Human action is both intentional directed toward goals, and it is subj
ect to criteria of right and wrong. The problem of a simulation of int
entionality has been much discussed but the issue of normativity has a
ttracted less attention. It can be approached through the question of
the relation between simulations of rule-following and how normatively
constrained actions are performed in concrete situations. Writing out
explicit rules can express the norms of an activity. Skills express n
orms from the point of view of action. There are three kinds of regula
rities in behavior: causal responses, monitored actions, and habits. H
abits embody implicit rules. There are good and bad habits implemented
by the neural mechanisms produced by programming a machine. But they
also embedded in normative constraints of institutional or social orde
rs extrinsic to the action and its foundations in the ''innards'' of o
rganism or machine. Extrinsic norms are never complete, or completable
. New situations call for new normative decisions. The Turing Test is
limited in that successfully passing it can never resolve the question
of the normativity of a simulated action patten.