On Memory Supersets (was RE: (level seven) Further Discussion of Identity

From: Lee Corbin (lcorbin@tsoft.com)
Date: Tue Nov 12 2002 - 17:16:05 MST


Eliezer (for whom personal survival has never been an issue
and so cannot properly be said to occupy any level in the
identity hierarchy) writes

> Lee Corbin wrote:
> >
> > Also, what do you think of the maxim, "Any memory superset
> > of me is me"?
>
> I'd deny this one. For example, if Lee Corbin has access to Eliezer
> Yudkowsky's memories but the Lee Corbin memories/skills/philosophy are
> solely in control of decisionmaking, including reflective decisionmaking,
> then I would call you "Lee Corbin with access to Eliezer Yudkowsky's
> memories" but not "Eliezer Yudkowsky".

I agree, basically, but you are supposing that I would have
a way to quarantine the Eliezer memories in order to avoid
them contaminating my thinking process. Suppose that I am
just as likely to reflect upon, and be influenced by, an
Eliezer memory as a Lee memory. What then?

> (At least until my static memories took over your personality using
> your own goal system, but that would be an essentially friendly process.)

Being much more recent than many of my own much longer term
memories, the resultant entity

> On the other hand, if my current personality is "only a memory"

I'm not really sure what that means

> but I [a form of "present-time"-you, speaking about some time in
> the future] grew out of that personality by a continuous process
> of a kind that I regard as valid(*), I would regard that entity
> as being Eliezer Yudkowsky.

As an example, barring a Singularity and so on, you would
consider a 92 year old Eliezer in the 2070's as E.Y.?
Do you think that the 92 year old E.Y. will regard himself
as the same person you are today?

As for me, I can only put my identity on a continuum and
say that I'm not the same person that I was at age 8, but
am mostly the same person I was at age 25, and definitely
the same person I was ten years ago.

> But there are some kinds of disruptive supervention that
> could destroy continuity of identity without destroying
> continuity of memory.

I conjecture that we may speak consistently of a human's
behavior dispositions and basic_personality independently
of that human's memories. If I discovered a long lost
identical twin, for example, while he'd not have my
memories, evidence suggests that he'd share my basic
personality traits and behaviorial tendencies.

> (*) Yes, I realize how much buried territory there is inside this
> statement. But let's just take it that there are some kinds of dynamic
> changes that I regard as valid, and some that I don't.

It's clear that there is a class of transformations that
would leave you the same person, and there is a class that
doesn't. There is also a shady class that sort-of does,
and sort-of doesn't. The language of the continuum is
what we're stuck with, and is most appropriate, IMO.

Lee



This archive was generated by hypermail 2.1.5 : Wed Jan 15 2003 - 17:58:04 MST