Re: personal identity

Nick Bostrom (
Sat, 10 Jan 1998 00:07:25 +0000

Wei Dai < wrote:

> On Thu, Jan 08, 1998 at 10:53:18AM -0800, Hal Finney wrote:
> > I see a couple of problems. The first is a trivial matter of language.
> > It is very confusing to write that A is B but B is not A. Normally we
> > use the term "is" in a symmetrical way. It's not even clear which
> > way to interpret a question like "are you the same person you were a
> > second ago?". So you might want to use a different phrasing.
> I don't think using "is" in an asymmetric way is necessarily confusing. We
> handle sentences like "a penguin is a bird, but a bird is not necessarily
> a penguin" without too much trouble. The reason I use "is" is that we
> often hear questions like "is an upload of me really me?" from people who
> have no clear idea of what "is" means in that context. I am suggesting
> such a meaning.

There is a difference. If I say "Fido is a dog" that is a of the
logical form Dog (Fido). The "is a" is a way to indicate predicate
instantiation and is really superfluous. That is very different from
"3+2=5" or "The morning star is the evening star." which are of the
form a=b, or equivalently, Identity(a,b); here the "is" is a
predicate itself. Here we are using the identity predicate
which is per definition symmetric. And that is presumably what Hal
thinks it looks as if you are doing when you say about two
person-segments that they are the same person, that the person
to which the former person-segment belongs is the same person as
that to which the latter person-segment belongs. This looks like a=b.

Not that we might not need a new grammar for personal identity. Our
present notion seems to be hopelessly vague and confused. But it
might be best to use new terms to signify the new usage you propose.
Stop talking about personal identity and start talking about personal
containment, for example?

Nick Bostrom

*Visit my transhumanist web site*