Lambda the Ultimate

inactiveTopic No "theory of types" in legal systems
started 9/4/2003; 5:35:34 AM - last post 9/4/2003; 4:54:18 PM
Andris Birkmanis - No "theory of types" in legal systems  blueArrow
9/4/2003; 5:35:34 AM (reads: 274, responses: 4)
While not (directly) related to PL, this book discusses paradoxes in legal systems, mostly in terms of reflexive laws, clauses, etc. As some paradoxes in math (and PL design?) were avoided using some type system, one could wonder, is there a way to do something similar for law.

I think Section 20 is the best place to start with.

By the way, did anybody hear about Nomic recently?

Ehud Lamm - Re: No "theory of types" in legal systems  blueArrow
9/4/2003; 5:47:26 AM (reads: 288, responses: 0)
When I saw the link ("this book") I just knew you are talking about Peter Suber...

Neel Krishnaswami - Re: No  blueArrow
9/4/2003; 4:48:01 PM (reads: 260, responses: 1)
According to economics, contracts are inherently unformalizable using the standard machinery of formal semantics.

This is because there are costs associated with specifying a contract, which rise with the specificity of the contract. So at some point it becomes economically attractive to specify an incomplete contract, because the benefits of clearer specification are outweighed by the costs of specification. (A contract that specified all possible contingencies would be infinitely long, for example.) Instead, the parties to a contract rely on mutual understandings to avoid disputes, and if there is a dispute they go to court (or use a pre-agreed mediator, pay a fee to terminate the contract, etc) to resolve it. So most court disputes over contractual obligations are over things that weren't in the contract to start with! So a semantics for contracts would require you to include the expectations in each party's head, as well as their strategic assessment of what the other side thinks, and so on. It would look like game theory, not like a PL-style operational or denotational semantics.

There is a sub-field of transaction cost economics called incomplete contract theory, which studies how people behave, given the various institutional and contractual frameworks they find themselves in.

Neel Krishnaswami - Re: No  blueArrow
9/4/2003; 4:54:18 PM (reads: 250, responses: 0)
Addendum: You can get a feel for the flavor of incomplete contract theory by reading the paper "Foundations of Incomplete Contracts" by Hart and Moore. It's chock full of game theory, as promised. :)

Frank Atanassow - Re: No  blueArrow
9/5/2003; 5:41:14 AM (reads: 274, responses: 0)
So a semantics for contracts would require you to include the expectations in each party's head, as well as their strategic assessment of what the other side thinks, and so on. It would look like game theory, not like a PL-style operational or denotational semantics.

I don't find that very convincing. By that argument, to give a PL's semantics you would need to describe not only what is expressed in a program but also the environmental dependencies (like the OS, network disposition, etc.) which the programmer relies on when writing a program. Maybe there is a place for that, but it's not what we call PL semantics.

Similarly, one ought to be able to describe the semantics of a written contract independently of the external assumptions in the minds of the parties involved. You may argue that such a semantics wouldn't characterize the entire situation, but it would still be useful to those parties, and particularly to a judge who needs to resolve a dispute.

And, BTW, there are programming languages (such as the lambda-mu calculus) which have been given game-theoretic models.