The Size of Random Variations to a Natural Information Store

The Size of Random Variations to a Natural Information Store
Natural information stores have mechanisms to ensure that variations to
the store are small. If, in order to deal with a very complex, variable
environment, a store is very large, then relative to its size, any usable
alterations will constitute a minute proportion of the total store. A
large variation in the store will almost certainly disrupt essential functions
and so is incompatible with the continuation of a natural store in a natural
environment.
Individual mutations and genetic recombination that permit continuation
of a species constitute a very small proportion of a genetic code. A
substantial genetic shift will take many thousands or even millions of years.
The huge overlap in the genetic code of species that separated millions of
years ago is a testament to the stability of genetic codes. Changes over short
periods are minute. Only such small variations are viable. Large variations
do not survive. Similarly, as indicated by the working memory limitations
Evolution of human cognitive architecture 239continuum, human working memory ensures that alterations to the long-
term memory store are relatively slow and small.
In summary, mutation and sexual recombination result in quite random
variations analogous to the random choice of moves faced by a person
solving a problem for which schema-based solutions are not available. The
usefulness or otherwise of a genetic variation can only be assessed after it
has occurred. If it is successful, information in the genetic code will be
passed on to subsequent generations, whereas a failure will result in a
genetic dead end with the information not passed to subsequent generations.
Similarly, when limited or no knowledge is available to a problem solver,
moves must be chosen randomly. Successful moves may be incorporated in
schemas that then can be used indefinitely when faced with similar
circumstances. Unsuccessful moves result in dead ends with information
not incorporated in schemas and not used subsequently.
Under this formula, a schema encapsulates psychological information in
the same way that a gene encapsulates genetic information. Both can be
reproduced indefinitely, providing the environment supports the use of that
information. Nevertheless, alternative schemas/genes may be more appro-
priate for environmental conditions. If inappropriate, the structure of the
information encapsulated in schemas or genes must change. Changes or
variations are generated randomly and tested against the environment. If
successful, a new schema or gene will be constructed and used in future.
Thus, natural selection and the processing of information by human
cognitive architecture can be characterized as identical ways of handling
very complex information.