by Hubert J.M. Hermans

As argued earlier (Hermans, Kempen, & Van Loon, 1992), the proposed conception is a step beyond individualism and rationalism and differs essentially from the Cartesian cogito. The Cartesian conception of the self is traditionally phrased in terms of the expression ‘I think’. This expression assumes that there is one centralized I responsible for the steps in reasoning or thinking. Moreover, the Cartesian ‘I think’ is based on a disembodied mental process assumed to be essentially different from the body and other material extended in space. (For a comparison of the separated Cartesian self and the dialogical self, see also Fogel, 1993.)

In contrast to the individualistic self, the dialogical self is based on the assumption that there are many I-positions that can be occupied by the same person.{/quotes} The I in the one position, moreover, can agree, disagree, understand, misunderstand, oppose, contradict, question, challenge and even ridicule the I in another position. In contrast to the rationalistic self, the dialogical self is always tied to a particular position in space and time. As Merleau-Ponty (1945/1962) would have it, there is no ‘God’s eye view’. As an embodied being, the person is not able to freely ‘fly above’ his or her position in space and time, but he or she is always located at some point in space and time. Even the most advanced arithmetical problem involves a system of numbers originally based on the counting of 10 fingers, which is in turn indispensable for the child to understand the activity of counting at all (we also measure in ‘feet’). The dialogical self is ‘social’, not in the sense that a self-contained individual enters into social interactions with other outside people, but in the sense that other people occupy positions in a multivoiced self. The self is not only ‘here’ but also ‘there’, and, owing to the power of imagination, the person can act as if he or she were the other and the other were him- or herself. This is not the same as ‘taking the role of the other’, as Mead (1934) meant by this expression that the self is taking the actual perspective of the other. Rather, I’m able to construe another person or being as a position that I can occupy and as a position that creates an alternative perspective on the world and myself. The constructed perspective may or may not be congruent with the perspective that is defined as the perspective of the ‘actual’ other (which can be checked by entering into conversation with the other). It should be emphasized, however, that the other may be partly the product of my imagination, closely intertwined with the ‘actual’ other, and can be even completely imaginary.

Mind as Society

In computer brain sciences there are developments that use ‘society’ as a model to comprehend the complexities of the brain. Such models may contribute to an understanding of the multivoiced and dialogical nature of the self. Computer scientist Minsky (1985), for example, considers the mind as a hierarchically organized network of interconnected parts that together function as a ‘society’. In his model the mind consists of a host of smaller minds, called agents. Many of these agents don’t comprehend one another because most agents are not able to communicate with each other at all. In this respect, the mind resembles a human society in which many agents have their own action programs and simply do their job without knowing all the other agents who are part of the community. However, at the higher levels of organization, agents may be involved in direct communication.

Minsky developed a computer program for block building in which he describes the conflict between two agents at the same level of organization:a Builder and a Wrecker, who is only interested in breaking down what Builder has achieved. At this level, agents may agree or disagree with one another:

Only larger agencies could be resourceful enough to do such things. Inside an actual child, the agencies responsible for Building and Wrecking might indeed become versatile enough to negotiate by offering support for one another’s goals. ‘Please, Wrecker, wait a moment more till Builder adds just one more block: it’s worth it for a louder crash!’ (Minsky, 1985, p. 33) In this model, conflicts between agents tend to migrate upward to higher levels in the society of mind. If the conflicts between the disagreeing agents are not solved, the higher-level agent under which they are subordinated is weakened. If Builder and Wrecker, in the above example, cannot solve their conflict, they reduce the strength of their mutual superior (e.g. Play), with the result that this superior will then be surpassed by competing agencies on the same level (e.g. Sleep or Eat). If Builder and Wrecker are not able to settle their disagreements, the child stops playing and wants to sleep.

Another computer scientist, Hofstadter (1986), also uses the notions of voice and dialogue in his attempts to comprehend the workings of the mind. In his model the mind, with its billions of neurons, resembles a community made up of smaller communities, each in turn made up of smaller ones. The highest-level communities are called ‘subselves’ or ‘inner voices’. In Hofstadter’s view, each inner voice is composed of millions of smaller parts, each of which is an active part of a community.

Under specific circumstances, these smaller parts are all ‘pointing in the same direction’, and at that moment an inner voice crystallizes and undergoes a ‘phase transition’. The voice proclaims itself an active member of the community of subselves. If it is strong enough, it exerts pressure in order to be recognized and to get in touch with other voices. A hypothetical dialogue may take place: ‘. . . a dialogue between two persons both of whom are inside me, both of whom are genuinely myself, but who are at odds, in some sense, with each other’ (Hofstadter, 1986, p. 782). If the disagreeing voices are able to solve their conflict, or when one of the voices becomes stronger than the other, the person is able to take a ‘decision’. Hofstadter (1986) and Minsky (1985) share the idea that the brain is a community of agents or voices that, at its higher levels, may entertain mutual dialogical relationships, with one voice being more dominant or active than the other voice. The multiplicity of voices, as postulated by computer scientists, is well in agreement with the original formulations by James (1890) on the ‘rivalry and conflict of the different selves’ (p. 309) and with the metaphor of the polyphonic novel as proposed by Bakhtin (1929/1973). Moreover, the models of the two computer scientists share the idea that the decisions that are reached do not result from a centralized and unified ego or I that, as an authoritarian leader, keeps its followers under control. Decisions are taken ‘from the bottom up’ rather than ‘from the top down’.

Despite the apparent commonalities between the two computer scientists, it should be kept in mind that they use the notion of society, voice and dialogue more as metaphors for comprehending the workings of the brain than as means for understanding the social processes in actual communities of people. Dialogicality in the Bakhtinian sense is not restricted to ‘inner voices’ within the individual mind but also includes ‘external voices’. Both dialogical forms are needed for a model of self and culture. As featured on