Oral contracts, implicit contracts, and why they're needed

Jeff Bone jbone@jump.net
Tue, 05 Jun 2001 00:25:34 -0500

Russell Turpin wrote:

> You're missing an important point. ALL contracts
> are implicit to some degree, and contain a degree
> of uncertainty. The language of commerce and law
> is not -- and cannot be -- complete.

It doesn't have to be complete as long as it's consistent, and
supports contracts that are consistent and complete with respect to
themselves.  Back to that argument...

> But
> it cannot give a complete meaning to words like
> "due," "potable," "reasonable," "cover," and all
> the others that often occur in contracts, because
> *we* cannot give these complete meanings.

So again you're asserting that there's something fundamentally
magical in meatware that's not duplicable in hardware / software?  If
*we* cannot assign meaning to the terms in our contracts, then
contracts are meaningless.  Or rather, they are assigned meaning only
when tested in tort, a sort of semiotic Schroedinger's Cat,
half-alive until the waveform is collapsed.  God may well play dice
with the universe, Russell, and we may do the same today --- but I
feel that we can do better.


Need more functionality?  Assuming the existance of human
equivalence, automate the whole process:  initial handshake,
negotiation, execution of contracts, complaint, tort, appeal...  ad
infinitum.  The agents hypothetically handle the whole thing in the
seconds before you order and after you depart.

Forgetting the science fiction for a moment...  (Well, let's not
forget it, let's put it aside:  engineering after all is a creative
way of using science to turn science fiction into everyday fact.)
You and I continually miss on one assumption.  You (and Penrose)
always seem to assume that there are things a human can do that a
machine never will.  I don't make that assumption.  So if human
beings today can deal with contract ambiguity, machines will be able
to do so in the future.  Much more expeditiously and efficiently, one

> Mathematics and formal semantics can create
> rigid connections between terms.

I'm not suggesting a particular set of tools.  Go ahead and critique
the tools.  Maybe we need fuzzy logic.  Maybe we need some kind of
Bayesian Zermelo-Frankel with choice.  Maybe we need something
completely unknown.  Point is, there's something about this mess of
sticky stuff in our skulls that can cope with this.  Machines will
eventually be able to as well.

And can it be *more* formal?  Even ignoring the machines?  Almost

> Potentially,
> this creates completeness when the bottom terms
> are themselves purely formal, i.e., in mathematics.

Very interesting to think that an admittedly incomplete formal
foundation (math) somehow yields completeness.  I think there are
some people waiting to disabuse you of your illusions, Russell...

> What I'm pointing out is an important theoretical
> limit: that there is not and cannot be an end to the
> process.

This is only true if every contingency must be covered, i.e. if you
covet completeness at the expense of consistency.  I agree that
completeness is unachievable.  I'm not sure that I believe
consistency is impossible.

> But the halting problem is unsolvable.

And I'm not trying to solve it, either.

> So is complete precision in contracts.

Um, weren't you the one arguing for it before?  (:makes a quick
position check.)

> But I think the
> courts will have contract disputes to resolve,
> even after Addie and Hal are combined to
> produce Hattie, the automated attorney.

So automate the courts.