Logic and Law

[pycyn@aol.com]

Reading a thread on the international use of Lojban, specifically as a 
language for patents and so a part of the legal apparatus, reminded me of one 
important fact about legal reasoning.  Quite aside from the use of inventive 
equivocation and ad hoc reinterpretation which play major roles both in 
drafting legal documents and in presenting cases in court, there is the fact 
that the logic of the law is not classical first order (or higher) predicate 
logic, but rather some kind of local relevance logic.  The classic example of 
this comes from the MULL project of a couple of decades ago.  In classical 
logic, an inconsistent set of premises validly entails absolutely any 
sentence at all.  The MULL project demonstrated that the US Tax Code was in 
fact inconsistent.  BUT it also found that there were conclusions (court 
verdicts, indeed) that could not be derived from that Code, even though they 
could be made to follow from it.  The failure was typically the result of the 
demonstration taking into account  parts of the code that were not judged to 
be germane to the issue at hand -- even though a logically tight (classical) 
connection could be made.  Many of the unaccepted demonstrations did not even 
take advantage of the inconsistency of teh Code.
So, on the one hand, having the law in a logical format (one of the 
achievements of MULL), allows us to see these facts and also allows us to 
begin to work out the rules that underlie the relevant sense of relevance, 
but, on the other hand, they do not allow us simply to use the apparatus to 
draw conclusions.