||[Feb. 11th, 2006|12:49 am]
Matthias Felleisen wrote a paper exactly on that subject:
"On the Expressive Power of Programming Languages"
[ http://www.cs.rice.edu/CS/PLT/Publications/ ] He came up
with an interesting, formal definition that lets you tell
when one language is strictly more expressive than the other.
He _proved_ for example that a call-by-value lambda-calculus
is strictly less expressive than call-by-name lambda-calculus
(if I remember correctly).