Hacker News Reader

Comment by sixo | Hacker News Reader

If you make 3 subsequent measurements of different but related variables (like a vector component wrt to 3 different axes) the 3rd's probability distribution will depend on the value you observed in the 2nd observation in a way that cannot be replicated without the 2nd.

Meaning it's not just matter of poor precision. See the Stern-gorlach experiment.

View on HN

A simple version of this (not the exact experiment) is to take three polarizing filters. Put one down, and see it reduces the amount of light passing through. Put a second one down, in line with the first, then rotate it to 90 degrees off. You will observe the light dim until the second filter is completely dark, blocking all the light which passed the first filter. Finally, take the third filter, and place it between the first two, at 45 degrees off, and suddenly light will pass through all three.

This very clearly demonstrates that each filter reshapes the wave's polarization, not to a single orientation, but to a narrow band.

The inaccuracy persists, e.g. in your comment - "the wave" I can image along literal hand waving. I'm being cynic because this polarization filter experiment had me dumbfounded.

Blocking the way between the second and third filter will show no light, as if the third filter wasn't there. It appears as if there is a wave but polarized in a way that doesn't reflect from the hand and the third filter would polarize it again. I believe that's also what you are saying. Only, what is "the wave"?

> a narrow band.

Are you agreeing with Scotty79 above?

What band?

But, anyway, I don't see the analogy to Stern-Gerlach (which I only skimmed quickly on wikipedia).

I think the problem he points out is that the measurement (e.g.the polarizer/slit/field) is never perfect. In fact the measurements are also made of materials at finite temperature with quantum effects so the subsequent probabilistic results won't be "exact". Of course you can design the measurements to be exponentially insensitive to those variations. However, when you build a sequence of measurements each exponentially sensitive to that variation you make setting up each individual measurement ever more difficult.

It still doesn't prevent the thought experiment.

Yes. Because 2nd measurement will reshape the probability wave of measuresed object.

There can't be measurement without interaction and no interaction is without consequences. Also uncertainty principle.

Reply on news.ycombinator.com