The Illusion of Conscious Will (29 page)

We don’t normally think of memory for the present as a big problem. The present is here and now, and if we forget what it is, we can just look around us and, yes, gosh, there it is. However, the loss of synchronous act memory can have some interesting and even tragic repercussions. When does this happen? The lack of awareness of an intention that is observed in some of the classic automatisms appears to signal a failure of current act memory. A person may be engaged in an action and have forgotten what was being done. As a result, the action seems to be occurring with-out awareness. This possibility was specifically noted in one study of automatic writing: “As the experiment progressed we came to realize more and more the extent to which a reported lapse of awareness might be a lapse of memory instead” (Downey and Anderson 1915, 193). The determination that a particular moment of writing is automatic depends on the person’s current memory for what is being done. Not remembering what was intended as it is being done counts quite clearly as unawareness of action. This inability to remember what one is doing during the action can translate into a failure to recall the nature of the action afterwards as well. So the absence of an intention in the current act register pretty much dooms the person to postaction interpretation of intention.

Without a memory of what the self is currently doing, a person is left with some fairly odd options for postaction interpretation. As we saw in the case of posthypnotic suggestion, the person could deny that any action took place or could simply declare the source of the action unknown and the action inexplicable. Another clear alternative would be inventing intentions. These options hold in common the assumption that the person continues to view the body’s behaviors as emanating from the self and that explanations therefore will be made in terms of that actor.

Ongoing behavior for which there is no entry in the current act register could also be accounted for in a very different way—as the doing of an agent other than the self. This possibility seems highly unlikely for normal people under normal conditions, but it does occur in certain cases of schizophrenia. Christopher Frith (1987; 1992; 1994) has proposed that there is an impairment of memory for current action in some people with schizophrenia. He has suggested that such impairment is responsible for the phenomena of alien control that they experience when they hear voices or experience their thoughts or actions as those of someone else. We have already discussed some of the evidence regarding auditory hallucinations, but it is instructive to examine, too, the research on synchronous intent memory in schizophrenia.

In one study, Frith and Done (1989) arranged for schizophrenic patients suffering from alien control, and other patients without this experience, to play a video game that measured memory for current action. The player used a joystick to shoot at birds coming from the left or the right, and the video bullet took a long time to reach the target (2.8 seconds). At any time, the player could correct an error by changing aim to shoot the bird coming from the other direction. Although all players did equally well under these conditions, the alien control patients experienced a noteworthy failure when a small variation was made in the game: Barriers were placed to hide the trajectory of their bullet for the first 2 seconds after the shot, leaving them with only a brief interval in which to see where they had shot and to correct any wrong moves. Under these conditions, the patients with alien control symptoms performed far less well than did the comparison group. Mlakar, Jensterle, and Frith (1994) observed much the same thing among alien control patients who were asked to make drawings without immediate visual feedback on their performance. They had great difficulty keeping track of what they were doing.

This characteristic of schizophrenia suggests that people with this problem might be deeply distractible because they constantly fail to remember what they’re doing. In fact, this is a common observation. Back in the heyday of research on the Zeigarnik effect, a set of studies by Rickers-Ovsiankina (1937) focused specifically on the degree to which people with schizophrenia showed the magnified memory for incomplete tasks that is common in normal individuals. She observed that people with schizophrenia “tend towards undirected playful activities, which do not lead to any definite outcome. Striking on a xylophone without attempting to produce a melody is an example” (179). And, indeed, she found a clear tendency in normal individuals to resume tasks that were interrupted, whereas this did not occur nearly as often for people with schizophrenia. The lack of a continued memory of current action, as suggested by Frith’s hypothesis, could account for such findings. This conclusion depends on the idea that the same psychological cause may underlie different symptoms simply because they co-occur in this disorder, of course, and this interpretation is tempting but not beyond doubt. The analysis of a wider range of disorders in terms of this model, as undertaken by Frith, Blakemore, and Wolpert (1999), holds promise for understanding how breakdowns in current act memory might influence a variety of human abilities.

Taken together, these observations suggest that the current act register is a piece of mental architecture that allows us to produce action with the illusion of conscious will. Without a well-functioning memory of what one is doing now, any actions which ensue that are hard to predict from past thoughts might seem instead to come from elsewhere, perhaps even outside the self. Actions that are not known as they are being produced— that is, that are not coded in the current register—might be entirely disowned and experienced as alien or perhaps simply disregarded. Knowing that our actions are our own seems so entirely natural and automatic that it is startling to realize that such ownership only occurs as a result of a finely tuned system that adheres the knowledge of action to the action it-self. Breaks in this system can range from the profound in schizophrenia to the minor lapses that occur in absentminded behavior among normal individuals (Reason 1984). Just as people with schizophrenia may hear voices, the rest of us may wonder who rearranged our sock drawer or moved the teapot when we don’t recall doing these things ourselves.

The final type of intention memory is retrospective memory for intention—knowing what we did after the action is done. Presumably, of course, we all expect that our retrospective memory of intention will correspond nicely with our prospective and synchronous memories. We presume that the same intention will be remembered after action that we anticipated beforehand and that we embraced as the action occurred. As it happens, however, a number of observations indicate that what we think afterwards about what we were doing can diverge quite dramatically from what we thought we were doing before or during the action. The matter of retrospective memory for intention is the focus of the rest of this chapter.

Our discussion of unconscious thought can be summed up simply: We don’t always know what we are doing. Whether our thoughts of action are unconscious because of shifting action identification, because of action instigation through thoughts that are only accessible and not conscious, or because of lapses in memory for intention, these cases provide serious challenges to our conception of ourselves as pretenders to ideal agency. We can’t be ideal agents if we didn’t consciously intend
each and every action we come to understand we have performed
. This means that we must respond to the challenge of unconscious action creatively—by finding, inventing, or constructing notions of what our intentions must have been whenever we find ourselves falling short as ideal agents.

The Confabulation of Intentions

Aesop knew that people, or foxes who act like them, may change what they say about their intentions (
fig. 5.3
). The fox whose dearly desired snack turned into “sour grapes” reminds us that initial intentions (“I’d love some grapes”) can be readily overwritten by replacements (“These grapes are no doubt sour, and I didn’t want them in the first place”). As it turns out, such intention invention stands at the center of several psychological theories. Studies of cognitive dissonance suggest that intention confabulation occurs from conflict between old and new intentions; self-perception theory alerts us to the confabulation of intention that may flow from the apparent absence of old intentions. Finally, the most unusual circumstances of invented intention occur when brain damage leaves the left side of the brain interpreting what the right side is doing.
¹⁰
Each of these perspectives illustrates the confabulation of intention through a different conceptual lens, but together they have provided a consistent array of evidence to show that people often revise what they think they intended to do after their action is complete. This confabulation protects the ideal of conscious agency from the fact that actions are caused by unconscious processes.

Figure 5.3

The fox in Aesop’s fable wanted the grapes but couldn’t reach them, so she decided that they were sour and not worth having. This revision of her initial intention suggests that it was easily clouded in her mind. Marcus Gheerhaerts’ 1674 engraving for L’Estrange’s translation of the fables.

Cognitive Dissonance

With the theory of cognitive dissonance, Leon Festinger (1957) proposed that people will revise their attitudes to justify their action. In a nutshell, the theory says this happens because people are motivated to avoid having their thoughts in a dissonant relationship, and they feel uncomfortable when dissonance occurs. The strongest dissonance arises when a person does something that is inconsistent with a preexisting attitude or desire. So, for instance, buying a house that has an aging, decrepit furnace would create cognitive dissonance. The person goes ahead and buys the house knowing full well that there will be a big furnace problem. The theory proposes that when this happens, the person will change his or her attitude to make it consonant with the behavior. The attitude (in this case, of neutrality toward the house) is relatively changeable because it exists only in the person’s private thoughts, whereas the behavior of buying the house (with its broken heating plant) is typically public and much more difficult to undo. The person comes to like the house that was purchased, strangely enough, for the very reason that the furnace is faulty. The furnace remains unloved, but the house is seen as desirable because it was purchased in spite of that furnace.

10.
In
Altered Egos,
the neurologist Todd Feinberg (2001) has woven a variety of case studies of brain-damaged patients into a lucid account of the maintenance of a sense of self through confabulation. This book is a key piece in the puzzle of how the agent self is constructed.

In everyday parlance, the theory merely says that people justify the things they do. This much had been observed in research that predated Festinger’s theory. For instance, in studies of laboratory role playing by Janis and King (1954), participants who agreed to make a speech for which they played the role of someone who believed strongly in an issue were found afterwards to have come to believe in the issue themselves, especially in comparison to those who had simply to listen to a speech on the topic by someone else. The role players’ attitudes moved in the direction of the speech and so became consistent with their behavior. Imagine: You might get someone to pretend to be an opera lover (no one actually starts out life that way) and the person would come to like opera more as a result. Festinger’s theory took such observations a step further by noting that such changes are particularly likely to occur
the more the person feels in advance that there were good reasons not to perform the behavior,
and
the more the person feels he or she chose the behavior and was responsible for it.
¹¹

These principles are illustrated nicely in an experiment on a kind of role playing by Linder, Cooper, and Jones (1967). Participants for this study were students who were asked to write essays praising a ban on speakers at their college—a silly position that all disagreed with in advance. For some of them, the pay for writing the essay was announced as 50 cents, whereas for others it was $2.50 (at that time, although the latter sum was actually a fair wage for writing an essay, 50 cents was still cheap). A further variation in the experiment manipulated the perception of choice: Some participants were led to believe that they had considerable personal choice on whether to write an essay—the experimenter explained at the outset that after he had described the study they could decide for themselves whether or not to write the essay. For others, this emphasis on choice was not made.

11.
Festinger (1957) didn’t say precisely these things about dissonance, but this gloss on his theory follows the understanding of these processes that follows from current research. For reviews of this literature, see Cooper and Fazio (1984), Eagly and Chaiken (1993), and Wicklund and Brehm (1976).