Re: origins of "x self" construals

Devi Jankowicz (
Mon, 30 Mar 98 23:39:22 +0100

Dear Mike,

> This is interesting. Can you explain a little bit about what you
>mean by the following:
>> a self
>>cannot model a self (it will run up against undecidable, i.e. logically
>>contradictory assertions, Godel etc.); but provided we allow a notion of
>>a metalanguage, a self can certainly model a "self" which, perhaps, your
>>analysis doesn't allow for?

Sure. Godel (1962) pointed out that any symbol system (= language) is
going to encounter logical contradictions when it is used to make
statements about itself; Berrand Russel, I think it was, encountered this
problem in set theory and invented the notion of "logical types" as a way
of overcoming it. Generalising from his idea, it is possible to think of
a superordinate metalanguage for talking about a subordinate language,
thus eliminating the logical contradictions of self-referential
statements made in the subordinae language.

Consider a Library Index (LI) which lists library indexes. Focus on those
which don't mention themselves in their listing. What would a list of
such indexes be called? If it's called "the index of indexes which don't
contain themselves, we immediately contradict ourselves: since we've just
named them, and hence in some sense. they're "contained". BUT if we
consider the Library Index (LI) as being divided into two sections,
indexes which do mention themsleves and indexes which don't, then there's
no contradiction _so long as we view the LI index as being superordinate
to_, "talking about", the component lists.

Of course, this just removes the problem one step "higher", which is why
I'm interested in the Lefebvrian notion that there are limits to the
number of meta-meta-recursions that human beings are capable of
apperceiving (see some old correspondence on this list I had with Jack
Adams-Webber, who knows much more about all this than I do).

But, for practical terms, I find the notion of linguistic levels and
metalevels very useful in talking about the distinction between strategy
and operations in the management of business organizations. The current
debate on "learning organizations" is muddled and vague, IMHO, because it
does not distinguish between organizations which merely "learn", and
those which are "adaptive", i.e., "proper" learning organizations as I'd
define it; and, in order to be "adaptive", they have to be able to talk
about themselevs in a metalanguage as described above.

My comments came as I was finishing off a paper on this topic. Though
it's about organizations, there's lots there relevant to human learning
compatible (very compatible!) with a Kellian framework, which you'd be
very welcome to have a copy of, in draft form. Please send me your
snail-mail address and I'll post it to you: (PC users have problems,
sometimes, in decoding the diagrams which I enclose as attachments to
e-mails since I use a Macintosh, so I won't append it directly here and

Kindest regards,


In its list will be indexes which mention themselves, and indexes which
don't mention themselves. What are we to make of the latter?

Godel K.(1962) 'On Formally Undecidable Propositions of 'Principia
Mathematica' and Related Systems' trans B. Meltzer, Edinburgh: Oliver &