Is there any special reason why any? and every? in SRFI-1
   work only on clists, i.e., not on improper lists?  This bit me
   on a code walker for lambda formals lists.

Check the discussion archives. I, personally, was pushing for a uniform
approach wherein *every* value is a list, and any non-pair is an empty
list. I wrote some long messages explaining why I felt this was the right
thing to do, in the Scheme world. But there was major pushback on the
discussion list; the general community didn't sign up for this. The major
counterargument, as I recall, was losing the ability to catch common errors
(such as flipping a list & an elt parameter on a function call) in exchange
for non-common generality.

So I acceded, and there you are.
    -Olin