The theory of subjective expected lexicographic utility brings togethe
r two classical developments in expected utility theory. The first is
Hausner's theory of expected lexicographic utility in decision under r
isk. The second is a lottery-based theory of subjective expected utili
ty in decision under uncertainty that was first axiomatized by Anscomb
e and Aumann. Our synthesis of the two produces representations of pre
ference in decision under uncertainty in which utilities are finite-di
mensional real vectors ordered lexicographically and subjective probab
ilities are real matrices. Axiomatizations of subjective expected lexi
cographic utility are described for finite and infinite sets of states
. Procedures for assessing vector utilities and matrix probabilities a
re outlined.