Jesse, you need to fix up your email client to follow the usual
quoting conventions, wrap lines etc. Below is how your text appears in mine:

On Sun, Feb 08, 2009 at 06:46:04AM -0500, Jesse Mazer wrote:
> Russell Standish wrote:> > According to Wikipedia, Born's rule is that the 
> probability of an> observed result \lambda_i is given by <\psi|P_i|\psi>, 
> where P_i is the> projection onto the eigenspace corresponding to \lambda_i 
> of the> observable. > > This formula is only correct if \psi is normalised. 
> More correctly,> the above formula should be divided by <\psi|\psi>.> > This 
> probability can be interpreted as a conditional probability - the> 
> probability of observing outcome \lambda_i for some observation A,> _given_ a 
> pre-measurment state \psi.> > What is important here is that it says nothing 
> about what the state> vector is after the measurement occurs. There is a (von 
> Neumann)> projection postulate, which says that after measurement, the 
> system> will be found in the state P_i|\psi>, but as I said before, this is> 
> independent of the Born rule, and also it does not state what the> 
> "amplitude" (ie magnitude) of the state is. The v-N PP is also distinctly 
> not> a feature of the MWI (it is basically the Copenhagen collapse).The 
> projection postulate needs to work *effectively* in the MWI though, in order 
> for it to make the same predictions about actual experimental results as the 
> Copenhagen interpretation. If an observer makes one measurement and then 
> makes a later measurement of the same system, they can collapse the system's 
> quantum state onto an eigenstate at the moment of the first measurement and 
> evolve that new state forward, and this will give correct predictions about 
> the probabilities of different outcomes when the second measurement is made. 
> I think part of the difficulty with connecting the MWI to actual observed 
> probabilities is explaining *why* this rule should work in an effective 
> sense.> > > I think the quote I was responding to was the following:> > "In 
> an ordinary quantum mechanical situation (without deaths), and> assuming the 
> Born Rule holds, the effective probability is proportional> to the total 
> squared amplitude of a branch."> > If you compare it with the description of 
> the Born rule above (which> computes a conditional probability), there is no 
> sense in which one> can say that "the effective probability is proportional 
> to the total> squared amplitude of a branch" follows directly from the Born> 
> rule. Jacques is assuming something else entirely - perhaps> einselection?The 
> notion that probability is proportional to squared amplitude is equivalent to 
> the wikipedia definition you posted above. The projection operator P_i onto a 
> given eigenstate |\lambda_i> is really just |\lambda_i><\lambda_i| (see the 
> text immediately above 'Postulate 5' near the bottom of the page at 
> ), which means that when 
> this operator acts on a given state vector |\psi>, it gives 
> (|\lambda_i><\lambda_i|)|\psi> = |\lambda_i>(<\lambda_i|\psi>), and 
> (<\lambda_i|\psi>) is just a scalar c_i, so this becomes c_i * |\lambda_i>. 
> This c_i is referred to as the "amplitude" that the original state |\psi> 
> assigns to the eigenstate |\lambda_i> (by something called the 'expansion 
> postulate' it's possible to write |psi> as a weighted sum of all the 
> eigenstates of a measurement operator, and the weights on each eigenstate are 
> just the amplitudes for each eigenstate). So, if the probability is 
> <\psi|P_i|\psi> (which I think assumes that |\psi> has been normalized), then 
> since P_i = |\lambda_i><\lambda_i|, this is the same as saying the 
> probability is <\psi|\lambda_i><\lambda_i|\psi>. If <\lambda_i|\psi> is the 
> amplitude c_i, then <\psi|\lambda_i> is just the complex conjugate (c_i)*, so 
> the probability is just the amplitude times its own complex conjugate, often 
> referred to as the "amplitude-squared".Jesse

By working *effectively*, I think you mean the subjective experience (or
1st person experience) should be as though \psi is transformed to
\psi_i, where \psi_i is an eigenvector of the relevant observable.

Of course no such thing happens in the MWI, ie the 3rd person
description. \psi remains unaltered in this case, and is undergoing
regular unitary evolution.

In any case, there is no requirement for <\psi_i|\psi_i> to be given
by the product of the Born rule probability with <\psi|\psi>. This is
an additional process that Jacques Mallah is adding, and I would like
to see his reasoning for doing so. I'm not entirely convinced it can
be done consistently without additional hypotheses (such as eg
einselection) - for example in an experiment where some
observers in the MV are measuring circular polarisation and others are
measuring linear polarisation, the Born rule probabilities treat the
circular polarisers differently to the linear polarisers - they're in
different sample spaces.

As I mentioned before, the conventional treatment is to set
<\psi_i|\psi_i>=1, but this is a calculational convenience only. There
is, AFAICS, no constraint on the value of <\psi_i|\psi_i> in
conventional QM.



A/Prof Russell Standish                  Phone 0425 253119 (mobile)
UNSW SYDNEY 2052               

You received this message because you are subscribed to the Google Groups 
"Everything List" group.
To post to this group, send email to
To unsubscribe from this group, send email to
For more options, visit this group at

Reply via email to