Ipseity at the intersection of phenomenology and the natural philosophy of mind

Talk from post AAP Zahavi-fest. Kristina Musholt is first author on this.


Wegner and colleague’s model of the sense of agency probably won’t explain delusions of control but there are some surprising successes

A blog on a recent paper over at Imperfect Cognitions

many thanks to Lisa and Ema for including me.

with love, DrNPC

Do we elicit a sense of agency by inferring that our intentions cause actions?

It’s doubtful, is the short answer, but the long answer is more interesting. What I’m asking about today is the account of the sense of agency put forward by Daniel Wegner and various collaborators since the late ‘90’s (Aarts, Custers, & Wegner, 2005; Wegner, 2002; Wegner, Sparrow, & Winerman, 2004; Wegner & Wheatley, 1999). Their account can be read as either an alternative to or an addition to the comparator model account of the sense of agency which we have met in previous posts. Unlike the comparator model which hypothesises that the sense of agency is elicited by the same mechanisms that are responsible for action control, Wegner and colleagues’ account suggests that the sense of agency is elicited by an inference as to the internal causes of action. In its most basic form if I infer that one or other of mental states, usually on of my intentions, causes me to act in a certain way, then I experience a sense of agency for that action. But, that’s getting ahead of ourselves for the moment, so let’s remind ourselves what the sense of agency is.

“Imagine that you are moving swiftly down a flight of stairs. At the appropriate floor, you slow and reach for the door handle that will allow you egress from the stairwell. As you open the door, you find it moves far more rapidly than you had intended. Someone else is opening the door from the other side! Call this the simultaneous door-opening effect. This effect involves the feeling that one is the agent behind an action being suddenly replaced by the feeling that one is not the agent due to the interference of another.” (Carruthers, 2010, p. 345)

In this example, which is supposed to be fairly ordinary and mundane the sense of agency is just the feeling of being the agent who is performing/controlling/initiating (different authors prefer different descriptors here) action. That is the feeling which Wegner and colleagues’ are seeking to explain. So what, on their account, might be happening in the simultaneous door opening case? It seems that the agent of this action will, for a while, infer that their mental states, their intention to open the door say, is causing their action, and so they experience a sense of agency. But, after experiencing the effects of another on the door this inference is no longer plausible and so their sense of agency is lost or reduced.

In order to complete the model Wegner and colleagues owe us an account of how such inferences are made. Well that’s probably asking a bit much, but we would like an account that tells us something about what information the inference is based on and that makes predictions about when it is made. This is just what we get in the articles cited above. To be able to infer that one or other of one’s mental states are the cause of an action one needs to represent the action and a mental state that is a potential cause. Say opening a door and one’s intention to open the door. When do we infer that one’s intention caused the door to open? According to Wegner and colleagues we do so automatically when 3 further conditions are met. First, the intention must appear an appropriate time prior to the action, e.g. a memory of previously opening a door won’t be inferred as the cause of one currently opening the door. Call this the principle of priority. Second, the intention must be consistent with the action, i.e. it should specify that action, e.g. my desire to get a coke represents a different action than my intention to open the door so won’t be inferred as a cause of that action. Call this the principle of consistency. Third, the intention must be represented as the exclusive cause of the action. Call this the principle of exclusivity. This is what is violated in the door opening scenario, in that case the subject perceives that another agent is also opening the door and so their intention to open the door isn’t an exclusive cause of it opening. As such their sense of agency is reduced.

Wegner and colleagues’ studies looked to confirm this model by attempting to elicit a sense of agency for actions not controlled by the subject. They do so by creating circumstances where to the subject it appears that the principles of priority and consistency are met. Typically it is clear to the subject that the principle of exclusivity is not met. However, as each of the three principles are considered to contribute to the sense of agency, and the sense itself is conceived of as being formed by continuum, it is hypothesised that subjects in these studies should report a weak sense of agency rather than a full or absent sense.

If we allow the assumption that the inference that one’s intention is the cause of an action is mandatory, i.e. that the subject must make it when the principles are met, then combined with the above assumption about a continuum of senses of agency, some studies do support Wegner and colleagues’ model. For example, in the helping hands study (Wegner et al., 2004), subjects reported a weak sense of agency for other people’s actions:

“In this study, one subject (the participant) stood with their back against the second subject (the helper). The participant stood with their arms by their side, whilst the helper reached their arms forward underneath the participant’s arms. A screen obscured the helper in such a way that their arms appeared (from the front) to be those of the participant. Both subjects were given head phones. The helpers heard a series of instructions to perform a set of hand movements (e.g. make the ok sign with both hands). The participants were divided into three groups depending on what they heard. One group (the preview group) heard the instructions to the helper whilst the other groups (the control groups) heard either nothing or an instruction to perform a movement other than what the helper heard. Those in the preview group reported a greater sense of agency over the movements of the helper than either of the control groups (Wegner et al. 2004, pp. 841 & 842).” (Carruthers, 2010, p. 346).

A variety of studies, mostly coming from Wegner and his collaborators, seem to confirm that subject’s experiences of their own agency can be altered by playing with the principles of priority, consistency and exclusivity. As such we have good reason to consider this model a viable explanation of the sense of agency.

As I said above, though, I think it is doubtful that it is making such an inference that causes subjects to experience a sense of agency. The reason for this, is that such an account ties the sense of agency to the subject’s knowledge of their mental states, in particular their intentions. This, it seems, we have good reason to question as some subjects seem to experience a sense of agency even when they don’t know what their intentions are.

A study by Montgomery and Lightner (2004) provides an illustrative example. They began by showing young children (3-4 years old) a picture of a ball. They then copied the picture by holding their child’s drawing hand (with the child’s eyes closed) and moving the hand in a circle. The child then watched the experimenter alter the picture to be a picture of a clock. In comparison conditions the child either produced a copy of the picture of a ball themselves or produced the copy themselves and then watched as the experimenter changed it to a picture of a clock. After the picture of a clock or ball was produced in this way the child was asked who had produced the (final) picture. This was relatively easy for 3-4 year olds, most of whom were able to accurately state who drew the final picture. That is they were able to identify the agent of the action, suggesting they experienced a sense of agency. Despite this, in the conditions where the experimenter changed the picture of a ball to a picture of the clock, the children tended to claim that they had tried (i.e. intended) to produce a picture of a clock. In other words, although the children where good at keeping track of who the agent of the action was, they were poor at knowing their own intentions. Even though they knew it was the experimenter and not themselves who had drawn the clock, they nevertheless claimed that they had tried to produce the picture of a clock.

Lost? Fair enough. The point is in this study children knew who had done what (so they had a sense of agency) but they did not know what their intentions where. So it seems that a sense of agency is not dependent on knowing what one’s intentions are, as predicted by Wegner and colleagues’ model. So it is, as I say, doubtful that we elicit a sense of agency by inferring that our intentions cause our actions.

with love


A more complete version of this argument is published in: Carruthers, G. (2010). A problem for Wegner and colleagues’ model of the sense of agency. Phenomenology and the Cognitive Sciences, 9, 341–357.


Aarts, H., Custers, R., & Wegner, D. M. (2005). On the inference of personal authorship: enhancing experienced agency by priming effect information. Consciousness and Cognition, 14, 439–458.

Carruthers, G. (2010). A problem for Wegner and colleagues’ model of the sense of agency. Phenomenology and the Cognitive Sciences, 9, 341–357.

Montgomery, D. E., & Lightner, M. (2004). Children’s developing understanding of differences between their own intentional action and passive movement. British Journal of Developmental Psychology, 22, 417–438.

Wegner, D. M. (2002). The Illusion of Conscious Will. MIT Press.

Wegner, D. M., Sparrow, B., & Winerman, L. (2004). Vicarious Agency: Experiencing control over the movements of others. Journal of Personality and Social Psychology, 86(6), 838–848.

Wegner, D. M., & Wheatley, T. (1999). Apparent mental causation: sources of the experience of will. American Psychologist, 54(7), 480–492.

Delusions of Thought Insertion and the Sense of Agency, or Disorder’s of Self-Consciousness part 2

So here is a weird thing to say:

I look out of the window and I think that the garden looks nice and the grass looks cool, but the thoughts of Eamon Andrews come into my mind. There are no other thoughts there, only his…He treats my mind like a screen and flashes his thoughts onto it like you flash a picture (Mellor 1970 p. 17).

Now there’s lots to be baffled by here (not the least of which being: who the hell is Eamon Andrews? – he’s an old British TV presenter), and it’s just that sense of bafflement that I want to highlight. What is this person claiming? That someone else is thinking in their mind? What on Earth would that be like? This isn’t trivial. No one else ever thinks using my mind and only you ever think with yours. How can we even understand what this person is saying?

Well, when I tell you that this statement comes from a person suffering from schizophrenia you might be tempted to dismiss the statement as a semi-random sequence of words, which certainly sounds like an English sentence, but not one which in anyway reflects the experiences the subject is actually having. Maybe this isn’t a report of an experience the patient has, but rather what we hear from them is an empty sequence of words- perhaps the result of thought derailment- more closely related to what is insensitively referred to as “incomprehensible ramblings” than the kind of delusion we saw last time out.

There are a number of considerations which speak against this. The first is the sheer consistency of patient’s reports. Patients don’t just say this once; it is something they claim over extended periods of time. The second is the number of patients making claims around this theme; it just isn’t likely that so many people would randomly produce such sentences. The third is that patients make this kind of claim even when they don’t show other signs of symptoms such as thought derailment. Overall, then, it seems the burden is on us to try and understand what these patients are saying about their experiences, rather than dismiss them as incomprehensible.

Well how might we start to do so? Last time out with delusions of alien control the problem didn’t arise; I just highlighted a common experience and suggested that the delusion arose from a bizarre form of that experience. But what kind of experience, familiar experience, could be altered to give rise to this kind of belief, these “delusions of thought insertion”.

One popular approach has actually been to relate these delusions to delusions of alien control. As delusions of alien control involve a deficit in the sense of agency over bodily actions, perhaps delusions of thought insertion involve a deficit in the sense of agency over mental actions. Just as delusions of alien control arise because the patient doesn’t feel like they control their own actions, perhaps delusions of thought insertion arise because the patient doesn’t feel like they control their thoughts?

However an explanation such as this may work it cannot be the same as what we saw last time out for alien control. Recall the mechanism which we thought was important for this delusion involved a representation of actual sensory feedback, however, in the case of thoughts there isn’t actual sensory feedback to be had. So we’re looking for a different, but perhaps in some way analogous mechanism.

All well and good, but there is a bigger problem for trying to extend this kind of account for delusions of alien control to explain thought insertion. I left a part of the explanation last time out hidden. In order to come up with something as weird as feeling that someone else is controlling your actions it is supposed that lacking a feeling of agency is somehow anomalous. The patient is thought to come up with the delusion to explain their experience because the experience is weird. It’s unexpected. This stage is needed to get us from “feeling that I didn’t do it” to “feeling person x did it”.

Is this move plausible for accounts of delusions of thought insertion? Well there are a few worries to be had. First, if you’re anything like me, you might not find it so weird to think that such and such a thought you have wasn’t under your control, maybe it just popped into your head, out of the blue as it were. If that sort of consideration is right then just how noticeable would thoughts for which the patient lacks a sense of agency be noticeable? Maybe thinking thoughts one doesn’t seem to control isn’t as anomalous as performing actions one doesn’t seem to control. Second, we might worry about the nature of the delusional belief the patient ends up with. It certainly makes sense to attribute your actions to someone else- this really happens, as in the simultaneous door opening effect we saw last time. But it is in fact the case that no one else can put their thoughts in your mind (i.e. put them in directly, without an act of communication- the patient quotes above certainly doesn’t seem to be saying that Eamon Andrews is communicating his thoughts to her, it’s something more direct than that.)

Well I must admit that the second of these problems still baffles me, but I think some progress has been made on the first. The more radical proposal I have made around this is to suppose that patients suffering from these delusions don’t lack a sense of agency when they ought to have it; rather they lack a sense of agency over their thoughts and this accurately reflects the fact that they have lost, a rather specific kind of, control over some of their thoughts. In particular it seems that what they lose, sometimes for some thoughts, is the capacity to try not to have the thought.

Trying not to have a thought, or trying not to think about something, is in one way very hard to do… “God I want the Thomas the Tank Engine theme song out of my head”, so I don’t want to be read as saying that normally we are good at directing our thoughts away from certain topics. Rather what is important to note is that usually what we can do is pick out thoughts we don’t want to have and then trying to supress it. This is the capacity which seems to be failing in delusions of thought insertion.

The evidence for this comes from studies where experimenters try and create particular thoughts in their subjects and then direct them not to act on the thought. We might do this be giving subjects a list of sentence stems, like “The dough was put in the hot ____” and then directing subject’s to finish the sentence in a way that doesn’t make sense. Here the experimenter tries to make the subject think the word “oven” but then the subject is required to inhibit this word, and come up with another. It turns out that patients suffering from delusions of thought insertion (and some related symptoms, just to make it complex!) are worse at these sorts of tasks than people who don’t suffer from these symptoms (Waters et al. 2003).

I still don’t know why patient’s go from this to attributing the thought to someone else, but I think progress is being made here.

You can read my full account of this in my paper:

“A metacognitive model of the sense of agency over thoughts.” Cognitive Neuropsychiatry 17(4): 291-314.

And that’s something else that I do.

With love



Mellor, C. S. (1970). “First Rank Symptoms of Schizophrenia.” The British Journal of Psychiatry 117(536): 15-23.

Waters, F. A. V., J. C. Badcock, M. T. Maybery and P. T. Michie (2003). “Inhibition in schizophrenia: Association with auditory hallucinations.” Schizophrenia Research 62(3): 275-280.