Scientific Methodology PH351F/PH451F

1
Scientific Methodology PH351F/PH451F

• Luca Moretti
• Department of Philosophy
• Aberdeen University
• l.moretti_at_abdn.ac.uk
• http//lucamoretti.org

2
Lecture 7Lakatos methodology I

• Requested readings
• I. Lakatos, Falsificationism and the methodology
  of scientific research programmes in his
  Philosophical Papers, vol 1, pp. 31-52, pp.
  68-73, and pp. 86-90.

3
7.1 Overview (what we do in these two lectures)

• I will first outline Lakatos methodology of
  scientific research programs.
• Then, I will apply this new methodology to the
  historical case of the Copernican revolution.
• I will briefly describe the Ptolemaic research
  program and the Copernican research program.
• We will learn why, according to Lakatos, the
  second became rationally preferable to the first
  in the first half of the XVII century.
• Although Lakatos explanation is illuminating, it
  is not completely convincing.
• The problem of explaining the Copernican
  revolution from the point of view of scientific
  methodology remains partly open.
• Some of the problems depend on Feyerabends sharp
  criticism of Lakatos view.

4
7.2 Lakatos project

• Lakatos aims essentially to provide a
  reformulation of falsificationism capable to
  lessen the subjective and arbitrary elements of
  Poppers methodology, which were introduced to
  cope with the Duhem-Quine thesis.
• For Popper, a theory can be considered falsified
  and thus eliminated on the grounds of
  incompatible observations when the scientific
  community unanimously stipulates (?) that the
  (well corroborated) auxiliary assumptions
  necessary for the falsification are true.
• Since any well corroborated auxiliary assumption
  might prove false, true theories might be
  eliminated!
• To avoid methodological anarchism (i.e. the view
  that no theory can ever be eliminated), Poppers
  falsificationism would seem to turn science into
  an arbitrary enterprise.
• Lakatos intends to re-formulate falsificationism
  to the effect that the elimination of theories
  does not rest on subjective decisions but is
  empirically justified.
• This requires focusing, not on theories, but on
  research programs conceived of as dynamic
  sequences or progressions of theories.

5
7.3 The decisions necessary to falsify a theory
for the falsificationist

• On Poppers falsificationism (at least in
  Lakatos interpretation of it), scientists can
  falsify a theory on the grounds of observations
  only after taking certain methodological
  decisions.
• There are three types of decisions (or
  assumptions)
• (i) Decisions about which type of statements are
  observation statements
• As the range of what is observable has vague
  boundaries (something observable only by
  telescope is really observable?), a specific
  class of statements should be assumed to be the
  one of the observation statements.
• (ii) Decisions about which asserted observation
  statements are true
• This requires deciding that the tools (e.g.
  telescopes) used to make the relevant
  observations are reliable tools.
• This in turn requires deciding that the theories
  about these tools are correct (e.g. optics).
  These theories must be corroborated by
  independent tests.
• This also involves deciding that the conditions
  of correct functioning of the relevant
  observation instruments (e.g. absence of high
  temperature) are satisfied in the specific cases.
  These conditions must have been repeatedly
  checked.
• Finally, the accepted observations must have been
  repeated several times.

6
7.4 The decisions necessary to falsify a theory
for the falsificationist

• (iii) Decisions about which theories are
  falsified by true observational statements
• Suppose T (e.g. a theory of comets) entails O
  (e.g. The comet C will be seen in x at time t).
  Suppose Not-O is true. in this case O is false!
• T entails O only in conjunction with auxiliary
  assumptions specifying at least
• The initial conditions (A1) (e.g. the location
  of C in the past).
• That there is no disturbing factor (A2) (e.g.
  no body close to C that disturbs C motion).
• That the observation tools are reliable (A3)
  (e.g. telescope is reliable).
• To conclude that the falsification of O, entailed
  by T A1 A2 A3 does falsify just T,
  scientists have to assume that A1, A2 and A3 are
  all true.
• A3 has been already considered while speaking
  of decisions of type (ii).
•   Scientists will assume that A1 is true after
  repeated tests of it.
•   Scientists will assume that A2 is true after
  testing many conjectures about possible
  influencing factors.

7
7.5 Lakatos objections to falsificationism

• According to Lakatos, falsificationism is
  affected by two general problems
• (a) Falsificationism asks scientists to take
  decisions that may prove misleading.
• Popper seems to think of that these decisions
  are irreversible.
• Popper wants to secure the possibility of
  rejecting some theories to
• allow for the possibility of scientific
  progress, but the risk is rejecting
• the wrong theories!
• (b) Falsificationism does not reflect what
  scientists really do.
• In particular
• The history of science shows that the
  methodological decisions of
• type (i), (ii) and (iii) are continuously
  questioned and re-discussed in time.
• Real scientists are much slower and more prudent
  than Popper believes.
• Furthermore, real science is more complex than
  in Poppers picture because scientific disputes
  typically involve not one theory but two rival
  theories.
• For instance
• Ptolemys theory versus Copernicus theory,
• Cartesian mechanics versus Newtonian mechanics,

8
7.6 How Falsificationism should be reformed
according to Lakatos

• Lakatos proposes the following rules
• (Theory acceptance) A new theory T should be
  accepted if and only if (a) the unrefuted (but
  not the refuted) predictions of its immediate
  predecessor T are derivable from T, (b) T
  makes new predictions with respect to T, (c) some
  of these new predictions are verified.
• (Theory rejection) A theory T should be rejected
  if and only if there is new theory T such that
  (a) the unrefuted (but not the refuted)
  predictions of T are derivable from T, (b) T
  makes new predictions with respect to T, (c) some
  of these new predictions are verified.
• Note that the falsification of T is neither
  sufficient nor necessary for the rejection of T.
• It is not sufficient because T cannot be rejected
  on the grounds of incompatible observation
  statements if no better theory T is available.
• It is not necessary because T can be rejected
  just because there is a new theory T with
  additional and verified predictions.
• Note also that the new theory T can be accepted
  even if some of its new predictions prove false.

9
7.7 How falsificationism should be reformed
according to Lakatos

• To lessen the arbitrary or subjective component
  of falsificationism and to explain how
  methodological decisions are continuously
  re-discussed by scientists, the rules of
  acceptance and rejection should be applied to
  theories in conjunction with all their auxiliary
  assumptions.
• Let us call the theories of this type theoretical
  systems or just systems.
• Consider a series of systems S1, S2, S3, of
  this type, and suppose that each system results
  from changing only the auxiliary assumptions of
  the former system in order to accommodate its
  empirical anomalies.
• A series of this type is theoretically
  progressive if each system in the series makes
  new predictions.
• A theoretically progressive series is also
  empirically progressive if, for each system in
  the series, some of the new predictions are
  verified.
• A series of systems that is both theoretically
  progressive and (at least intermittently)
  empirically progressive is said progressive,
  otherwise it is said degenerating.
• Note that if a series of systems is progressive,
  it has been constructed in perfect accordance
  with Lakatos rules of theory acceptance and
  rejection (applied to systems).

10
7.8 How falsificationism should be reformed
according to Lakatos

• Lakatos suggests that, to cope with
  falsifications, scientists should construct (and
  that they do construct) a series of systems of
  this kind in which the auxiliary assumptions are
  continuously modified with the purpose of making
  the series progressive.
• (Note that this seems to correspond to what Kuhn
  calls puzzle solving activity of scientists,
  which characterises normal science).
• For instance, consider a system S equivalent to
  the conjunction T A1 A2 A3, where A1, A2
  and A3 are the auxiliary assumptions. Suppose
  that S entails the observational statement O,
  which turns out to be false.
• In this case, S is false too.
• To cope with the falsification of S, the
  scientists will modify A1, A2 or A3 to prevent
  the deduction of O, but they will try not to
  touch the theory T, central to the whole system.
• Modifications of A1, A2 or A3 are tolerable as
  non-ad hoc if they make the series of
  succeeding systems at least theoretically
  progressive.
• Any series of systems which is not at least
  theoretically progressive should be considered
  pseudoscientific.

11
7.9 How this makes falsificationism less arbitrary

• If falsificationism is reformulated in this
  fashion, the only arbitrary decisions that
  scientists have to take concern the
  identification of the class of the observation
  statements.
• All other assumptions and decisions of scientists
  will be empirically tested again and again, and
  so (to a good extent) justified, while developing
  the series of systems to cope with
  falsifications.
• Suppose again T (e.g. a theory of comets) entails
  O (e.g. The comet C will be seen in x at time
  t) in conjunction with the auxiliary assumptions
  specifying
• (A1) The initial conditions. (Earlier locations
  of C).
• (A2) That there is no disturbing factor. (No
  unknown body is close to C).
• (A3) That the observation tools are reliable.
  (Telescope is reliable).
• Suppose O is false.
• To cope with this falsification, scientists will
  replace A1, A2 and A3 in turn with alternative
  assumptions and they will test each new systems
  so obtained.
• Any rejected auxiliary can always be reconsidered
  in combination with new auxiliaries
• The ideal goal is to maintain the sequence of
  systems progressive.

12
7.10 Lakatos demarcation criterion

• While Poppers demarcation criterion focuses on
  single theories (a theory is scientific if and
  only if it could be falsified by observations
  describable with precision in advance), Lakatos
  demarcation criterion focuses on sequences of
  theories or systems
• Any series of systems which is not at least
  theoretically progressive is pseudoscientific.
• One can already guess from this why Lakatos
  thought that the scientists were rationally
  justified in switching from the old Ptolemaic
  view of the heavens to the new Copernican.
• No system of the series
• Eudoxus theory -gt Apollonius theory
  -gt Hipparchus theory -gt Ptolemys theory
  
• probably predicted any new fact.
• The whole series became very soon
  pseudoscientific an enormous sequence of ad-hoc
  adjustments.
• This explains Copernicus feeling that the
  Ptolemaic system was nothing but a (rational)
  monster.

13

• End Lecture 7