Actions vs. Accusations {#sec4.1.5} An action-based view of the world has been proposed as a cognitive mechanism that explains how one understands the intentional mental states of others (Searle [@cit0069]; see also Goldman [@cit0022], for a general discussion). According to this view, people represent the world in terms of their intentions, goals, and beliefs. When someone tries to do something, we interpret his behavior as part of his attempt to achieve his goal, and in particular we interpret his behavior as intended (rather than merely foreseen). For instance, when we observe someone trying to pass through a narrow door, we understand his behavior as attempting to go through the door rather than simply being aware of the obstacle. This view has received an interesting defense in recent work on the *doings* of consciousness (Pacherie [@cit0056]), where it is argued that it is useful to think about the mental states of others in terms of their *doings*, or the actions they attempt to perform (such as going to the library or stealing the cup). One reason for the usefulness of the doings view is that it provides a simple explanation for how we understand intentional mental states: For instance, if someone asks me whether Mary is at home, I will readily understand that I need to consider the doings of Mary rather than her beliefs, even though the latter are also relevant for my answer. A similar view holds for speech acts. When someone says "He is here," we are ready to interpret the sentence as an act of saying and not as a mere announcement. The same holds for speech acts like telling. This ability to interpret someone\'s speech as part of a genuine attempt to perform an act presupposes a common ground about the actions the speaker attempts to perform (such as going to school or playing basketball) and the actions the hearer attempts to perform (such as going to class or playing football). This common ground can be understood as part of the larger social context, which is required for interpretation of intentional mental states and speech acts. Thus, a common ground is arguably not just a constraint on interpretation, but it also explains the possibility of the interpretations. This suggests that an adequate understanding of intentional states requires that the interpreter and the person who attempts to perform the action share a common ground. This is analogous to cases of disagreement where both participants have a shared common ground. For instance, when Alfa claims that she believes that the sky is blue, and Beta disagrees, Beta does not need to understand that her opponent holds this belief; she only needs to know that she herself believes that the sky is not blue, and also that Alfa\'s belief is based on Alfa\'s perceptual evidence (rather than on a mere guess, say). The shared common ground then provides the basis for their disagreement. ### 5.1.6. Reflective self-ascription and disagreement {#s5.1.6} In light of our previous discussion, I will propose the following picture of the cognitive mechanisms that support the understanding of the intentional states of others. On this picture, the problem of other minds is understood as a problem of reflective self-ascription. There is a tendency to view the cognitive mechanisms in terms of two distinct levels: (1) the initial level, where the behavior of an agent is described in terms of his perceptions and intentions, and (2) the reflective level, where we describe our perception and intentions in terms of beliefs and desires (see [Figure 2](#f0002){ref-type="fig"}). The problem of other minds, I propose, arises when these two levels are collapsed (see [Figure 3](#f0003){ref-type="fig"}). When we reflectively ascribe intentional states to others (i.e., when we reflect on our own behavior and come to believe that it is caused by our beliefs, desires, and intentions), we tend to view our own behavior as caused by our own beliefs, desires, and intentions, which makes the interpretation of other agents\' mental states seem like our own personal beliefs. However, if our beliefs about our own mental states are different from what we actually think about the world (i.e., our beliefs are misguided), then this interpretation seems flawed. This problem is then easily generalized to our understanding of other agents and their mental states. We could describe our own beliefs about the world (e.g., as being caused by our perceptual experiences and beliefs), and the perception-action mechanism is then designed in a way that fits with these beliefs. When our beliefs about the world deviate from how the world really is (e.g., by not taking into account information about the reliability of our perceptual faculties or by overvaluing the reliability of perceptual faculties when they are actually flawed), then the agent will be mistaken about the world, which will make it difficult to understand others. When we attempt to interpret the intentional states of others, we try to do so on the basis of the beliefs and desires we have about our own intentional states. When our beliefs about our own mental states deviate from our behavior, however, then this interpretation can seem like misguided or even ridiculous overinterpretation of other agents\' behavior. The same mechanism that underlies the problem of understanding our own behavior can then explain the difficulty in understanding others\' behavior. It follows that if there is a common ground for the interpretation of our own behavior, there is also a common ground for the interpretation of other agents\' behavior. This common ground is part of the larger social context, which also includes a shared understanding of the possibilities (of perception, memory, action, etc.) that is relevant for interpreting others\' behavior. In order to understand others, we must identify a person as someone whose behavior is possible for him. This shared understanding of the possibilities requires a shared understanding of the physical laws of the world, which can be understood as the external context, and we must rely on this context in order to understand the behavior of others. This context serves as a constraint on the understanding of other people\'s behavior. If one person\'s behavior is prevented by the laws of physics, then it is more difficult for him to understand the behavior of another person. For instance, if there were a magical law in place that prevents Alice from passing through door A, then her belief about her own behavior would not be considered a misguided belief. But since this is not the case, she does not view the behavior of Peter as misguided. These considerations support the claim that when we try to understand the intentional states of others, we do so on the basis of a common ground that includes shared beliefs and desires and also common understanding of the possibilities. In other words, there is a shared understanding of the external context, which includes a shared understanding of the possibilities of perception, memory, action, etc. According to this proposal, misguided interpretations can result when our beliefs about the world deviate from the external context. These deviations can take place either because our beliefs are false (e.g., because our perceptual experiences are flawed) or because our beliefs are correct but the world does not actually have the properties we take it to have (e.g., because our perceptual experiences are veridical but our memory fails us). In cases where the misguided beliefs result from false perceptions or flawed memories, the belief of the agent can be considered a misleading overinterpretation of a physical state (such as the redness of an object). However, I argue that the misguided beliefs might also result from other cases where the person has a correct interpretation of the world, as long as there is a mismatch between the interpretation of this state and the actual properties of the world (such as false beliefs and desires in a world where perceptual veridicality is independent of the truth of our beliefs about the world). I want to now argue that misguided interpretations can result when our beliefs about the world deviate from the external context and when they are correct. Misguided interpretations result from other cases when we mistake the state of the world for the effect of our beliefs and desires, as when we misinterpret the redness of an object as caused by its being red. On this view, a person can make a correct interpretation of the world if he possesses beliefs that are consistent with what we take the world to be (such as a color realism). Misguided interpretations result, for instance, when we have a false belief about the properties of an object that is independent of our perceptual experiences. This can happen when we overinterpret a physical state (e.g., because we overinterpret the reliability of our sensory systems or we overinterpret the causal connection between perceptual states). One can then think that because we make this error we will not be able to understand the behavior of another person, but the situation can be different: There is a common ground for the interpretation of the behavior of others that includes shared beliefs and desires (which can explain the difficulty in understanding others), and also a shared context of possibilities (which helps in understanding others).[^2^](#fn0002){ref-type="fn"} I suggest that when we understand other people\'s behavior in the common ground, we need to think about both what it is like to be a person, that is, about the properties of that person, as well as what the environment would look like for such a person. I have already argued that in the situation of an agent, the physical laws (the context) of the world constrain what properties are real and relevant for the agent. As a result, it follows that the physical context of the world has a crucial role in our understanding of other people\'s mental states. For instance, if our mental states are caused by physical states,