Peter Bennett   An Essay concerning Personal Identity     20


 3   A particular type of counterfactual situation is discussed in the

 philosophy of personal identity.  This section aims at describing what

 this situation is.

      According to  Wilkes, thought experiments are characterized

 firstly by being imaginary - a thought experiment cannot be realized

 in this world - and secondly by observing the constraints by which

 actual experiments are constrained.  There is a problem here since

 one constraint on a real experiment is that it is actually carried out.

 To get round this let us say that the second requirement covers

 what is called 'experimental design' rather than the actual execution

 of the experiment.  (The problem of not actually being able to do

 your experiment will be discussed shortly.)  One constraint which will

 be important to my argument is that thought experiments, just as

 much as real experiments, always have an aim of confirming or

 rejecting some hypothesis.  Typically there is something you control -

 the dependent variable - and something you measure - the independent

 variable.  One way for real experiments to go wrong is for an

 uncontrolled variable to creep in.  This claim is often made against a

 thought experiment.  For example someone who  believes  that

 continuity of character is important for personal identity might argue

 against a brain transplant argument by saying that the individuals

 coming out of a brain transplant must have characters radically

 different to those of the individuals going in.  Anybody who tries to

 ask 'Which is Jack and which is Jill?' after the experiment, which

 has not been controlled for character, has to deal with the fact

 that as far as the character-theorist is concerned, the answer

 is neither.

      Bernard Williams's solution to the problem posed by TE2 and

 TE3 is similar.  In TE1 Locke was very quick to conclude that

 psychological continuity was what was important for personal

 identity.  (The next chapter deals with Locke in greater detail and

 precision.)  But we could say that he failed to control for the fact

 that bodily continuity might be part of _




 Peter Bennett   An Essay concerning Personal Identity     21


 what is involved.  TE2 and TE3 bring this out.  Williams has

 concluded elsewhere11 that bodily continuity must be part of what is

 involved in personal identity.  Such a conclusion could be used to

 solve the problem presented by TE2 and TE3.  But this is not the

 only conclusion that one can draw from these thought experiments.  I

 shall eventually argue something different.

      But this chapter is concerned with the methodology of thought

 experiments, and as we have just seen the first question to ask

 about a thought experiment is:


      Q1:  Does the putative thought experiment satisfy the
           constraints required of real experiments?



      There are two kinds of 'experiment' which Wilkes excludes from

 further discussion.  The first are those which 'take place in thought'.

 Wilkes gives the example of mentally parsing the sentence 'colourless

 green thoughts sleep furiously' to see if it is grammatically correct.

 One could also put Descartes's cogito  and  William  James's

 introspection under this heading.  Such activities have peculiar

 problems of their own, but like Wilkes, I want to exclude them from

 further discussion, since these cases are not like the things which go

 on in the philosophical discussion of personal identity.  I shall

 therefore record the exclusion of this kind of thought experiment as

 our second question to ask about thought experiments:

      Q2:  Does this thought  experiment  merely  take  place
           in thought?


 _

 11   'Personal identity and individuation', also in Problems of the
      Self, pp. 1-18.  The purpose of 'The Self and the Future' is
      more to present a puzzle than to offer a solution.  My solution
      will be offered in the conclusion to this thesis.




 Peter Bennett   An Essay concerning Personal Identity     22


      The second kind of thought experiment which Wilkes excludes

 from the discussion are those which are 'merely imagined', i.e. could

 have been done but were not.  Wilkes gives the example that it is

 likely that Galileo merely imagined what would happen if he dropped a

 cannon ball and a musket ball simultaneously, rather than actually

 doing it.  If this is the case then Galileo might be accused of sloppy

 science for not actually trying it out, but the imaginary scenario is

 not a thought experiment in the  sense  which  Wilkes  wants

 to discuss.12

      If a thought experiment is not 'merely imagined' then there is at

 least a doubt as to whether the hypothetical situation could come

 about in this world.  I shall call such impossible thought experiments,

 which do not merely 'take place in thought' and are not 'merely

 imagined', fantastic thought experiments.  This terminology, if mildly

 abusive, is used in the same kind of spirit as is talk of 'mad

 scientists'.  The terminology will alert us to the fact that we are

 moving outside the circle of experience, and so 'can be sure of not

 being contradicted by experience'.13

      But already there may be cause for caution, since in real

 experiments it often happens that things do not work out in quite the

 way that was expected.  Indeed there would be little point in doing

 the experiment if we could be absolutely certain of the outcome.

 There may be unforeseen deep impossibilities in the experiment.  (I

 shall say what a deep impossibility is shortly.)  So the third question

 to ask about a thought experiment is:


      Q3:  Is the thought experiment merely imagined?

 _


 12   For more details about Galileo's thought experiment and further
      examples of thought experiments from physics, see J. R. Brown,
      'Thought experiments since the Scientific Revolution'.  Wilkes
      mentions Galileo's thought experiment on p. 3 of Real People.

 13   I. Kant, Critique of Pure Reason A4/B8.




 Peter Bennett   An Essay concerning Personal Identity     23


      The fourth question to ask is:


      Q4:  Is the thought experiment merely illustrative?

 Consider:


 TE4:  Travelling at the speed of light


      Einstein, contemplating Maxwell's electrodynamics,
      wondered what someone travelling beside the front of a
      beam of light would see.  No individual could travel at
      this speed - the experiment could not be performed.  But
      if he did, then according to Maxwell's theory, he should
      see something that could not exist:  a  stationary
      oscillatory field.  Hence, from this thought experiment
      alone, there are at least serious difficulties with
      Maxwell's theory.  The imaginary state of affairs is an
      individual travelling at the speed of light; the implication
      drawn, that the theory is inconsistent.14


 Wilkes portrays this as an example of a good scientific thought

 experiment and Parfit describes it as 'worthwhile',15 which in a way

 it is.  But one must not suggest that this thought experiment by

 itself brought about the downfall of Maxwell's electrodynamics.  To

 do so would be to promulgate an implausible myth.  What the thought

 experiment did do was suggest a line of research which turned out

 to be fruitful.  Such thought experiments are useful to science as

 informal reasoning is useful in mathematics.  But at the end of the

 day these informal techniques need to be backed up by rigorous

 methods, and this cannot always be done using thought experiments.

 _


 14   Real People pp. 3-4.

 15   Reasons and Persons p. 219.



next
back
contents

Copyright Peter Bennett 1992