by Stephen A. Butterfill
118(3), pp. 413-5
--- links: external [doi: 10.1215/00318108-2009-013]
The first-person is "expressive of self-consciousness" (64). Someone who uses the first-person, for example in saying or thinking "Where am I?", not only succeeds in referring to herself but also knows that she refers to herself. In Self-Knowing Agents, O'Brien's primary aim is to explain how it is that using the first-person guarantees this knowledge-that is, to explain "how it is that 'I' expresses self-consciousness" (57).
The explanation hinges on agent's awareness, the awareness an agent has of an action by virtue of controlling it. Control consists in this: the agent evaluates possibilities for action which she is aware of as possibilities; then she selects, endorses or accepts one (116-7, 183). The possibilities in question are basic actions, that is, actions the subject can carry out without having to do anything else (163-5). The process of evaluation and selection immediately determines how the agent will act. For this reason, an agent who knows that some action resulted from her evaluation and selection is in a position to know which basic action resulted (165). Further, nobody else's actions can be immediately determined by her process of evaluation and selection, only the agent's (184). For this reason, if an agent knows that some action occurred as an immediate consequence of her process of evaluation and selection, then she is able to know that the action is her action (119). The general idea, then, is that agent's awareness does not involve first-person reference but does enable agents to know with respect to their own actions both who is acting and what the action is (123).
The explanation of how 'I' expresses self-consciousness is now straightforward. O'Brien holds that there are mental actions and that judging, wondering and supposing are all actions (89). Using the first-person in thought or talk is a basic action (or perhaps a component of one) and so can be done with agent's awareness. Where the first-person is used with agent's awareness, the agent can know that her action is one of using the first-person and that she is the agent of this action. Now users of the first-person also know that the first-person is governed by the rule that it refers to the subject who produced it (77). Knowledge of this rule in combination with agent's awareness enables subjects who use the first-person to judge that they have referred to themselves.
In short, given some general knowledge about the term 'I', what enables us to know who we are referring to when using 'I' is the act of using 'I' itself.
The book also explores a set of related claims about self-knowledge. These centre on the question, How do we know which bodily action we are performing and, in the case of mental action, which thought we are thinking? As sketched above, agent's awareness provides an answer. When an action results from evaluating and selecting possibilities that we are aware of as possibilities, the agent can know what she is thinking or doing because she selected it. So it is the fact of acting with agent's awareness and not any perception or representation of our actions that enables us to know what we are thinking and doing.
The positive claims of Self-Knowing Agents are supported by critical discussion of alternative positions. One type of alternative is to reject the datum O'Brien sets out to explain, that the first-person is expressive of self-consciousness. This might be rejected either on the grounds that 'I' does not refer (Anscombe) or the grounds that using the first-person need not involve knowing that one refers to oneself (Mellor). Both views are countered in detail (Chapter 2 and 59-65). Perceptual accounts are another type of alternative to O'Brien's position. In Chapter 3, O'Brien discusses the view that self-conscious self-reference might be grounded in bodily awareness and Gareth Evans' view that self-conscious self-reference depends on being able to use perceptual information to track one's own location. Both views imply that perception, or at least memory based on perception, is necessary for self-conscious self-reference (47-8). O'Brien objects that we could use the first-person self-consciously while suffering complete memory loss and full sensory deprivation (4-5, 34, 46).
Is it true, as O'Brien and others claim, that agents would know who their own uses of 'I' referred to even if deprived of memory and perception? Although this claim carries much weight in Self-Knowing Agents, it is not defended. Unless this claim is obviously true, there is a gap in the main argument against perceptual accounts.
O'Brien interprets others' views in the most charitable way possible and faces objections to her own squarely. She also identifies a counterexample to her own position. Recall that agent's awareness was characterised in terms of control over action. This control takes the form of evaluating possibilities for action and selecting one. When the action is judging, O'Brien illustrates the general idea by suggesting that we "accept or endorse a given thought in the light of our awareness of the possible judgements we could make, and the reasons in favour of one over another" (116). But consider being suddenly struck by the thought, "I am supposed to be outside the school gates right now" (90). Thinking this thought seems to involve self-reference as expressive of self-consciousness as any other use of the first-person. But being struck by a thought involves no process of evaluating possible thoughts and so does not involve agent's awareness. On the face of it, then, O'Brien's account cannot explain self-conscious uses of the first-person in all cases.
The problem also affects O'Brien's account of self-knowledge. Many of our actions, including our judgements, are triggered by things others say or do, and by changes in our situation. The rude remark provokes a swift retort, the infant's smile induces a hug. Agents don't generally evaluate and select possibilities in such cases, but they do exercise control to the extent that they can inhibit their actions. Control by inhibition cannot explain self-knowledge; indeed, inhibition sometimes requires awareness of which action one is inhibiting. So agent's awareness construed as evaluation and selection cannot explain an agent's knowledge of which action she is performing in these cases.
O'Brien's main response to these counterexamples is to suggest that they are derivative in the sense that self-knowledge in cases lacking agent's awareness depends on the existence of other cases in which there is agent's awareness (90-2). This suggestion is hard to evaluate without an account of how the dependence works. When an infant's smile induces hugging (say), how does the agent's knowledge of who is acting now depend on her knowledge of who acted on other occasions? O'Brien's core idea is that the fact of acting enables an agent to know who is acting. This seems equally plausible in cases where action is not the outcome of a process evaluation and selection. Perhaps, then, the substantive account of agent's awareness needs modification. As it stands, O'Brien's account makes no appeal to connections among a subject's thoughts and actions. The account works as well for an agent who acts just once as it does for an agent who enjoys a long and eventful life. If the lives of self-consciousness agents necessarily form more than a series of isolated actions, one way to modify the account of agent's awareness would be to appeal to how an agent's judgements fit with her other thoughts and how her bodily actions affect her plans.
Self-Knowing Agents is a deep and ambitious book which develops and defends a new thesis about the role of agency in self-reference and self-knowledge. Readers will be grateful that O'Brien sets the scene for her account with sympathetic and rigorous discussion of competing and connected positions. And there is a richness to the book which this review entirely fails capture, for O'Brien deftly weaves the main arguments into larger-scale views about the nature of action, bodily awareness and agency.