🚀 go-pugleaf

  John Baez writes:

(stuff from and about my previous post deleted)

The transition in question is
(|v1>|v2> +|h1>|h2>)/sqrt2 -> (|v1>e^(i phi) +
|h1>)|v2>/sqrt2
It is elementary to show that this is norm preserving in that
the initial state has the same norm as the final state (and
is equal to one) so that probability is conserved.

-----
Okay, maybe this is a case of a typo but yet again I have to note that
this transformation is NOT unitary, since the h2 that appears on the
left side is simply MISSING from the right side.  A transformation that
is not 1-1 cannot be unitary.

To which I respond

But WAIT (sorry for shouting :>), there is one vector in, and one vector out.
Both have the same norm.  Without additional assumptions, nothing can be said
about the transformation.  That assumption is of course that the transformation
acts ONLY on system 2, which I believe we agree is desireable on physical
grounds.  WITHOUT that assertion, we could construct a unitary transformation
on the _4_-dimensional hilbert space which which reproduces the indicated
transition.  It would be nonlocal as hell in the proposed circumstances,
and ugly as sin in the basis of choice (two photon states), but in and
of itself, the transition is not an indication that unitarity or linearity
has been tossed out.  I believe that this is part of which is fooling Sarfatti.
The point of my post was that, with the indicated transtion, one had a
choice, unitarity or locality, but not both.  By retaining some of the
characteristics of a unitary evolution, at least through part of his
his calculations,  Sarfatti is not getting the early indications
that his model has significant problems.

Sorry for the excessive use of CAPS but I've been discussing this with
Sarfatti directly and feel a little frustrated.
-mike gallis