Idea verus implementation of language requirement

A number of opinions have been expressed on the language requirement
 -- for the most part saying how good it is and that it should be
 retained.  There are, however, a couple of other factors which I think
 are important to consider.
 The usual _implementation_ of the PhD language requirement in the US
 is a test and/or course based on translating technical material into
 English.  As such, a "language requirement" does not insure that
 ability to communicate verbally in the language and does not give any
 experience in the culture associated with the language.
 Very often, the list of languages is restricted -- for example French,
 German, or Russian is probably the most common set of "allowed"
 languages.  This further destroys the notion of the language
 requirement as a broadening experience, making the PhD more worldly or
 more philosophical.
 Although there are few (if any) journals, conferences, etc. in
 computer languages, understanding them is also important to many
 peoples' research.  As a theoretical chemist, it is not suprising that
 I deal quite routinely with computer codes from other sources.  My
 wife, however, is a spectroscopist.  She too is required to write
 programs or (worse) understand other people's programs fairly
 frequently.  Although she had Fortran 101 in her undergraduate work,
 she has little practical experience and is not at all comfortable with
 it, and this slows her work.  This is a much bigger problem for her
 than needing to read papers in other languages.
 In these times of tight budgets, there are also much more practical
 reasons why languages requirements can be a problem.  Here at UF,
 there were only two departments that required a foreign language for
 the PhD (Chemistry and Music).  The German department decided that
 they could no longer afford to teach the service course that satisfies
 the requirement.  I don't know how close other departments were to
 coming to the same decision, but the problem is obvious.
 Please note that I am distinguishing between the IDEA of knowing other
 languages and the IMPLEMENTATION of the language requirement as found
 in most US universities that still have one.  The IMPLEMENTATION does
 not provide the broadening experience that was perhaps intended; nor
 does it it serve much practical use since technical materials can
 quite often be translated with the aid of a dictionary and a basic
 knowledge of the grammar -- without formal education.
 I think the IDEA remains a very valid one.  As someone else pointed
 out, however, I believe the real failing comes before the PhD level in
 the US educational system.  The idea of a "liberal arts" education
 is better implemented at the college and earlier levels.  And I have
 yet to be convinced that the PhD is an appropriate time to rectify
 this failing.
 "Between the idea and the reailty falls the shadow" -- T.S. Eliot