[FOM] Simmons Denotation paradox again

[Sandy Hodges]

I posted a while ago concerning Simmon's denotation paradox.   I derived
a contradiction from this following axiom schema, concerning the notion
"Denotes(g,x)", which has the intended meaning that g is the Gödel
number of an expression denoting x.   The false axiom schema (called the
Naive law of denotation) is:

   Let g be the Gödel number of an expression {x e S | phi(x)},
   with s the Gn. of S.   Then
    (Ey) Denotes(s,y) => Denotes(g,{x e S | phi(x)})
   is an axiom.

Here is a briefer derivation of a contradiction from this schema.
Besides the schema I use only two axioms regarding Denotation, the Axiom
of unique denotation,

    (\/ g,x,y) ( ( Denotes(g,x) & Denotes(g,y) ) => x=y )

and the Integers Gn. axiom

   Denotes(8210,Integers)

which expresses the fact that 8210 is the Gn. for the symbol "Integers",
in the numbering I am using.

Let phi(x) be ~(E y) (Denotes(x,y) & x e y).    Let g be the Gn. of "{x
e Integers | ~(E y) (Denotes(x,y) & x e y) }".

1.   (E y) Denotes(8210,y) =>
    Denotes(g,{x e Integers | ~(E y) (Denotes(x,y) & x e y) })
    ; naive law of denotation
2.   Denotes(8210,Integers)  ; Integers Gn. Axiom
3.   (E y) Denotes(8210,y)
4.   Denotes(g,{x e Integers | ~(E y) (Denotes(x,y) & x e y) })
5.   g e Integers
6.   Assume: g e {x e Integers | ~(E y) (Denotes(x,y) & x e y) }
7.   (E y) ( Denotes(g,y) & g e y)  ; from 4 and 6, E-intro
8.   ~(E y) ( Denotes(g,y) & g e y ) ; from 6 and 5
9.   contradiction. ; 7 and 8
10   ~( g e {x e Integers | ~(E y) (Denotes(x,y) & x e y) } )
    ; discharging assumption 6.
11.   (E y) ( Denotes(g,y) & g e y ) ; from 10 and 5
12.   Denotes(g, u) & g e u ; from 11, E-elim
13.    u = {x e Integers | ~(E y) (Denotes(x,y) & x e y) }
    ; from 4 and lhs of 12, Axiom of unique denotation
14.  g e {x e Integers | ~(E y) (Denotes(x,y) & x e y) }
    ; rhs of 12 and 13, Leibniz's law
15.   Contradiction ; 10 and 14 ; QED
---------------------------------------------
Discussion:   A standard way of dealing with paradoxes is with a
hierarchy of meta-languages.    Suppose in some object language "64"
denotes sixty-four.    In the meta-language, supposing 5654 to be the
Gn. of the object language expression "64", we can say:

A)    Denotes(5654,64)

Now "64" denotes sixty-four in the meta-language as well.   If we use
the same Gödel numbering for object and meta-language, then in a
meta-meta-language, we could assert

B)    Denotes(5654,64)

to say that "64" in the meta-language denotes sixty-four.    It is no
coincidence that "64" denotes the same number in the meta-language, as
it does in the object language.   The object language is the same as (or
at least isomorphic to) a part of the meta-language.   Thus B follows
from A.   Nevertheless, when we assert A, we are not asserting B, since
when we assert A we are speaking the meta-language, and therefore can't
discuss what the meta-language's own expressions denote.

Therefore, when we assert A, we are not saying that "64" denotes
sixty-four everywhere, we are only saying that "64" denotes sixty-four
in certain contexts, namely object-language ones.   "64" may in fact
denote sixty-four in other contexts, but we are not saying so.   If
someone wants to derive B from A, that's their business.

Let OLC denote the set of contexts that are object-language contexts.
Thus when we assert A, what we mean is:

C)   (\/ u e OLC) Designates(u,5654,64)

and if MLC are meta-language contexts, then B means:

D)   (\/ u e MLC) Designates(u,5654,64)

[Here I am making the distinction (an invention of my own) that an
expression "Denotes", while an expression-in-context "Designates". ]

Suppose we do have a concept "Designates."   That is, suppose it is
indeed sensible to say that an expression denotes in some set of
contexts; that it is possible to say this without thereby saying that it
denotes in all contexts.    Well, if we have such a concept, we no
longer need a hierarchy of meta-languages (at least as far as denotation
is concerned).     We can do within a single system what the hierarchy
does.  (After all, a hierarchy is a system, even if it can also be
regarded as an infinite set of systems.)
------- -- ---- - --- -- --------- -----
Sandy Hodges / Alameda,  California,   USA
mail to SandyHodges at attbi.com will reach me.