40
that all possible concatenations of the models are potentially represented in the mind as morally
equal. There may be rules of combination for the models of an absolute, non-parametric sort
such that some concatenations are better formed than others. Perhaps the extreme reliance on
Market Pricing in the West and the extreme reliance on Authority Ranking in India are both
examples of poorly formed concatenations and hence are inferior to other possible constructions.
There need be nothing paradoxical in finding actual social structures that violate an internalized
rule constraining the combinatorial powers of the models. Once again, the distinction between
competence and performance is relevant: Just because there is a certain mental representation of
social life does not mean that actual social life will always precisely resemble it. Other factors
may intervene. One possibility is that large population densities, larger than those found in the
human ancestral environment, have forced people to rely on poorly formed concatenations,
concatenations which would not be found in small groups. One may have to rely on
concatenations which violate some internal principle in order, for example, to coordinate
interactions in extremely large groups. Overapplying Market Pricing or Authority Ranking may
be a practical necessity and yet, at the same time, deeply morally unsatisfying. If one could find
a means of mass coordination which would avoid these concatenations, it would better cohere
with our deepest values. This is speculative, but to dismiss it as mere speculation would be to
miss the point: The psychological reality of the four models does not rule out there being some
unconditional moral principles which are absolute rather than parametric.
Furthermore, there is no reason why one couldn’t accept the psychology of relational
models as only part of the truth about moral competence. For Chomsky, many linguistic
universals exhibit a P&P structure, but not all do. The first linguistic universal mentioned in this
paper, namely that all movement is structural, is not parametric. The matter calls for further
investigation, but even if the relational models take a P&P form, there may be other mental