Re: Linear Logic Maybe?
From: Jan Burse (janburse_at_fastmail.fm)
Date: 02/11/05
- Next message: Andrew Eremin: "Re: Commercial uses of prolog (to convince my boss it's valid)"
- Previous message: Justin Pearson: "Re: Linear Logic Maybe?"
- In reply to: tmp123: "Re: Linear Logic Maybe?"
- Next in thread: tmp123: "Re: Linear Logic Maybe?"
- Reply: tmp123: "Re: Linear Logic Maybe?"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Date: Fri, 11 Feb 2005 11:43:32 +0100 To: tmp123 <tmp123@menta.net>
Hi
tmp123 wrote:
> Probably I do not understand something, because if
> "!" means "->" ("then")
> "&" means "/\" ("and")
> "v" means "\/" ("or")
With ! I have more a composition/application
in mind then an implikation.
Let @ denote the application. It takes an
argument f and g, where f is a function A->B
and g is an element of A. Then f@g is an element
of B. Namely:
f:A->B, g:A
-----------
f@g:B
Then ! can denote the inverse composition as
follows. It takes argument h and f, where h
is a function B->C and f is a function A->B.
Then h!f is a function A->C. Namely:
h:B->C, f:A->B
--------------
h!f:A->C
We can model ! by @. Namely h!f = \x.h@(f@x)
where \x. is the lambda abstraction.
Now we can assume that Nouns are "Merkmale",
thus they map Objects to Subobjects. Namely
a Noun is a function O->O, where is O is
the set of a objects, an infinite set.
Further we can assume that Adjects are
"Ausprägungen", thus they map Objects to
Truthvalues. Namely an Adjective is a
function O->2. Where 2 ist the set {0,1},
and 2 and O are disjoint.
In the example we had apples and bananas.
Lets assume apples and bananas refers not
to apples and bananas in general, but we
have a situation which is modeled as a
record {apple:x,banna:y}.
apple:O->O
apple{apple:x,..}=x
banana:O->O
banana{banna:y,..}=y
In the example we had also red and green.
Lets assume that the x and y are futher
records of the form {color:z,...}. Then
we can define:
red:O->T
red{color:"red",..}=true
red{color:"green",..}=false
green:O->T
green{color:"red",..}=false
gren{color:"green",..}=true
Now red!apple means that the apple in the
current situation is red. The expression
is of the type O->T:
red:O->T apple:O->O
--------------------
red!apple:O->T
Now (red v green) can be easily interpreted
as the truth function which results from
the truth function of red and green by
the following rule:
p:O->T, q:O->T
--------------------------------------
(pvq)(x) := p(x) v q(x), (pvq):O->T
But what happens if we make a disjunction
or conjunction of Nouns?
p:O->O, q:O->O
-----------------
pvq ?? q&p ??
What disjunction and conjunction should
be defined. How could a calculus look like?
Best Regards
--- Spuntik III
- Next message: Andrew Eremin: "Re: Commercial uses of prolog (to convince my boss it's valid)"
- Previous message: Justin Pearson: "Re: Linear Logic Maybe?"
- In reply to: tmp123: "Re: Linear Logic Maybe?"
- Next in thread: tmp123: "Re: Linear Logic Maybe?"
- Reply: tmp123: "Re: Linear Logic Maybe?"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Relevant Pages
|
|