Re: [sl4] normative language is rational

From: Vladimir Nesov (
Date: Mon Feb 23 2009 - 12:46:53 MST

On Fri, Feb 13, 2009 at 7:09 AM, Johnicholas Hines
<> wrote:
> What do we mean by "should"? I'm not a philosopher of deontic logic,
> and I'm not going to try to write down any formalization of the
> concept. I don't think it's necessary or important for the moment.
> If I advocate for some action, like so:
> "I think we should (donate to the Methuselah Foundation's M prize)."
> I think everyone on this list understands, in a commonsense way, what
> I am trying to communicate.
> I hope that everyone also agrees that the future is not written. We
> can build it. Even if there are opposing forces and success is not
> guaranteed, we should strive to a better future.

The future can be considered to be at the same time written, usually
unknown, and determined by our actions. While you are deciding what to
do, and don't yet know what you'll do, you don't know the outcome of
your actions. When you decided, you do know the outcome. All the while
you know the outcome conditional on your action.

"Should" refers to what you'd like the world to be. The truth about
human "should" is in some kind of agreement for what we'd like to do
with the world, and as agreement it needs to take into account deep
knowledge about what possible outcomes entail, and how they get
evaluated on reflection. The person-independent ideal precision in
"should" comes from the same principles as precision in math, which
not limited to what an average mathematician thinks. A person can be
wrong, every living person may be wrong, but the math is defined by
what you'd reach on reflection, based on information about many
imperfect implementation of an idea.


You should read the whole sequence, and more generally most of the
Eliezer's OB posts if you weren't following or came to the party late.
See the list here (also useful for reference):

Vladimir Nesov

This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:01:04 MDT