>>>    BOOLEAN is_x_larger( double x[ ], double y[ ], int n )

> Dishonest editing undone:

> [ >> Pretend double is an infinite precision real number. ]


I didn't see that, Rob.  Great use of your time, by the way.  I'm sure
dumping out programming code has really cleared up to all concerned in
what circumstances preference relations have utility representations.
Particularly when you have to add as an afterthought that the code you
provided as an example of a pathological case without representation
actually, unless we "pretend," does not define that case.


> Assumptions of "rationality" are insufficient for preferences to be
> represented by a utility function.


True.  Now, we are still all patiently waiting for you to provide just
one little example of either a study which assumed away lexicographic
preferences in a context where you think that a better basis for the model.
Or even a contrived example of lexicographic preferences.  Recall,
lexicographic preferences are "perfectly reasonable" but "discarded by
assumption" in neoclassical economics nonetheless, so surely you can come
up with many, many important examples of such commonplace, perfectly
reasonable preferences?