify what constitutes agency. But they do draw attention to a particularly important type of agency: agency that individuals are answerable for. Central executive control is plausibly not sufficient for holding individuals responsible for their acts – monkeys will hardly be held responsible for their doings. But central executive control may well be necessary for the kinds of criticisms and assessments that the responsibility-based accounts invoke to explicate agency. By addressing the attributability-constraint in the way sketched, my proposal promises to integrate important lessons from all major accounts of agency.
5.3 Addressing the Comprehensiveness-Constraint
Appeal to central executive control will most likely not, however, help us meet the comprehensiveness-constraint. This constraint requires that an account of agency should endeavor to explain all instances of agency. An account of agency should at least hold some promise to account for exercises of agency across all species. Guidance in Primates promises to account for guidance in a much wider range of animals than most of the proposals that have recently been made in action theory. We know that primates have a central executive system. Many animals, such as mammals and possibly birds, exhibit behavior that is best explained by appeal to executive functions. But for many other actual agents it is at best unclear whether they have a central executive system. We do not know, at this point, whether the behavior of sharks and spiders exhibits the three executive functions switching, maintenance, and inhibition. Animals such as jellyfish, worms, and ciliates may not even have psychologies. So it seems that these animals could not have the psychological competencies that characterize the central executive system. Reflection on guidance in primates does, however, motivate a conjecture about the constitution of agency, which respects the comprehensiveness-constraint:
Constitution of Guidance
The control of a process by an individual’s central structures (partly) constitutes that individual’s guidance.
‘Central structures’ are structures that function to integrate and coordinate the activity of individuals’ sub-systems. Such structures ‘control’ a process if they actually determine and shape an individual-level process by integrating and coordinating the activity of sub-systems. The conjecture leaves open whether such central structures are psychological structures or of some other nature. The conjecture hence involves no commitment to specific psychological competences on an agent’s part. The conjecture allows for agents that do not have a psychology at all. Indeed, the conjecture allows that the nature of the relevant central structures and the ways in which they control activity may differ from kind of agent to kind of agent. So the conjecture meets the comprehensiveness-constraint in that it at least promises to apply to all agents, across species.
Why might we think that such central structures’ control might constitute individuals’ guidance? First, reflection on actual agents provides initial support for the claim. Attributions of a central executive system to primates, possibly mammals, and maybe even birds, seem empirically plausible. Mammals and birds certainly seem to have a psychological system functioning as an integrator and coordinator of these animals’ activities. Central structures of sorts can be found in arthropods, cephalopods, and many fish. One may even argue that the membrane of single-cell organisms such as ciliates constitutes a central structure. A cursory overview of actual agents at least suggests that they all have central structures of the kind that the conjecture appeals to.
Second, reflection on the evolution of actual agents provides some support for the claim that they all have central structures. The agents that we know seem to exhibit some degree of complexity. These animals have different types of behaviors available. They typically are capable of processing different types of stimulation. Which stimulus they process and which behavior they exhibit typically depends on their internal state. From an evolutionary point of view, it would seem beneficial for these agents, if they had central structures that integrate and coordinate stimuli from different sources with their internal state. The agents would seem to benefit from these structures’ selecting some specific type of behavior on that basis. So it would seem beneficial if these animals had central integrating and coordinating structures.
Third, armchair reflection suggests that agents should exhibit a certain unity. Actions are doings by agents. Many agents – and plausibly all actual agents – draw upon different capacities for their actions. Such agents draw upon different behavioral abilities, different perceptual capacities, and different types of motivation. In virtue of what are these different processes unified so as to yield the agent’s act? It seems plausible that some central structure that coordinates and integrates the activities of relevant sub-systems should yield this unification. Different sub-systems can contribute to each of an individual’s acts. But the contributions of different sub-systems would be coordinated and integrated by the same central structure, in each act. Appeal to such a central structure would help us explain a sense in which agents exhibit unity.
I offer Constitution of Guidance as a conjecture about the constitution of agency. I hope to verify the conjecture in future work.
In this paper I have argued that we can explicate individuals’ guidance by appeal to central executive control. I have suggested that exercises of primates’ guidance constitutively involve central executive control. And I have conjectured that control by integrating and coordinating central structures is a constitutive condition on all agency. I have developed my explication(s) by reflecting on a close examination of empirical research on episodes in which individuals guide their acts. I thus hope to have illustrated the benefits of such a methodology in action theory.
Anderson, B. 2013. “A Value-Driven Mechanism of Attentional Selection.” Journal of Vision 13(3): 1-16
Anscombe, G.E.M. 1953. Intention. Cambridge, MA: Harvard University Press
Baddeley, A. D. 2007. Working Memory, Thought, and Action. Oxford: Oxford University Press.
Barth, F. 2002. A Spider’s World. Senses and Behavior. Springer: Heidelberg.
Bishop, J. 1989. Natural Agency. An Essay on the Causal Theory of Action. Cambridge: Cambridge University Press.
Block, N. & Stalknaker, R. 1999, ”Conceptual Analysis, Dualism, and the Explanatory Gap.” Philosophical Review 108: 1-46
Brand, M. 1984. Intending and Acting. Cambridge, MA: MIT Press.
Bratman, M. 1987. Intentions, Plans, and Practical Reason. Cambridge, MA: Harvard University Press.
Buehler, D. [MS]. “Animal Agency.”
Buehler, D. [MS]. “Unsophisticated Agency.”
Buehler, D. [MS]. “The Individual/Sub-System Distinction.”
Buehler, D. 2014, Psychological Agency – Guidance of Visual Attention, UCLA Dissertation
Burge, T. 1979. “Individualism and the Mental.” in T. Burge (2007), Foundations of Mind (pp. 392-419), OUP: Oxford
Burge, T. 2007, Foundations of Mind, OUP: Oxford
Burge, T. 2009, “Primitive Agency and Natural Norms,’ Philosophy and Phenomenological Research 79: 251-287
Burge, T. 2010, Origins of Objectivity, OUP: Oxford
Burge, T. 2010a. “Steps Towards Origins of Propositional Thought.” Disputatio 4(29): 39-67
Carnap, R. 1947, Meaning and Necessity, UCP: Chicago
Carrasco, M. 2011. “Visual Attention: The Past 25 Years.” Vision Research (51): 1488
Chisholm, R. 1976. Person and Object. A Metaphysical Study. La Salle, IL: Open Court Publishing.
Chisholm, R. 1978. “Comments and Replies.” Philosophia 7: 597-636
Chun, M. & Nakayama, K. 2000. “On the Functional Role of Implicit Visual Memory for the Adaptive Deployment of Attention across Scenes.” Visual Cognition 7(1/2/3): 65-81
Davidson, D. 1980. Essays on Actions and Events. Oxford: Oxford University Press.
Emery, N. & Clayton, N. 2004. “Comparing the Complex Cognition of Birds and Primates.” in L. Rogers & G. Kaplan, Comparative Vertebrate Cognition (pp. 3-55), Springer: New York
Folk, C., Remington, R. & Johnston, C. 1992. “Involuntary Covert Orienting is Contingent on Attentional Control Settings.” Journal of Experimental Psychology: Human Perception and Performance 18(4): 1030-1044
Frankfurt, H. 1978, “The Problem of Action,” in his The Importance of What We Care About, CUP: New York (1988)
Gazzaley, A. & D’Esposito, M. 2006. “Unifying the Prefrontal Cortex: Executive Control, Neural Networks, and Top-Down Modulation.” In B. Miller & J. Cummings (eds.), The Human Frontal Lobes. Functions and Disorders. (pp. 187-207) New York: Guilford Press.
Geisler, W. & Cormack, L. 2011. “Models of Overt Attention.” In S. Liversedge, I. Gilchrist & S. Everling (eds.), The Oxford Handbook of Eye Movements (pp. 439 – 454), OUP: New York
Godijn, R. & Theeuwes, J. 2003. “The Relationship between Exogenous and Endogenous Saccades and Attention.” In J. Hyona, R. Radach & H. Deubel (eds.), The Mind’s Eyes: Cognitive and Applied Aspects of Eye Movements (pp. 3-26), Elsevier: Amsterdam
Goldman, A. 1970. A Theory of Human Action. Princeton, NJ: Princeton University Press
Güntürkün, O. 2005. ”The Avian ’Prefrontal Cortex’ and Cognition.” Current Opinion in Neurobiology 15: 686-693
Hieronymi, P. 2009. “Two Kinds of Mental Agency.” In Lucy O’Brien & Matthew Soteriou (eds.), Mental Actions, Oxford: Oxford University Press.
Hollingworth, A. 2014. “Guidance of Visual Search by Memory and Knowledge.” In M. Dodd & J. Flowers (eds.), The Influence of Attention, Learning, and Motivation on Visual Search, Nebraska Symposion on Motivation
Hornsby, J. 1980. Actions. London: Routledge.
Kripke, S. 1980. Naming and Necessity. Harvard University Press: Cambridge, MA
Kristjansson, A. & Campana, G. 2010. “Where Perception Meets Memory: A Review of Repetition Priming in Visual Search Tasks.” Attention, Perception & Psychophysics 72(1): 5-18
Korsgaard, C. 2009. Self-Constitution. Agency, Identity, and Integrity. Oxford: Oxford University Press.
Lavie, N., Hirst, A., De Fockert, J. & Viding, E. 2004. “Load Theory of Attention and Cognitive Control.” Journal of Experimental Psychology: General 133(3): 339-354
Lavie, N. & Dalton, P. 2014. “Load Theory of Attention and Cognitive Control.” In A. Nobre & S. Kastner (eds.), The Oxford Handbook of Attention (pp. 56-75), Oxford: Oxford University Press.
Leonard, C. & Egeth, H. 2008. “Attentional Guidance in Singleton Search: An Examination of Top-Down, Bottom-Up, and Intertrial Factors.” Visual Cognition 16(8): 1078-1091
McDowell, J. 1994. Mind and World. Cambridge, MA: Harvard University Press.
McDowell, J. 2007. “What Myth?” Inquiry 50(4): 338-351
McDowell, J. 2010. “What is the Content of an Intention in Action?” Ratio 23: 415-432
Mele, A. 1992. The Springs of Action. Understanding Intentional Behavior. Oxford: Oxford University Press
Miller, E. & Cohen, J. 2001. ”An Integrative Theory of Prefrontal Cortex Function.” Annual Review of Neuroscience 24: 167-202
Miyake, A., Friedman, N., Emerson, M., Witzki, A., Howerter, A. & Wager, T. 2000. “The Unity and Diversity of Executive Functions and Their Contributions to Complex “Frontal Lobe” Tasks: A Latent Variable Analysis.” Cognitive Psychology 41: 49-1
Munakata, Y., Hard, S., Chatham, C., Depue, B., Banich, M. & O’Reilly, R. 2011. ”A Unified Frameworke for Inhibitory Control.” Trends in Cognitive Sciences 15: 453-459
Munakata, Y., Snyder, H. & Chatham, C. 2012. ”Developing Cognitive Control: Three Key Transitions,” Current Directions in Psychological Science 21(2): 71-77
Nanay, B. 2014, Between Perception and Action, OUP: Oxford.
Nixon, M. & Young, J. 2003. The Brains and Lives of Cephalopods. OUP: Oxford.
Oh, S. & Kim, M. 2004. “The Role of Spatial Working Memory in Visual Search Efficiency.” Psychonomic Bulletin & Review 11(2): 275-281
Olivers, C. & Eimer, M. 2011. “On the Difference between Working Memory and Attentional Set.” Neuropsychologia 49: 1553-1558
Olivers, C., Peters, J., Houtkamp, R. & Roelfsma, P. 2011. “Different States in Visual Working Memory: When it Guides Attention and When it does Not.” Trends in Cognitive Sciences 15(7): 327-334
O’Reilly, R. & Frank, M. 2006. ”Making Working Memory Work: A Computational Model of Learning in the Frontal Cortex and Basal Ganglia.” Neural Computation 18: 283-328
O’Shaughnessy, B. 1973. “Trying (as the mental “pineal gland”)” The Journal of Philosophy 70(13): 365-386
O’Shaughnessy, B. 1980. The Will. A Dual Aspect Theory. Cambridge: Cambridge University Press.
Peacocke, C. 2007. “Mental Action and Self-Awareness (I).” In J. Cohen & B. McLaughlin (eds.), Contemporary Debates in the Philosophy of Mind, Oxford: Blackwell.
Pears, D. 1975. “Sketch for a Causal Theory of Wanting and Doing.” In D. Pears, Questions in the Philosophy of Mind (pp. 97-141), London: Duckworth.
Posner, M. 1980. “Orienting of Attention.” Quarterly Journal of Experimental Psychology 32
Putnam, H. 1970. “Is Semantics Possible?” In H. Putnam (1975), Mind, Language, and Reality. Philosophical Papers Vol. 2 (139-152), CUP
Quine, W.V.O. 1960, Word and Object, MIT Press: Cambridge, MA
Searle, J. 1983. Intentionality. An Essay in the Philosophy of Mind. Cambridge: Cambridge University Press.
Shadmehr, R. & Wise, S. 2005. The Computational Neurobiology of Reaching and Pointing. A Foundation for Motor Learning. Cambridge: MIT Press.
Sleigh, M. 1975. The Biology of Protozoa. Edward Arnold: London.
Soto, D., Humphreys, G. & Heinke, D. 2006. “Working Memory can Guide Pop-Out Search.” Vision Research 46: 1010-1018
Soto, D., Hodsoll, J. Rotshtein, P. & Humphreys, G. 2008. “Automatic Guidance of Attention from Working Memory.” Trends in Cognitive Science 12(9): 342-348
Stokes, M. & Duncan, J. 2014. “Dynamic Brain States for Preparatory Attention and Working Memory.” In A. Nobre & S. Kastner (eds.), The Oxford Handbook of Attention (pp. 152-182), Oxford: Oxford University Press
Theeuwes, J. 1991. “Exogenous and Endogenous Control of Attention: The Effect of Visual Onsets and Offsets.” Perception & Psychophysics 49(1): 83-90
Todd P., Hills, T. & Robbins, T. (eds.) 2012, Cognitive Search: Evolution, Algorithms, and the Brain, MIT Press: Cambridge, MA
Velleman, D. J. 2000. The Possibility of Practical Reason. Oxford: Oxford University Press
Walther, D. & Fei-Fei, Li 2007. “Task-Set Switching with Natural Scenes: Measuring the Cost of Deploying Top-Down Attention.” Journal of Vision 7(11): 1-12
 My special thanks to Tyler Burge. I have benefitted from comments by Harry G. Frankfurt, Pamela Hieronymi, Michael Rescorla, Miguel Ángel Sebastián, James Stazicker, and participants of the UCLA Mind and Language Workshop and UNAM-IIF’s TEC discussion group. Thanks also to audiences at UCLA, Indiana University Bloomington, and at the National Research University in Moscow.
 I indicate representational contents by underlining them.
 Frankfurt, H. 1978
 Burge, T. 2007, 163ff.; Cf. also Carnap, R. 1947; Quine, W.V.O. 1960, Chapter V.
 Here are two examples of empirically grounded explications that deepen conceptual understanding: Some contracts are oral and Whales are mammals. In both cases, the explication of the concepts provides a condition on being the relevant kind contract or whale. In both cases, the relevant explication depends on empirical information about whales and contracts for its warrant.
 Harry Frankfurt (personal communication) agrees that the lack of progress for the exclusively armchair methodology suggests appeal to empirical knowledge as a viable alternative.
 James Stazicker helpfully suggested describing the method as an attempt to achieve a reflective equilibrium between intuitions about cases, armchair reflection on the concept, and empirical research.
 This methodology has led to major advances in other parts of philosophy, especially in the philosophy of perception. Cf. Burge, T. 2010. So far, we lack an application of this methodology in action theory.
 A longer version of this paper contains a section addressing several objections.
 Frankfurt, H. 1978
 For the most powerful recent criticism along these lines, see Burge, T. 2009.
 Velleman, D. 2000, 30; Korsgaard, C. 2009, 26.
 Anscombe, G.E.M. 1953, 9; see also Hieronymi, P. 2009, 138
 Davidson, D. 1980, 46. Ibid., Essays 1 and 5.; Searle, J. 1983; Goldman, A. 1970; McDowell, J. 1994; McDowell, J. 2007; Mele, A. 1992; Bratman, M. 1987; Brand, M. 1984; Bishop, J. 1989; Pears, D. 1975. I intend for this family to include accounts on which the act-constituting state is a belief-desire pair, a conceptual desire, or a reason.
 Mele, A. 1992, 222; cf. also Nanay, B. 2014.
 Cf. Velleman, D. 2000, 127& 137. Our deepest intuition about agency is that actions are doings by agents, that actions have agents as their source. We find this intuition expressed throughout the history of philosophy, e.g. by Aristotle (DA II.4, 415b10/2; MA 9, 703a1/3), or Kant (B567/9; Lectures on Metaphysics, 29:822 & 29:903).
 Burge 2010, 469ff.
 Shadmehr, R. & Wise, S. 2005, Chapter 12, especially p. 225.
 The same demand applies to reflection- and intention-based accounts. It is related to the problem of deviant chains. Cf. Bishop, J. 1989.
 Chisholm, R. 1978, 627; also his Chisholm, R. 1976
 O’Shaughnessy, B. 1973, 369; O’Shaughnessy, B. 1980; Peacocke, C. 2007; Hornsby, J. 1980.
 Miyake, A. et. al. 2000; Miller, E. & Cohen, J. 2001; Munakata, Y. et al. 2011
 Burge, T. 2010a
 Cf. e.g. O’Reilly, R. & Frank, M. 2006; Gazzaley, A. & D’Esposito, M. 2006
 Posner, M. 1980
 Carrasco, M. 2011
 Geisler, W. & Cormack, L. 2011
 Cf. e.g. Theeuwes, J. 1991
 Folk, C., Remington, R. & Johnston, C. 1992
 The speed of their shifts toward the diamond exceeds that in cases of capture, where the suddenly appearing stimulus does not coincide with the search goal. Cf. Leonard, C. & Egeth, H. 2008.
 Godijn, R. & Theeuwes, J. 2003
 Kristjansson, A. & Campana, G. 2010
 Chun, M. & Nakayama, K. 2000
 Anderson, B. 2013; Hollingworth, A. 2014
 Soto, D., Humphreys, G. & Heinke, D. 2006; Soto, D., Hodsoll, J. Rotshtein, P. & Humphreys, G. 2008.
 Olivers, C. & Eimer, M. 2011; cf. also Olivers, C., Peters, J., Houtkamp, R. & Roelfsma, P. 2011; Stokes, M. & Duncan, J. 2014
 Walther, D. & Fei-Fei, Li 2007
 Oh, S. & Kim, M. 2004
 Lavie, N., Hirst, A., De Fockert, J. & Viding, E. 2004; Lavie, N. & Dalton, P. 2014
 Burge, T. 2010, 58
 Cf. Baddeley, A. D. 2007; Miller, E. & Cohen, J. 2001; Munakata, Y. et al. 2011; Munakata, Y. et al. 2012.
 Section 3.2
 I cannot here do justice to the complexities of the relation between central executive control and attributability to individuals. I address the issue in my “The Individual/Sub-System Distinction [MS].”
 Todd P. et al. 2012; Güntürkün, O. 2005; Emery, N. & Clayton, N. 2004
 Burge, T. 2010
 Cf. Burge, T. 2009. I more fully investigate this conjecture in my ‘Animal Agency [MS].’
 Nixon, M. & Young, J. 2003; Barth, F. 2002; Sleigh, M. 1975