Question about Entropy/Shannon Information

  • Thread starter Thread starter celal777
  • Start date Start date
  • Tags Tags
    Information
AI Thread Summary
The discussion centers on the apparent contradiction between increased entropy and the claim of increased information in an experimental setup involving magnets. The original poster questions how entropy can increase while information simultaneously increases, highlighting a conflict in interpretations. A respondent suggests that the original poster is using an intuitive rather than a quantitative definition of entropy, asserting that the system is more orderly despite increased entropy. However, other participants argue that the scattering of magnets leads to increased entropy and a loss of useful information, emphasizing that while data quantity may rise, meaningful information decreases. The conversation reflects ongoing confusion and debate regarding the relationship between entropy and information theory.
celal777
Messages
11
Reaction score
0
Hello,

I would like second (and more) opinions on comments i received from a physicist in response to my observations on a certain experimental setup designed to make a particular point about information theory in the context of entropy.

The experimental narrative is found on pages 9 & 10 of the following document :
http://www.colorado.edu/philosophy/vstenger/Found/04Message.pdf

A diagram of what is discussed in given at : http://www.colorado.edu/philosophy/vstenger/Found/042Barmag.pdf

What follows is the discussion that ensued between "me" and "my respondent":

ME :
1) You started out with a more highly ordered state (magnets in contact N-S & S-N) .
2) You opened the window
3) Magnets scattered - so magnets in a more disordered state
i.e. (entropy is increased =information is decreased)
4) Result : yet you say "information in the system has increased by
one bit"(p.10) i.e. (entropy is decreased= information is increased)

How is this possible ? Why are the conclusions in #3 & #4
different ?

The bits at the end of the experiment tell me there is more
information yet the system is also more disordered hence there is more
entropy which by definition means there is less information. The
contradiction in #3 & #4 below remains, doesn't it ? What am i missing
here ?

MY RESPONDENT:
You are using an intuitive definition of entropy rather than a
quantitative one, assuming the system is less orderly. Since entropy is
negative information, the entropy has decreased and so the system is
more orderly.
==========

My comments and questions for the Physics Help Forum :
-------------------------------------------------------------------------
Why is my respondant calling my interpretation "intuitive" ? Does he mean "wrong" ?
Is the system's final state really "more orderly" as he says ?
Is it possible that his experiment is really not demonstrating what he thinks it demonstrates either in terms of entropy or information theory ?

Many Thanks in advance for all comments received,

Celal Berker
London, England
 
Last edited by a moderator:
Science news on Phys.org
Celal,

I think your respondent is wrong when saying that the entropy has decreased. It has increased because it's the result of random action.

But I doubt that the information has increased with the scattering. Wouldn't there be some way to retrieve the orientation of both magnets when they are on top of each other?

And in general, it is true that entropy in general does decrease information. It may increase the quantity of data, but not necessarily information. Noise on the transmission cables, for example, usually will corrupt the bits. They may still remain as data, but the correct information has decreased as Shannon's law predicts. In your experiment, suppose the magnets fall on the opposite orientation. Isn't that a corruption of data, and hence a decrease in information?

Regards
 
Last edited:
As I already said, I don't know if 4 bits of information result in the final system. They are random and don't make any sense. The original bits are lost by the event, so there is also a resulting loss from 2 bits to 0 bits of useful infromation.
I'm not an expert in Information Theory and it's formulas, but if they relate entropy with any pattern of bits, I begin to doubt them.
 
I was watching a Khan Academy video on entropy called: Reconciling thermodynamic and state definitions of entropy. So in the video it says: Let's say I have a container. And in that container, I have gas particles and they're bouncing around like gas particles tend to do, creating some pressure on the container of a certain volume. And let's say I have n particles. Now, each of these particles could be in x different states. Now, if each of them can be in x different states, how many total...
Thread 'Why work is PdV and not (P+dP)dV in an isothermal process?'
Let's say we have a cylinder of volume V1 with a frictionless movable piston and some gas trapped inside with pressure P1 and temperature T1. On top of the piston lay some small pebbles that add weight and essentially create the pressure P1. Also the system is inside a reservoir of water that keeps its temperature constant at T1. The system is in equilibrium at V1, P1, T1. Now let's say i put another very small pebble on top of the piston (0,00001kg) and after some seconds the system...
Back
Top