Of Legal Language and CS Notation

["Peter J. Wasilko" <address@hidden>]

Hi Shriram,

> (As a personal aside, do you eschew legal language when speaking about
> legal matters to a fellow member of the bar?  If so, I'm rather
> curious about how well this works.  Do you think legal language exists
> *solely* to create a barrier to comprehension by people not already
> members of the club?)

The short answer is yes, lawyers don't really talk that way other than
to key in on issues. And no, legal language is really more of a
historical accident.

But this should come as no surprise, after all, how do you talk about
types if you can't write down the two dimensional type equations we see
in the literature? 

Actually, the whole notion of legal language has shifted dramatically
over the last twenty years and perhaps our experience in law might be of
value to the CS community.

What we found was that our "legal language" was acting as an impediment
to our serving our clients. So legal educators made a conscious decision
to break future lawyers of the worst habits in this regard by offering
formal courses in legal writing during the first year and to follow up
on that work in subsequent "clinic" courses that combine legal writing
and substantive law.

The emphasis of these courses is on using Plain English. Students who
pepper their papers with archaic latinisms and pursuant's and whereas's
are punished with bad grades and quickly broken of the desire to string
an entire logical proof into a single sentence.

We were even made to read books like Veda R. Charrow and Myra K.
Erhardt's "Clear & Effective Legal Writing" ISBN 0-316-13771-5.

As a result, about the only legal language we use consists of the names
of legal principles and famous cases that act as shorthand index
pointers. But once we are on the same page, other than referencing
specific sections of a bill or statue we are using ordinary language,
although we are extremely careful to spell out our arguments and banish
any traces of ambiguity from what we say.

So we use the same logical rigour that one would in CS while restricting
ourselves to the ASCII character set.  :-)  At least until you start
looking at papers authored by the Law and AI crowd.

Interestingly, I would say that most artifacts of legal language were
common usage when they originated. Then they were maintained in their
historical form as the vernacular evolved around them. This may have
helped inflate lawyers egos when we would still grok the old language
and lay people couldn't. But eventually forces in the marketplace led
firms and educators to realize the using clear language was a
competitive advantage. So newly authored legal scholarship strives for
plain English while standardized older documents often don't.

Ironically, this is why lay people find that it is harder to understand
the legal boiler plate they come into contact with, when they would have
considerably less difficulty penetrating cutting edge law review
articles.

Assuming that the underlying ideas of CS are of the same order of
complexity as those in law, a plain English movement in CS curriculum
design might be worth exploring. In any case, I would argue that the
problems of effective notation and effective exposition have been
conflated in more than one CS subdiscipline. Just because a notation
arises from the historical evolution of a field does not mean that it
serves its practitioners well.

--- Peter