Why is agency so important

How can one explain reasoning about abstract concepts, counterfactuals, and theoretical generalizations? And how can one explain that our agency is to a significant extent motivated, guided, and constrained by our long-terms plans and commitments?

There is, as we have seen, good reason to distinguish between different kinds of agency. The standard theory offers an account of what is, arguably, the most central kind of agency: intentional agency and the kind of unintentional agency that derives from it; see section 2. This can be distinguished from higher or more refined kinds of agency, such as self-controlled, autonomous, and free agency, and it can be distinguished from more basic kinds of agency that do not require the ascription of representational mental states.

Apart from that, there are several candidates for further kinds of agency. They include mental agency, shared agency, collective agency, relational agency, and artificial agency. In each case, we can ask whether the agency in question can be explained in terms of the standard conception and theory, or whether it is indeed a different kind of agency.

The main focus in this section will be on mental agency, and we will address the other candidates only very briefly. It may seem obvious that our mental lives are filled with mental action. We attend, consider, judge, reason, deliberate, accept, endorse, decide, try, and so on. It may seem that these are all things that we do. If we consider such cases through the standard theory of agency, we encounter immediately two difficulties. First, it seems that such mental occurrences are hardly ever, if ever, intentional actions.

According to the standard theory, an event is an intentional action of the type A only if the agent has an intention that includes A in its content. In the basic case, this would be an intention to A. In an instrumental case, this would be an intention to perform some other action B in order to A. Now, thoughts are individuated in part by their contents.

Take the thought that p. This is rather odd and problematic, because we would have to have the intention to think a certain thought before we think it. Second, there are problems with the central case of decision-making.

This seems, again, rather odd and problematic. Further, our reasons for making a decision to A are usually our reasons to A —they are reasons for performing the action. As reasons are usually reasons for action, it is again difficult to see how making a decision can ever be an action.

Considerations of this kind may lead one to conclude that thoughts are hardly ever, if ever, mental actions see Strawson It is not difficult to avoid this conclusion, as Mele , Ch. Consider again the central case of decision-making, and assume that making a decision consists in the formation of an intention.

According to the standard theory, the formation of an intention is an action if it is an intentional action under some description or if it is either identical with or generated by an intentional action; see section 2. Mele suggests that processes of decision-making are usually motivated by the intention to settle the practical question at hand.

This proposal avoids the problem outlined above. Suppose the agent decides to A. For this to be an action, it is not required that the agent has the intention to decide to A. For if the agent has the intention to settle the question by making a decision, making the decision is intentional under a description.

In particular, making a decision is then an intentional action and making the decision to A is then an unintentional action that is either identical with or generated by the intentional action of making a decision.

Mele b argues that remembering something is never an intentional action, because no one has ever the intention to remember the particular content in question. But there is nevertheless a closely associated intentional mental action that one might perform: intentionally trying to bring it about that one remembers the particular content in question.

See Shepherd for a defense of the view that decisions are intentional actions by construing them as extensions and conclusions of deliberative activity. Hieronymi takes a very different line. She thinks that we engage in mental agency whenever we settle the question of whether to do or whether to believe something, and she argues that this kind of mental agency differs from ordinary intentional agency, primarily due to a difference in control.

According to volitionist theories of agency, mental acts of willing choosing or trying are also different in kind from overt bodily actions. On such views, mental acts of willing are furthermore fundamental, in the sense that they are the source of overt agency Ginet ; McCann ; Lowe ; more on this in section 3.

Epistemic agency concerns the control that agents may exercise over their beliefs and other doxastic states. It is common to distinguish between two main positions: indirect doxastic voluntarism and direct doxastic voluntarism.

The former concerns the ways in which we may acquire or revise beliefs by doing research, evaluating the evidence, considering opposing opinions, and so on. It is fairly uncontroversial that we can exercise control over our beliefs in such indirect ways. In contrast, direct doxastic voluntarism is very controversial. It says that we have direct voluntary control over some of our beliefs, where voluntary control is usually understood as the kind of control that agents exercise in the performance of intentional actions.

A main issue here is that direct doxastic voluntarism appears to be incompatible with the nature of beliefs. One may argue that there is no fundamental difference in the control over action and belief-formation, because in both cases the control consists basically in reason-responsiveness. But this proposal overlooks the central role of intentions. According to the standard theory, actions must be initiated and guided by intentions, in addition to being responsive to reasons. The challenge is to find beliefs-formations that are initiated and guided by intentions in the same or similar way as intentional actions.

For a more extensive overview and references see Vitz Shared agency occurs when two or more individuals do something together such as carry a piece of furniture or sing a song. Collective agency occurs when two or more individuals act as a group in accordance with certain principles or procedures that constitute and organize the group. Research on shared and collective agency has flourished over the past two decades or so.

One central question has been whether shared and collective agency can be reduced to the agency of the individuals involved, or whether they are constitutive of different kinds of agency—whether they are, in some sense, something over and above individual agency. An account of collective agency in terms of the standard theory raises the question of whether it makes sense to attribute mental states and events such as desires, beliefs, and intentions to groups of individuals. For references and discussion see the entries on shared agency and collective intentionality.

The notion of relational agency derives from relational accounts of autonomy. According to feminist critiques, traditional accounts of autonomy are overly individualistic, insofar as they overlook or neglect the importance of interpersonal relationships in the development and sustenance of an autonomous individual. As Westlund points out, however, most traditional accounts are compatible with the feminist emphasis on interpersonal relationships as long as relationships and dependence on others are construed as being causally necessary for the development and sustenance of an individual agent.

Autonomy is genuinely relational only if interpersonal relationships and dependence are constitutive of autonomy. On this account, autonomy is an irreducibly relational kind of agency. For more on this see the entry on feminist perspectives on autonomy.

Finally, we turn briefly to the question of whether robots and other systems of artificial intelligence are capable of agency. If one presumes the standard theory, one faces the question of whether it is appropriate to attribute mental states to artificial systems see section 2.

If one takes an instrumentalist stance Dennett Ch. According to realist positions, however, it is far from obvious whether or not this is justified, because it is far from obvious whether or not artificial systems have internal states that ground the ascription of representational mental states. If artificial systems are not capable of intentional agency, as construed by the standard theory, they may still be capable of some more basic kind of agency.

According to Barandiaran et al. What is the nature of agency? How should we construe the relation between agents and actions? How can agency be part of the event-causal order? In this section, we will first turn to the three main approaches in the metaphysics of agency that provide three different frameworks for how to think about such metaphysical questions the event-causal, the agent-causal, and the volitionist framework.

After considering some problems and objections, we turn to an alternative approach that rejects the project of providing a metaphysics of agency dual standpoint theory. Finally, we briefly consider the individuation of actions and some further issues in the metaphysics of agency. According to an event-causal approach, agency is to be explained in terms of event-causal relations between agent-involving states and events.

According to an agent-causal approach, agency is to be explained in terms of a kind of substance-causation: causation by the agent, construed as a persisting substance. On this view, actions are events, and an event is an action just in case it has the right agent-causal history. On this view, volitions are the source of agency: an overt movement is an action just in case it is caused, in the right way, by a volition.

Volitions themselves are entirely uncaused and they are sui generis acts: they are acts in virtue of their intrinsic properties, not in virtue of some extrinsic or relational property such as having the right causal history.

This is also a non-reductive approach to agency, but it differs sharply from both the event-causal and the agent-causal framework in the important respect that it rejects the suggestion that all actions are events with a certain causal history Ginet ; McCann ; see also Lowe The event-causal framework is by far the most widely accepted view in the contemporary philosophy of mind and action. One reason for this is that the commitment to the event-causal framework is tantamount to a commitment to a very minimal and widely endorsed kind of naturalism, according to which any appeal to irreducible substance-causation or teleology is to be avoided.

Further, this commitment to the event-causal framework is sustained by a widespread dissatisfaction with alternative agent-causal and volitionist theories of agency. Some objections to agent-causal theories derive from more general objections to the notion of substance-causation, others address more directly the agent-causal account of agency. It has been argued, for instance, that appeal to substances leaves both the timing and the manner of causation mysterious Broad Further, it has been argued that substance-causation collapses into event-causation, once it is acknowledged that a substance has its causal powers in virtue of its properties Clarke Ch.

A common objection to volitionist accounts is that they generate a regress of mental acts Ryle Arguably, though, this objection begs the question. The view holds that overt actions are to be explained in terms of volitions. This, however, points also to the reason why the view is widely rejected. Volitionist theories stipulate as primitive what appears to be in need of explanation.

Moreover, if, as most contemporary philosophers would assume, volitions are realized by events in the brain, the view appears to be in tension with the fact that there are no events in the brain that are entirely uncaused. In the s and 60s, several philosophers argued that the event-causal framework is incoherent. It is widely agreed now that this attack was unsuccessful the most influential reply is due to Davidson ; see also Goldman — In general, the problem is that it seems always possible that the relevant mental states and events cause the relevant event a certain movement, for instance in a deviant way: so that this event is clearly not an intentional action or not an action at all.

It is common to distinguish between cases of basic deviance and consequential deviance also called primary and secondary deviance. A murderous nephew intends to kill his uncle in order to inherit his fortune. As it turns out, this pedestrian is his uncle. This is a case of consequential deviance Chisholm In a standard case of basic deviance Davidson , a climber intends to rid himself of the weight and danger of holding another man on a rope by loosening his grip.

This intention unnerves him so that it causes him to loosen his hold on the rope. The difference between the cases is best explained in terms of the distinction between basic and non-basic actions. He successfully performs several basic actions, but it is a sheer coincidence that he brings about the intended end. The climber, in contrast, does not perform any action at all. The mental antecedent causes a movement that would have been a basic action, had the causal chain not been deviant.

Any event-causal theory of agency must require that the relevant mental attitudes cause the action in the right way. The right way of causation is non-deviant causation. The challenge is to spell out what non-deviant causation consists in within the event-causal framework; without, in particular, any appeal to some unanalyzed notion of agent-causation or control.

Davidson was pessimistic about the prospects for finding an event-causal account of non-deviant causation, and he suggested that the standard theory is best understood as providing only necessary conditions for agency.

Goldman suggested that giving an account of non-deviant causation is an empirical rather than a philosophical task. Since then, however, most proponents of the event-causal approach have acknowledged that the problem of deviant causal chains is a serious philosophical problem, and various solutions have been proposed see Peacocke ; Brand ; Bishop ; Mele ; Schlosser , ; Wu Sometimes it is suggested that the problem of deviant causal chains is merely a symptom of the deeper problem that event-causal theories altogether fail to capture agency, because they reduce actions to things that merely happen to us Lowe 9, for instance.

Put differently, this challenge says that the event-causal framework is deficient because it leaves out agents: all there is, on this view, is a nexus of causal pushes and pulls in which no one does anything Melden ; Nagel ; see also Velleman According to Mele Ch. However, sometimes the challenge is raised in order to motivate alternative agent-causal or volitionist theories of agency, and the main proponents of agent-causal and volitionist theories maintain that their views are compatible with naturalism.

They would argue that it is a mistake to presume that the event-causal order exhausts the natural order of things. Further, the disappearing agent objection is not always put forward as a general objection to the event-causal framework. As we have seen section 2. In his reply, Mele Ch. Instances of deficient agency can be explained in terms of various capacities or properties that the agent does not possess, exercise, or instantiate; capacities and properties such as conscious awareness, reflective awareness, reason-responsiveness, self-control, self-governance, and so on.

Further, doing so creates a rather implausible dichotomy between a kind of agency in which the agent does participate and a kind of agency in which the agent does not participate Schlosser Others, yet, press the disappearing agent objection in order to motivate a dual standpoint theory. According to dual standpoint theories, agency cannot be explained from any theoretical standpoint or metaphysical framework.

Agency can only be understood from a practical and normative standpoint Nagel ; Korsgaard ; Bilgrami , for instance. Usually, dual standpoint theories do not reject metaphysics as such, and they often provide a metaphysical framework of their own. But they reject both reductive and non-reductive theories of agency, and they reject, in general, the notion that we can have a metaphysical account of what the exercise of agency consists in.

They align themselves naturally with non-causal theories of reason explanation see section 2. Both views tend to emphasize the normative and irreducibly teleological nature of reason explanation and, hence, agency. Dual standpoint theories have received relatively little attention in the philosophy of action.

To many, it seems that such views are deeply unsatisfactory precisely because they refuse to face a central question in the metaphysics of agency: how can agents exercise control over their actions in a world in which all movements can be explained in terms of event-causation?

It seems that this is in need of explanation, and it seems that this requires a metaphysics of agency see Bishop ; Schlosser Nelkin has questioned the coherence of dual standpoint theories on the basis of an argument for the claim that they entail commitments to contradictory beliefs about free will. We now turn, in brief, to some further issues in the metaphysics of agency.

The first concerns the individuation of actions. You flick the switch, turn on the light, illuminate the room, and you thereby also alert the burglar. How many actions do you perform? According to coarse-grained or minimizing views on the individuation of actions, you perform one action under different descriptions Anscombe ; Davidson According to fine-grained or maximizing views, how many actions you perform depends on how many act-properties are instantiated.

If you instantiate four act-properties, then you perform four distinct actions Goldman ; see also Ginet According to a third alternative, actions can have other actions as their components or parts Thalberg ; Ginet According to all three views, actions are events, and the individuation of actions derives from different views on the individuation of events see the entry on events.

This is partly because it is now widely agreed that the individuation of actions has little or no bearing on other issues. To illustrate, the question of whether agency is to be explained within an event-causal or an agent-causal framework bears directly on various issues in the debate on free will and moral responsibility see the entry on free will.

But event-causal and agent-causal theories are both compatible with coarse-grained and fine-grained views on the individuation of actions. Similarly, it seems that the views on the individuation of actions have no substantial bearing on the question of whether or not reason explanations are causal explanations. A related issue is whether actions are to be identified with the outcomes of causal processes or with the processes themselves.

According to most versions of event-causal and agent-causal theories, an action is an event that is caused in the right way: the action is identical with or constituted by the outcome of that process.

This issue has also not received much attention. Again, this is mainly because it is widely assumed that this issue has little or no substantial bearing on more fundamental issues in the metaphysics of agency and on debates outside the philosophy of action. Another issue in the metaphysics of agency that has received more attention in the recent debate is the nature of omissions in particular, intentional omissions.

According to Sartorio , an intentional omission is the absence of an action that is caused by the absence of an intention.

She argues, on the basis of this account, that intentional omissions cannot be accommodated easily by the standard theory. In reply, Clarke a has argued that in cases of intentional omission the agent usually does have an intention not to act that plays an important causal role, and he has identified various parallels between intentional actions and intentional omissions.

On his view, there are no major obstacles to an account of intentional omissions that is compatible and continuous with the standard theory of intentional action.

Further, he argues that a failure to account for intentional omissions would not obviously be a shortcoming of a theory of intentional action. There are, after all, significant differences between actions and omissions, and so we should not expect that a theory of action provides all the resources that are required for an account of omissions. For more on this see Clarke This assumption is part and parcel of the standard theory and of numerous psychological theories of intentional action and motivation Fishbein and Ajzen ; Locke and Latham ; Heckhausen ; Gollwitzer ; Austin and Vancouver , for instance.

There are, however, various empirical findings from psychology and cognitive neuroscience that have been taken to show that this commonsense assumption is unwarranted, and that have raised interesting and challenging questions concerning the role of consciousness in the initiation and guidance of agency. This section provides an overview of the most relevant research. An early and highly influential source of the skepticism concerning the causal relevance of our reasons is a theoretical review by Nisbett and Wilson This article reports numerous experiments and studies in which participants appear to construct or confabulate rationalizing explanations by giving reasons that could not possibly have been the reasons they acted for.

Despite some rather serious methodological problems White , this research has achieved and retained the status of textbook knowledge within psychology and cognitive science. Moreover, it has been taken to show that ordinary reason explanations are not causal explanations, even though the authors themselves rejected this conclusion. On their view, the evidence shows, first and foremost, that verbal reports of mental states are based on self-interpretation theorizing or rationalization , rather than on direct or introspective access.

They noted that this epistemic view is perfectly compatible with the assumption that we can and often do give the actual causes of our actions when we give an ordinary reason explanation. The upshot is that, even if the proposed epistemic view is correct, there is nothing in the evidence which shows that reason explanations cannot be causal explanations, and there is nothing in the evidence which shows that reason explanations are usually not causal explanations. It seems that the empirical evidence in support of situationism raises a challenge for our commonsense conception of agency.

According to situationism, empirical research shows that commonsense explanations of actions in terms of character traits such as honesty, kindness, or courage are systematically mistaken or inaccurate, because this research shows that the actions in question are better explained in terms of situational features Ross and Nisbett ; Harman ; Doris Moreover, the interpretation of the empirical evidence in question and the argument for situationism have been controversial Sreenivasan , for instance.

It has been argued, however, that this evidence raises the further question of whether we are genuinely reason-responsive. The evidence suggests that our actions are, under certain conditions, driven by situational and morally irrelevant factors even when there are salient moral reasons to act otherwise.

This suggests that we or most of us are not as reason-responsive as we would like to think. But it is controversial whether or not the evidence supports any stronger claims than that for more on this see Nelkin ; Schlosser ; Vargas In the Libet experiment Libet , participants were instructed to initiate a simple and predefined movement when the wish or urge to do so arises.

During this, EEG measurements were taken to record the readiness potential, a brain potential that was known to precede intentional movements. The main finding was that the readiness potential precedes the occurrence of the conscious wish or urge to move by about ms. According to Libet, this shows that movements are not consciously initiated and that we do not have free will in the sense we commonly think we do Libet The methodology of this experiment has been scrutinized extensively and criticized on a number of points.

Some of those methodological issues have been addressed in follow-up experiments Soon et al. Further, it has been argued that the experiment creates a very unusual and artificial context in which participants are instructed to decide spontaneously. Due to this, it is questionable that the results of the experiment can be generalized Keller and Heckhausen ; Roskies ; Waller ; Schlosser Schurger et al. According to this model, the timing of the movement in the Libet experiment is determined by random threshold crossings in spontaneous fluctuations in neural activity.

In particular, the model says that a decision when to move is determined by random threshold crossings only when it is not constrained by any evidence or reasons for action. The fact that this model has been tested successfully supports the claim that the results from the Libet experiment and from similar follow-up studies do not generalize, because most of our everyday decisions clearly are constrained by evidence and by reasons for action.

Wegner provided evidence of dissociations between the sense of agency and the actual exercise of agency, and he argued that the model of apparent mental causation provides the best explanation of the data. This view has been strongly criticized for conceptual ambiguities and argumentative flaws see also section 4.

This is where CultureSpan Marketing can help your internal team like few agencies can. CultureSpan Marketing is a full-service firm specializing in reaching general and Hispanic markets in a wide range of industries, including: healthcare, banking and credit unions, highway development, airports, gaming and large-scale utilities just to name a few. Want to learn more about how CultureSpan can help you? Call me at or shoot an email to gabe.

Get the latest from CultureSpan Marketing! Skip navigation CultureSpan Marketing. Insights Blog Newsletters Whitepapers. I like that mention that working with a marketing or creative agency allows you to connect with resources that would be hard to come by without them. They typically have the tools and resources that any given business may not have but would need in order to help grow their publicity.

I think that when a marketing is smaller, they also have the ability to work with clients on an even more personal level, which benefits the business even more. Thanks for the comments Cameron. Much appreciated. We've seen something interesting in our industry that I describe as an expansion and contraction of creative talent.

Expansion being, "We need to hire an agency who can get our marketing, advertising and social done quickly and efficiently". The contraction is more like "We're bringing it all in house. Inevitably the client's marketing needs will reach critical mass, which will require help from the outside. Glad to see that companies like Utoka and CultureSpan are geared up and ready for when that happens.

I think that you make a good point that a good advertising agency will be able to adapt to your business's culture and go for your goals just like you do. I'd think that they'd also be able to do more for your business and make it look more professional in the ads they make for you.

It certainly sounds like it would be a big benefit to any company to work with an advertising agency. Its' interesting to note how you highlighted these agencies bring new levels of creativity, where they can solve problems and approach solutions in new and unexpected ways.

Anyone interested in working with these agencies will be happy to keep in mind the point you made here. I started looking into creative advertising agencies here in NYC after having a conversation with my brother who was looking to collaborate with an agency as such.

Instead, the brain appears to actively construct the sense of agency, and because of this, our experiences of agency can be quite divorced from the facts of agency. The second reason is that these lapses reveal something quite remarkable about our sense of agency: its impressive flexibility.

Beyond the examples I have given here, we see over and over again that people come to experience control over outcomes in many weird and wonderful situations.

The voodoo doll is another example; people taking part in this practice genuinely believe that sticking a pin in an effigy of someone causes actual physical harm in that person.

At first blush this inference seems irrational. However, examples like the voodoo doll actually hint at the adaptability and flexibility of the agency processing system. It is worth reminding ourselves that causal mechanisms are quite opaque in a lot of modern technology consider the simple act of tapping on a keyboard and seeing a letter appear on the screen in front of you — there are a lot of steps in this causal chain that are hidden from you.

Despite this causal opacity, we feel in control of these interactions. So the flexibility that might make us vulnerable to agency errors in things like placebo buttons and voodoo dolls, can also allow our experience of agency to extend into new domains and track the rapidly changing agentic structure of our environment.

Rather than our agency processing system breaking down with the development of tools, which have changed and extended our agentic capabilities, it has been flexible and adaptable, allowing us to accommodate these changes. The number of scientific investigations of sense of agency has increased considerably over the past 20 years or so. This increase is despite the fact that experiments on sense of agency face certain methodological problems. A major one is that the sense of agency is phenomenologically thin Haggard, That is, when we make actions we are typically only minimally aware of our agentic experiences.

This is quite unlike conscious experience in other modalities, especially vision, where our experiences are typically phenomenologically strong and stable. What this means is that sense of agency can be difficult to measure. As a result of this, experimenters have had to be quite inventive in order to develop paradigms that capture this rather elusive experience.

You can generally group these paradigms into implicit or explicit measures. Implicit measures assess a correlate of voluntary action and infer something about the agentic experience on the basis of this.

In these paradigms no one is ever asked, directly, about their agentic experience. Probably the most widely used implicit measure of sense of agency is intentional binding for a review see Moore and Obhi, This was developed by Haggard et al. Haggard et al. This change in time perception is taken to be an implicit marker of sense of agency. Other implicit measures of sense of agency include sensory attenuation paradigms. It has been shown that the perceived intensity of the sensory consequences of voluntary action is lower than for passive movements Blakemore et al.

This can explain why we are unable to tickle ourselves Blakemore et al. Explicit measures, on the other hand, directly ask the participant to report something about their agentic experience. These measures are more intuitive but they can be vulnerable to problems like demand effects.

A number of these paradigms require participants to make action recognition judgments. Typically the participant makes an action, but does not directly see that action.

Instead they are shown some kind of feedback on a screen. Importantly, the experimenter ensures that there is some uncertainty over the agent of the action being displayed. An example of this kind of task was used by Farrer et al. They had participants perform regular finger tapping movements while wearing a glove. They could not directly see these movements, and instead they were shown video feedback of the movement on the screen. A delay was inserted between the movement and the feedback presented to the participant.

The participants were not aware that the movement was always their own, and instead were led to believe that the movement was either their own or an experimenter performing the same movement, and that this could switch at any time. Farrer et al. Instead, the participant is required to make a judgment about the feedback itself.

An example of this kind of action monitoring task can be seen in an experiment carried out by Synofzik et al. In this experiment participants made pointing movements under a screen, meaning that they could not directly see the movement. On the screen participants were shown a visual marker white disk that tracked the pointing movement. This marker was rotated by varying degrees relative to the actual movement. The participants had to indicate the direction in which the visual feedback was rotated relative to the actual movement.

This gave the experimenters a measure of action awareness and, more specifically, sensitivity to distortions in action-relevant feedback. A final kind of explicit measure requires participants to report on their feeling of agency for certain action outcomes that their movements might have caused.

A simple example of this would involve a key press that causes an outcome after a variable delay. Participants would then judge how much they felt their action caused the outcome. A common finding is that such causal judgments are stronger for shorter delays e. Interestingly, this kind of explicit measure taps into a slightly different aspect of the agentic experience compared with the other two kinds of explicit measure described in this section.

Although, both of these are central to the agentic experience, this difference is often overlooked and not very well-understood. The comparator model takes as its starting point the motor control system. We now know a great deal about the computational processes underpinning the control of voluntary movement see Wolpert and Miall, , for a review.

According to the comparator model, some of these processes also inform the sense of agency. On this view, our actions start with intentions or goals, which enables a representation to be formed of the desired state of the motor system.

Controllers within the motor control system then use this information about the desired states to generate a motor command. This motor command produces a movement, which changes the state of the motor system, and generates sensory feedback. On the basis of this information the new state of the system can be estimated.

This estimate is compared with the desired state at a comparator. If there is a mismatch then an updated motor command is issued. This process can continue until the desired state is achieved indicated by the absence of a mismatch at the comparator.

The issue with a motor system operating only in this way is that it is slow to respond to error. Because of this, the organism is vulnerable. The solution, it would appear, is to have an additional predictive component within the motor system, and it is this that is thought to be particularly relevant to sense of agency.

This includes predictions about changes to the motor system as well as the sensory consequences resulting from those changes. On the basis of these predictions, a representation of the predicted state of the system can be formed, and this representation can be compared both with the desired state of the system and with the actual state of the system.

The former comparison is important for motor control, as it allows the organism to rapidly adjust motor commands in advance of incorrect actions being performed. The latter comparison is thought to be important for sense of agency. According to the comparator model, the output of the comparison between predicted and actual states determines whether or not we feel a sense of agency.

If there is a match, then we feel a sense of agency; if there is mismatch then we do not. A number of studies support this idea that sense of agency is closely tied to sensorimotor processes. For example, as predicted by the comparator model, it has been shown that the perceived intensity of self-produced tactile sensations is attenuated relative to externally produced ones Blakemore et al.

It has also been shown that sensorimotor prediction contributes to intentional binding Moore and Haggard, Finally, mismatches between expected and actual sensory feedback influence action recognition judgments e. Whereas the comparator model places a heavy emphasis on the contribution of the motor system to sense of agency, the theory of apparent mental causation explicitly downplays this contribution.

Indeed, according to this theory, it is because we do not have conscious access to the motor control system that our sense of agency can, at times, be so misleading, as seen in phenomena like voodoo dolls and placebo buttons. According the theory of apparent mental causation when we make a voluntary action there is an unconscious causal pathway that is responsible for the action. This pathway corresponds to the workings of the motor control system.

There is also an unconscious causal pathway that is responsible for the associated thoughts about actions i. In addition to these unconscious causal pathways, there are certain events that we are conscious of, namely the intention to act and the act itself. If our intention to act happens before we act, is consistent with the action, and is the only plausible cause of the action, then we feel as though we have caused the action. For example, it has been shown that priming thoughts about an upcoming action fosters an illusory sense of agency for that action Wegner and Wheatley, ; Wegner et al.

It has also been shown that manipulating high-level contextual information about an action in the form of causal beliefs alters sense of agency, as measured by intentional binding Desantis et al. These two theories, the comparator model and the theory of apparent mental causation, offer competing accounts of sense of agency. They differ in terms of the sources of information thought to be most important for producing sense of agency. For the comparator model, sensorimotor processes are key.

For the theory of apparent mental causation, the emphasis, at least in the experimental work carried out to test it, has been on information that is external to the motor system, such as environmental and social cues for a more extensive discussion, see Wegner and Sparrow, The traditional assumption has been, therefore, that these two views are mutually exclusive. However, this assumption has challenged by a number of studies. For example, using the intentional binding measure Moore and Haggard showed that both internal sensorimotor prediction and external action outcomes contributed to the sense of agency.

It was found that binding of the action to the tone outcome was present when the probability of that outcome was high, even when it did not occur. This suggests that if sensorimotor prediction is sufficiently strong binding will occur. On the other hand, it was found that when sensorimotor prediction was weak, binding would occur but only when the key press actually caused the tone outcome. This would suggest that the presence of an external tone outcome retrospectively triggered the binding effect.

This helped us move beyond the debate over whether sense of agency was based on sensorimotor information comparator model or information external to the motor system theory of apparent mental causation. Instead, according to the cue integration theory both views have merit, and in fact the sense of agency is based on various different sources of information or agency cues.

We have also suggested that the relative influence of the different sources of information may be linked to their reliability, with the more reliable source of information dominating the agentic experience. We can also see evidence of this in patients with schizophrenia. Using an agency attribution paradigm, Synofzik et al.

This reliance on external visual feedback is consistent with the cue integration theory, as it has been shown that sensorimotor prediction is unreliable in people with schizophrenia a similar finding was also obtained by Voss et al. Although, a more thorough examination of this theory is needed, it does promise to help us understand the processes underpinning sense of agency in health and disease. The previous section provides an overview of sense of agency research and theory.

However, from this overview it would not be entirely clear why any of this matters, particularly from an impact point of view. In the following section I want to address this. I will look at the possible impact of sense of agency research in the context of health and well-being, human-computer-interaction, and the broader issues of free will and responsibility.

Schizophrenia is the classic disorder of sense of agency and has been the subject of more agency research than any other disorder. Positive symptoms, on the other hand, are defined by the abnormal presence of perceptions hallucinations or beliefs delusions. Abnormal experiences of agency fall within the positive symptom category. Although these abnormal experiences can take many forms, the most common are passivity symptoms or delusions of control.

A patient with passivity symptoms will feel as though his or her actions are not under their control. You can see this in the following patient reported by Mellor , p. What they do is nothing to do with me. Research on patients with schizophrenia has confirmed that these individuals have agency processing problems.

In one relatively early study by Daprati et al. They did not directly see their own movements. Instead they saw visual feedback of the movement on screen via a video link.

These movements were either a their own actual movements, b the same movements made by an experimenter in another room, or c the movement of that experimenter performing a different movement. The participants and the experimenter were wearing gloves to prevent any visual identity clues.

Compared with controls, patients — especially those experiencing passivity symptoms — made more errors in attributing the action to its correct source when the experimenter made the same movements as them. In this situation of agentic uncertainty, patients struggled to recognize their own movements.

These action recognition problems have since been confirmed in a number of other studies. For example, Franck et al. In this experiment they made movements and again only saw video feedback of the movement. In one condition different levels of spatial distortion were introduced. In another condition different time delays were introduced. After each trial participants had to say whether the hand movements on the screen matched their own. Healthy participants tended to say no earlier in both conditions than patients who took much longer to detect these mismatches.

Again this suggests abnormal action awareness in patients.