Skip to Content.
Sympa Menu

coq-club - Re: [Coq-Club] Is this apparently decidable existential provable?

coq-club AT inria.fr

Subject: The Coq mailing list

List archive

Re: [Coq-Club] Is this apparently decidable existential provable?


Chronological Thread 
  • From: Xavier Leroy <Xavier.Leroy AT inria.fr>
  • To: coq-club AT inria.fr
  • Subject: Re: [Coq-Club] Is this apparently decidable existential provable?
  • Date: Mon, 27 May 2013 15:33:33 +0200

On 2013-05-27 15:19, Floyd Lee wrote:
> Parameter A : Type.
> Parameter P : A -> Prop.
> Parameter Q : A -> Prop.
> Parameter P_dec : forall (x : A), {P x}+{~P x}.
> Parameter Q_dec : forall (x : A), {Q x}+{~Q x}.
>
> Definition R :=
> (exists r, P r /\ Q r).
>
> (* Both P and Q are decidable, so R should be too. *)
> Theorem R_dec : {R}+{~R}.

No, unless the type A is finite.

Example: let TM be an arbitrary Turing machine and take

A = nat
P = fun n => TM is halted after n steps
Q = fun n => True

P x is clearly decidable (just run TM for n steps and check whether it
is halted), but R is equivalent to "TM halts", so deciding R (for an
arbitrary TM) would be solving the halting problem.

- Xavier Leroy



Archive powered by MHonArc 2.6.18.

Top of Page