home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: sci.logic
- Path: sparky!uunet!stanford.edu!CSD-NewsHost.Stanford.EDU!Sunburn.Stanford.EDU!pratt
- From: pratt@Sunburn.Stanford.EDU (Vaughan R. Pratt)
- Subject: Re: Multiple Truth Values
- Message-ID: <1993Jan13.050806.7235@CSD-NewsHost.Stanford.EDU>
- Sender: news@CSD-NewsHost.Stanford.EDU
- Organization: Computer Science Department, Stanford University.
- References: <TORKEL.93Jan12110752@lludd.sics.se> <1993Jan12.201545.27599@CSD-NewsHost.Stanford.EDU> <TORKEL.93Jan12232944@bast.sics.se>
- Date: Wed, 13 Jan 1993 05:08:06 GMT
- Lines: 73
-
- In article <TORKEL.93Jan12232944@bast.sics.se> torkel@sics.se (Torkel Franzen) writes:
- >In article <1993Jan12.201545.27599@CSD-NewsHost.Stanford.EDU> pratt@Sunburn.
- >Stanford.EDU (Vaughan R. Pratt) writes:
- >
- > >Australian and American courts interrogate witnesses in a language that
- > >presumes intuitionistic logic, in that they draw a distinction between
- > >straight answers and hedged and insist on the former.
- >
- > The distinction between hedging and straight answers has nothing in
- >particular to do with intuitionistic logic,
-
- I don't see what I can add to my previous posts that would make it
- easier for you to see this connection, sorry.
-
- >Furthermore, natural language is not a formal system,
-
- Neither is the trajectory of a planet or the behavior of a wisp of
- cigarette smoke. This hasn't prevented optimists from attempting to
- describe seemingly complex natural phenomena in terms of mathematically
- tractable formal systems. I can see where a committed pessimist would
- have grave doubts.
-
- >This applies equally, of course, to classical propositional
- >logic. Formal logic doesn't tell us a great deal about what reasoning,
- >interrogation, justification, etc in natural language is like.
-
- This is like someone with a one-speed bicycle announcing that no
- bicycles can climb steep hills. While you sit around grumbling that
- natural language simply *can't* be modelled formally, other people are
- looking for models that balance realism and mathematical tractability.
-
- > More interesting is the other point you raised, concerning the
- >possibility of interpreting classical in intuitionistic theories via
- >e.g. a Godel translation, where you emphasized that the effect of this
- >is to introduce distinctions where there are none in classical logic.
- >But clearly this in itself tells us nothing about the usefulness or
- >interest of the intuitionistic versions of the theories. We can go on
- >to introduce a large number of logical distinctions, still with
- >Godel-type translations of classical theorems. Such distinctions have
- >no intrinsic value. Looking at the distinctions introduced by
- >intuitionistic logic, they are in part natural in the sense that they
- >have a connection with or correspond to distinctions that we make in
- >mathematics and that we have found useful, such as the distinction
- >between direct and indirect proofs of existential statements. But
- >intuitionistic logic involves very much more than such natural
- >distinctions. For example, the logical distinction between ~~ExP(x)
- >and ExP(x), with P a decidable number-theoretic predicate, corresponds
- >to nothing in ordinary mathematical experience.
-
- Let P(x) be the predicate "x is the least Mersenne prime with at least
- a million binary digits." This predicate is not only decidable but can
- be decided by a finite state automaton in the time taken to read the
- the input. Yet a constructivist of the appropriate stripe will today
- allow ~~ExP(x) while questioning ExP(x). It is reasonable to expect
- him to agree to ExP(x) in less than five years time, at the current
- rate of production of Mersenne primes (up to some 200,000 bits today I
- think). I'd say this distinction was not only well within mathematical
- experience but one that is very familiar to every amateur number
- theorist.
-
- >For another example,
- >the splitting up of the concept of an infinite set of natural numbers
- >into "infinite", "not not infinite", "not bounded" introduces
- >considerable complications, and we don't even know whether it is possible to
- >teach and learn mathematics without "conflating" these concepts. We need
- >to think about such matters as well in considering distinctions.
-
- If "infinite" means "not finite" then isn't "not not infinite" a
- synonym for "infinite"? But in these murkier waters involving hidden
- quantifiers I concede that I am a much less capable swimmer than many
- others, for all I know yourself included.
- --
- Vaughan Pratt There's safety in large condition numbers.
-