The program of the 10th TCCP was a presentation by Petter Johansson, a JSPS postdoctoral fellow (http://www.lucs.lu.se/petter.johansson). The title of his presentation was “Choice Blindness: The Incongruence of Intention, Action and Introspection”. This presentation was based on his PhD thesis, which concerns a new experimental method constructed by him and his colleagues: choice blindness. Through their experiments, they test the stability of people’s self-knowledge when making choices concerning their own preference in visual images (faces), tastes, smells, commodities and so on. The results show that, surprisingly, our self-knowledge is in fact far more limited and fragile than we usually suppose. Johansson holds that these results are better understood in the framework of interpretationism about mental states.

　　The basic idea of Johansson is that because of evolutionary restrictions, humans have only limited cognitive abilities. As for self-knowledge, we might obtain knowledge about our beliefs and desires through inference rather than introspection. Self-knowledge would be fallible just as knowledge about others’ mental states is. Thus, even when there is inconsistency in what we think we originally intended to do and the outcome of action, we might miss the inconsistency. For example, when we had intended to get A but actually got B, we might think that what we had intended to get was B. To see if this is the case, Johansson and his colleagues set situations where an experienced outcome deviates from what was intended.

　　Here is one example of the series of experiments they conducted: a subject is shown a pair of pictures of female faces and asked which one to prefer. After he makes a choice, the experimenter turns over pictures and hands one of them to the subject. It appears to the subject that he has been handed the picture he prefers. But, in manipulated trials, the picture is the other one, namely, the one he does not prefer. The experimenter uses a magician’s trick. Astonishingly, subjects often miss the switch. When they are asked reasons for preference, they give long and elaborate answers―even though the face is in fact the initially non-preferred one. The contents of reports for non-manipulated and manipulated trials are almost the same: both similarly provide emotional and specific reasons for choices. Besides, subjects in manipulated trials show confidence in their choices, just as subjects in non-manipulated trials do. When manipulation is repeated, subjects come to like the alternative they initially non-preferred. These results show the fragility of our recognition about our own reasoning and preference.

　　Johansson and his colleagues had experiments of choice blindness not only in choices of visual images. They also designed experiments to test whether choice blindness in choices of tastes and smells is possible. In the experiments, subjects are asked to choose a jam or tea from two candidates. Then, once again subjects are asked to taste the jam or smell the tea they “preferred” and to make some comments―yet in manipulated trials, subjects actually taste or smell the non-preferred one. The result is similar in the case of visual image: no difference in confidence between the manipulated and non-manipulated trials, and tendency to come to like the non-preferred one.

　　Similar research was conducted about the consumer decision making. When we make a decision about which item to buy, we take into consideration several attributes, and compare alternatives in terms of the attributes. In the experiment, such attributes were secretly changed after the choice. For example, in the choice of an apartment, an alternative which was shown as a lower-rent one is changed into a higher-rent one after the choice. There seemed to be effects like choice blindness; being displayed a non-preferred alternative as a preferred one, some subjects inferred or fabulated about the reasons for their own choice based on changed attributes. Johansson et al. are now planning to have the choice-blindness experiment about choice in gambling scenarios and moral dilemmas.

　　What kind of implication can these results have on philosophy? The phenomenon of choice blindness has much to do with the nature of folk psychology. There are two opposing views about the ontological status of mental states: interpretationism and realism. Interpretationism, represented by D. Dennett, is a view that mental states like beliefs and desires do not necessarily map onto some internal states. Mental states are patterns which are ascribed to people to interpret their actions as rational ones. On the other hand, a realist such as J. Fodor claims that there are some internal states which are identified as mental states. In the choice blindness experiment, participants ascribe to themselves beliefs and desires that they could not have when they made their choices. The result makes more sense within the framework of interpretationism, provided that knowing one’s own mental states through inference and self-ascription is in fact commonplace.

　　The phenomenon of choice blindness shows that our understanding of ourselves and the world owes itself much to our expectations about, and feedback from, our experiences. There is a famous experiment which shows the power of feedback. In the experiment, the researcher hides a subject’s hand in a box and places the rubber hand over the box. Then the researcher pushes the rubber hand and the real hand, in the same time and on the same place. Watching the rubber hand, the subject comes to feel as if the rubber hand is her real hand. This is “rubber-hand illusion”. The subject has an experience in which what happens to the rubber hand coincides with what she feels in her hand, and her recognition of her hand is influenced by the experience.

　　We are implicitly confident that other people, the world, and our senses hardly deceive us. Such trusting nature of us might be an outcome of evolutionary process. Most of the time, this raises no problem; we do not make significant mistakes about ourselves and the world. Yet, our success might be because of the fact that events in daily lives do not deviate so much from what we rationally think to occur. Examples like choice blindness or rubber-hand illusion suggest that.

　　As we have seen above, choice blindness is an intriguing phenomenon which might have some implications about the nature of our folk psychology. The research by Johansson et al. is outstanding in their unique experimental design which can reveal an unknown aspect of our mind. Their research shows that well-designed experiments can have interesting suggestions for philosophical problems.

Haruka Tsutsui (The University of Tokyo)