Anticipation by Analogy

Abstract

Why do we expect to find intelligent creatures on other planets in the Universe? Is there a general low stating it? Is there a theory that predicts it? Do we have many examples in order to generalize from them? No. Our anticipation to find intelligent beings is based on analogy with the only example we know - the planet Earth. Moreover, if we analyze the description of these potential creatures in the science fiction literature and movies, we will find out that all out imagination about these creatures is based on analogies with the human race or other animals on the Earth. Similarly, when the Europeans arrived for the first time at other continents, they expected the population there to have a culture analogous to the European and when this turned out not to be true, they announced ”the other” to be ”less developed”, exactly as the Romans declared all non Romans to be Barbarian. The same happens to each of us when traveling to a new country - we always expect to find patterns of behavior that are analogous to the patterns we are familiar with in our own culture and are highly surprised when these anticipations turn out to be wrong. Young children expect an object to breath or feel pain to the extent to which they consider it analogous to the human being (Inagaki and Hatano, 1987). All these examples demonstrate that when human cognition faces a new situation it usually uses analogy with a previously experienced or familiar situation to make predictions of what to expect. These analogies do not necessarily lead to correct predictions, but this is often the best the cognitive system can do under the given circumstances (especially in a new domain, where little knowledge is present).

From computational perspective analogy-making is a good heuristics that makes effective short-cuts in the exhaustive search of the state space. Suppose you have to search for your keys. Theoretically they can be everywhere at your place or even outside it. So, if you want to be sure you will find them you should search the whole space and look under, inside, and behind every single object around you. This will, however, keep you busy forever. Analogy-making will provide you with a heuristics where to look first (rather than searching for the key randomly) - it might be worth exploring the place where last time you found your keys - it is very probable that you put them there again. There is an assumption here that there is some regularity in your actions, that you do not behave randomly, which assumption makes sense.

Finally, from a cognitive perspective analogy-making is a very central and basic mechanism that is present from very early age if not from birth (Holyoak et al., 2001), (Hofstadter, 2001), (Goswami, 2001), so it is very likely that it is used for such an important function like anticipation. In addition, in this chapter we will explore the mechanisms of analogy-making and how they emerge from simpler mechanisms like spreading activation, marker passing, and elementary transfer and anticipation in the AMBR model. Thus the relation between analogy-making and anticipation is two-fold in this model: on one hand, analogy-making is used for high-level anticipation, on the other hand, low-level anticipations are used in the process of analogy-making, and more specifically in the process of representation-building (or perception).