Thought Experiment


  • What is the definition of a thought experiment ?
  • What distinquishes a good thought experiment from a bad thought experiment ?

  • Background

    IMO a thought experiment should be like a blue print of a real experiment.


    This Thought Experiment consists of a train and and Observers A. Observer A is on the ground, near the track.
    The following figure shows this.
         LS1                             LS2
         FD1 <-------------------------> FD2   ---->v  
               O                     O                     
    The train contains at each side: a Firing Device FD and a lightsource LS (a lamp).
    The Firing Device consists of two contacts: one contact on the ground and one contact on the train. The distance d between the two contacts on the ground and the two contacts on the train is identical, such that when the train (the train contacts) is above the ground contacts both lights are on (simultaneous) and when this is not the case both lamps off.
    Observer A is near FD1, as such that Observer A will see light immediate when the lightsource 1 is turned on and lightsource 2 always later.

    Now we perform two tests:
    First we slowly move the train above the track contacts. Observer A will observe that one light goes on immediate and the second one a little later. Difference dt1 in arriving time is a function of the length of the train (distance of contacts on the train)

    Second the train runs very fast over the track. Again the Observer registers the difference in arriving time dt2. However dt2 is larger as dt1.
    The Observer judges now that length contraction takes place.
    Suppose the train moves from LS1 to lamp LS2
    When LS1 is on LS2 is not on. (because the train is smaller)
    Train has still to move a little (Because it has contracted) Then B goes on (later compared to test 1) and A is off.


    I have some doubts about the validity of such a thought experiment in order to explain something. People could reply that also a different outcomes are possible.
    For example: dt1 = dt2
    Or even (when the distance between the track contacts is d) that dt1 <> d/c
    How do you decide who is right ?

    IMO only Real experiments can explain something i.e. improve our understanding


    In order to give feedback about this page you can use the Thought Experiment feedback form. An e-mail to the author
    is also appreciated.

    Last modified: 26 Augustus 1999 Back to my home page Contents of This Document