This is a question about selfreference: Has anyone established an abstract framework, maybe a certain kind of formal language with some extra structure, which makes it possible to define what is a selfreferential statement?

11$\begingroup$ If there is, it would be interesting to use it to decide if your question is self referential. :) $\endgroup$– Dick PalaisSep 22 '10 at 15:40

1$\begingroup$ If it turns out that not, I would have to rephrase my question :) $\endgroup$– Peter ArndtSep 22 '10 at 16:06

1$\begingroup$ Are you familiar with Vicious Circles by Barwise and Moss? $\endgroup$– Noam ZeilbergerSep 22 '10 at 16:24

3$\begingroup$ The "Recursion Theorem" of classical Recursion Theory, covered e.g. in Hartley Roger's "The Theory of Recusive Functions and Effective Computability" gives the beginnings of such an abstract framework. Also check out Smullyan's book "Diagonalization and SelfReference", which tries to abstract as much as possible from Godel numbering. $\endgroup$– Sidney RafferSep 22 '10 at 16:46

5$\begingroup$ Now Smullyan together with Lawvere's diagonalisation argument, as exposed at arxiv.org/abs/1006.0992, leaves me quite satisfied... $\endgroup$– Peter ArndtSep 23 '10 at 14:41
I am not quite sure if it fits the bill but you can also check out:
N. Yanofsky  A Universal Approach to SelfReferential Paradoxes, Incompleteness and Fixed Points

$\begingroup$ Ah, another exposition of Lawvere's argument  it certainly fits the bill, see my last comment to the question. Thanks! $\endgroup$ Sep 23 '10 at 23:40
For pleasure only I can at least give you the shortest definition of self reference.
You need only to look in a good dictionary ( from Borges world of course) it says:
Selfreference : see selfreference.

1
Raymond Smullyan, "Diagonalization and SelfReference", 1994

$\begingroup$ Thanks, I already got it, following the advice of SJR in the comments. It is indeed excellent! $\endgroup$ Sep 23 '10 at 14:19


$\begingroup$ No "sorry": It deserves an answer of its own:) $\endgroup$ Sep 25 '10 at 20:06
You might also be interested in Graham Priest's article "The Structure of the Paradoxes of SelfReference", Mind 103 (1994) pp. 2534. (Journal page ; JStor) and similar work by Priest. He has a general framework that he argues captures the various selfreferential paradoxes. I believe he also discusses this in some of his other work and monographs.
Perhaps the right question to ask is if the statement is expressible in any system whose prooftheoretic ordinal is smaller than the Feferman–Schütte ordinal.

$\begingroup$ Hm, could you clarify that? So, to make sense of prooftheoretic ordinals I need a language in which I can talk about some fragment of arithmetic. So maybe your proposal is to say a statement in a formal language is selfreferential if under some/any translation (to be defined, maybe via a Gödelization?) into the language of arithmetic it becomes selfreferential (this depends on a Gödelization of the language of Arithmetic)? $\endgroup$ Sep 22 '10 at 16:11

$\begingroup$ I don't quite like the dependence on a Gödelization, but maybe there is no other way. Maybe there is always a choice of Gödelization that renders a given statement selfreferential... $\endgroup$ Sep 22 '10 at 16:11

1$\begingroup$ I think that before I can make a more rigorous proposal about what a selfreferential statement is, we first need to set boundaries on what qualifies as a statement. If we need to take into account such sentences as "colorless green ideas sleep furiously", then I wouldn't know where to begin. $\endgroup$– dfrankeSep 22 '10 at 16:18
I found another one:
John Bell, “Incompleteness in a General Setting”. Bulletin of Symbolic Logic 13, 2007. It's paper number 66 here
Have you checked Craig Smorynski's work such as "Modal Logic and SelfReference" (Google Books)?
I have heard that Kapranov once said that he really wants to understand what is selfreference (i.e. is/was working on the question).
1) Let $C$,$P$, $T$, $R$ be sets. (as a mnemonic, you can think $C$ as the set of "claims", $P$ the set of "properties", $R$ the set of "binary relations" and $T$ the set of "terms").
2) We assume there is also
2.1) a map $\alpha_1: P\times T \to C $; (obviously builds a claim from a property and a term: this is purely syntactic)
2.2) a map $\alpha_2: R \times T \times T \to C$; (builds a claim form a relation and two terms; syntactic too)
2.3) a map $\#_0: C \to T$ (a Gödel numbering).
2.4) a map $\#_1: P \to T$ (idem)
2.5) an equivalence relationship $\equiv$ in $C$ (it will be often equality itself, or for instance when $a,b$ are logical formulas, $a \equiv b$ if and only if $\vdash a \leftrightarrow b$ ).
3) We also assume the following : for every $f \in P$, there exists $\overline f\in R$ such that for every $(u,x)\in P \times T$, $\alpha_1 \big (f , \#_0 \alpha_1 (u,x) \big)\equiv \alpha_2 \left (\overline f, \#_1 u,x \right)$
4) Finally we assume that for every $g\in R$, there exists $g^* \in P$ such that for all $v\in P$, $\alpha_2(g,\#_1 v,\#_1 v)\equiv \alpha_1(g^*,\#_1 v)$.
We have the following fixed point (recursion...) theorem: under all the assumptions 1) to 4) above, For every $h\in P$, there exists $e\in C$ such that $\alpha_1(h,\#_0 e)\equiv e$.
Proof: from 3) and 4), it follows that for every $w\in P$, we have:
$$\begin{align} \alpha_1 \big(h, \#_0 \alpha_1 (w,\#_1 w)\big) & \equiv \alpha_2 (\overline h ,\#_1 w, \#_1 w) \\ & \equiv \alpha_1 \big( (\overline h)^*, \#_1 w \big)\tag{I} \end{align}$$
now define $e:= \alpha_1 \big((\overline h)^*, \#_1 (\overline h)^*\big)$. We get immediately $\alpha_1(h,\#_0 e)\equiv e$ by replacing $w$ by $(\overline h)^*$ in $\text{(I)}$. Qed.
$e$ says the same as "my Godel number has property $h$".
How relevant this is to "an abstract framework, maybe a certain kind of formal language with some extra structure" in the OP, I am not quite sure, and I am only beginning to read Halbach's works, yet I think they are general enough to properly belong to this thread, and they weren't yet mentioned. Very very roughly, what I gather from Halbach's work on this topic so far is that (my interpretation)
the 'say' in the old idea of formulae which 'say' about themselves they are not provable is not an idea which is synthetic a priori (like e.g. the idea of necessity), rather is an idea consisting of many subideas. (For example, by filtering according to where the relevant formulae are placed in the arithmetical hierarchy.)
Philosophically, the approach to selfreference I am pointing to here could be called an 'analytic' approach to selfreference, as opposed to the 'synthetic' approach of Lawvere. (Both 'analytic' and 'synthetic' to be taken in their neutral technical sense, i.e. breaking a concept into parts in the former, combining the concept with other concepts in the latter.)
Some references on Halbach's work:
so new that it hasn't yet appeared: Volker Halbach has announced a book (with Graham Leigh) whose draft has the title Syntax and Circularity: A Study in SelfReference and Paradox.
so new that it lies after the answers given so far:
Volker Halbach, Albert Visser: Selfreference in arithmetic I. Vol. 7(4), 2014 , pp. 671691
Volker Halbach, Albert Visser: Selfreference in arithmetic II. Vol. 7(4), 2014 , pp. 692712
 a relevant lecture of Halbach's with a very general title, containing explanations on the work of Halbach and Visser:
There's a bunch of writing on this topic by the philosopher of mathematics Charles Chihara, includng a book called "Ontology and the Vicious Circle Principle". I haven't read that one but he also discusses the topic in his later book "Constructability and Mathematical Existence". You can probably find reviews of these books online that would help you decide if they are relevant to your interests.