Philosophy of Science

1
Philosophy of Science

• Class 7

2
Admin

• Want to move my office hours tomorrow to
  400-500. If you were already planning to come
  between 530 and 630, let me know.
• Can you e-mail Assignments? Generally, no (class
  too big!). Yes, if you are having printing
  problems- follow up with a hard-copy a.s.a.p (I
  will do a quick compare).
• Test- 30,30 or 20, 40

3
Review of Dates

• March 15- hand in outline/summary. You can use
  the drop box (special d.b. hours)
• March 16- outline/summary pick up (office hours),
  
• March 22- Final Test, outline/summary pick up
• March 29- Hand in Rough Draft
• April 2 (Friday)- pick up Rough Drafts (office)
• April 5, pick up Rough Drafts (class)
• April 8th- Final Draft Due- drop box

4
Quick Review of Last Class
5
Good theories have

• Testability in principle
• Explanatory power
• Deepens and Broadens understanding
• Predicts new phenomena
• (Used currently by scientists to accept or reject
  theories)

6
August Comte and Positivism
(1798-1857)
7
A Grown Up Science

• Science is limited to the observable
• and therefore
• Science should stick with what it is good at-
  observation and generalization!
• (You can probably see that Comte was a big fan of
  Bacon and Hume)

8
The Logical Positivists
9
Logical Positivists

• Do theoretical entities really exist?
• Vs
• What do theoretical terms mean?

10
Verificationism

• the meaning of any statement is its method of
  verification
• My blue ball bounces
• If I drop this blue ball on the floor it will
  rebound back upwards once it has hit the floor

11
Operationalism

• Just a further refinement of verificationism
• X is p means if O is performed on x then R
• The blue ball is bouncy
• The blue ball is bouncy means if I throw the blue
  ball at the floor from height y and with motion x
  and observe its trajectory then I will observer
  that it will come immediately back up off the
  floor to a height z

12
The Ceteris Paribus Problem

• (All things being equal)

13
The problem of material implication

• (the ifthen problem)

14
Truth Tables
I like dogs I like cats I like dogs and I like cats
True True True
True False False
False True False
False False False
15
Truth Tables
Statement A Statement B Statement A and Statement B
True True True
True False False
False True False
False False False
16
Happy Logicians

• Generalized from actual examples of English
• But then defined to be true
• A nice precise system.
• Or is it?

17
If then statement
Statement A Statement B If Statement A then Statement B
True True ?
True False ?
False True ?
False False ?
18
It is raining I (will) take my umbrella If it is raining then I (will) take my umbrella
True True ?
True False ?
False True ?
False False ?
Lets check our intuitions with a poll (first,
write down what you think)
19
More intuition checking
Owls eat mice Ice Freezes If Owls eat mice then ice freezes
True True ?
True False ?
False True ?
False False ?
20
Some Problems

• Particular instances dont seem adequate to
  completely verify the if then statement.
• It seems strange to put unrelated sentences in an
  if then statement and give it causal power.
• The last two lines of the truth table seem very
  ad hoc and unintuitive

21
The logicians must make a decision
(Studies in Logic,. Charles Pierce,
1883,Tractatus Logico-Philosophicus-
Wittgenstein, 1921)
Statement A Statement B If Statement A then Statement B
True True True
True False False
False True True
False False True
Remember- its a definition!
22
Side note

• In case you think this is some strange
  philosophical move that demonstrates the
  silliness of philosophers
• Mathematicians make moves like this in math, too
  (e.g. negative numbers)
• It is a problematic consequence of wanting to
  have consistent, rigorous systems.

23
x is fragile If (x is dropped on the floor)
then (x will be destroyed) Problem The fragile
sock problem A dog grabs the sock before it can
hit the floor- and eats it
24
Carnaps Correspondence Rule Patch

• Carnap (1891-1970)
• Carnaps patch iff
• If O is performed on x, then (x is p iff R)
• If x is brought near the leaf microscope at time
  t) then (x is (electrically charged) iff (the
  leaves of the electroscope diverge).
• (Note iff if and only if)

25
Pros and Cons

• Removes the problem of Material Implication- Yay!
  Definitions work again.
• The term of interest, p, is now buried
• If O is performed on x, then (x is p iff R)

26
Are these really serious problems?

• Remember
• Philosophers want their theories to work with
  logic all the time- they must be foolproof!
• Philosophy isnt about ignoring problems because
  they are too small
• The logical positivists wanted their theory to be
  solid and nit-pick free.

27
Logical Positivism- still popular!

• Why? Well
• Can you think of any alternatives to trying to
  define theoretical terms?
• (Maybe Atomism vs. HolismHempels network theory
  of meaning)

28
Review- Some possible test questions

• Hume, Experiments
• What was Humes argument? Why do you think
  philosophers took it so seriously? Why Was Humes
  argument relevant to science? Define causation.
  What role does it play in science? How have
  scientists responded to Humes problem?
• Why did scientists devise experiments? What are
  the important components of experiments? What do
  experimental methods give science? Explain how
  problems like Humes led to experiments. List
  four ways in which induction can go wrong How do
  experiments prevent science from falling into
  these traps?
• Devise an experiment. Discuss how your experiment
  is or is not vulnerable to Humes argument. As a
  scientist, what do you think you would do about
  this?

29
Review, Cont.

• Theories, Comte, Logical Positivists
• Consider the following theory. Explain why it is
  a good or bad theory (Hungs list) .
• Why was Comte studying the history of science?
  How did Comte describe the theoretical stage of
  science? What was the connection between Comte
  and the Logical Positivists? What is you opinion
  of Comtes Theory
• List the three stages that the logical
  positivists definitions went through.Why did
  they change? Create a definition for Eggs are
  fragile for each of these. What are some problems
  you might run into applying these definitions?
• What were some issues that the Logical
  Positivists faced? How did they deal with these
  issues? Where do the Logical Positivists stand
  today?

30
Current state of philosophy of science (and
science)

• Modern way to define theoretical terms
• Provided by the Logical Positivists
• Still used- e.g. psychology, operationalization
• Memoryperformance on flash card test

• Modern method for doing research
• How are scientists doing science these days?
• Popper!

31
Karl Popper1902-1994

• Popper wanted to know
• What is good science?
• How should we decide what is and what is not
  science?

32
To put it another way

• When should something be considered scientific?
• Can science find out the truth about everything?
• What is the scope of science?
• Should science be able to study everything?
• Should some topics be excluded from science?
• What does it mean if something cant be studied
  by science? If not, why not?
• When is an activity scientific?
• When is scientific activity good science? When
  is it bad science?
• How should science be done?
• What should we expect from science?

33
Some attempts to answer these questions

• Bacon Science should use my method. My method
  results in good science (because my method is
  objective)
• Comte Science should restrict itself to making
  generalizations from observations. Anything
  outside of this is outside bad science.
• Logical Positivists Science can study everything
  that is worth studying. Anything that science
  cant define through verification is meaningless.

34
Poppers Concerns

• Increasing strangeness of scientific theories
  (e.g. quantum physics)
• Dubious credentials of some theories claiming to
  be scientific (e.g. Freudian psychoanalysis,
  Marxism)

35
Why did Popper Care?

• Science is supposed to be about the truth (about
  objective reality)!
• Bad science is less likely to find the truth.
• Science has a certain status that could be
  misused.

36
Poppers proposed argument

• All good science is like x (i.e. has the property
  x)
• Your theory or story or explanation is not like x
• ---------
• Your theory is not good science!

37
Poppers Proposed Argument, more detailed

• All good science follows methodology x
• Your science does not follow methodology x
• ------
• Your science is not good science.

38
Popper and Induction
INDUCTION
39
The challenge

• Getting rid of induction!
• Aristotle said it could not be done!
• The Medieval Natural Philosophers couldnt do it!
• Hume couldnt do it!
• Can Popper do it?

40
Poppers Bright Idea

• Swan x is white
• Swan y is white
• Swan z is white
• -----------
• All swans are white.

It seems fairly easy to determine the falsity of
this statement
41
Poppers Method

• 1. Find a hypothesis (from somewhere,
  somehow).This hypothesis must be falsifiable.
  (Note theory)
• 2. Using deductive logic, derive an empirical
  implication that proves the hypothesis to be
  false ( like indirect tests)
• 3. Test the implication in some way-to see if it
  (the implication, not the hypothesis) is true or
  false.
• 4. a)If the test shows that the implication is
  true, discard the hypothesis and go to 1.
• b) If the test does shows the implication to be
  false, go back to step 2.
• Popper called this the hypothetico-deductive
  method.

42
Seems like backward science!

• Example
• Hypothesis Apples fall to the ground
• If it is false that apples fall to the ground
• Implication We should expect to see an apple
  hover in the air when released.
• If we carry out the test, the implication is
  falsified.
• Therefore we go back to step two and derive
  another implication

43
Some important features of Poppers method

• it makes falsification the main activity of
  science (black swan detector)
• the more tests a hypothesis passes, the more
  corroborated it is.
• science should progress by accumulating
  corroborated theories (i.e. hypotheses), while
  always being on the lookout for tests that might
  falsify these statements.

44
What counts as science

• All good science follows methodology x
• Your science does not follow methodology x
• ------
• Your science is not good science.
• Methodological failure If there appears to be no
  possible test that would prove the statement
  false, then that statement cannot be investigated
  by science.

45
Consider the following statements

• I like bananas or I dont like bananas.
• Apples are a healthy snack.
• The acceleration of gravity is 9.8 m/s2.
• God is good.
• The id controls your subconscious behaviours.

46
Metaphysics

• Popper describes theories that cannot be
  investigated by science as being metaphysical.
• Popper does not judge metaphysical theories. He
  does not, for example, say that they are morally
  wrong, or meaningless.
• Popper simply says that metaphysical theories
  cannot be evaluated by scientific criteria.
• Some other method of evaluation would need to be
  found if they were to be evaluated.

47
Group Discussion

• If you were in charge of training a group of new
  scientists, would you pick Bacons method or
  Poppers ( or do something different)?
• On the board
• Which would you pick (or what would you do)?
• One sentence stating why.

48
Popper and the Logical Positivists

• Different
• Popper is interested in falsification, while the
  logical positivists theory depends on the idea
  of verification
• logical positivism is concerned with meaning.
  Poppers theory does not involve philosophy of
  language or the meaning of statements.
• Popper was much less critical of metaphysical
  theories than logical positivists

49
Popper and the Logical Positivists

• The same
• Popper did think that observational statements
  were important, and could provide the basis for
  scientific theories.
• Popper wasnt too concerned with the discovery of
  hypotheses. What was important was how hypotheses
  were (or were not) justified.
• Both Popper and the Logical Positivists took Hume
  seriously.

50
Problems with Popper?

• 1. Find a hypothesis (from somewhere,
  somehow).This hypothesis must be falsifiable.
  (Note theory)
• 2. Using deductive logic, derive an empirical
  implication that proves the hypothesis to be
  false ( like indirect tests)
• 3. Test the implication in some way-to see if it
  (the implication, not the hypothesis) is true or
  false.
• 4. a)If the test shows that the implication is
  true, discard the hypothesis and go to 1.
• b) If the test does shows the implication to be
  false, go back to step 2.

51
Two problems

• The infinite loop problem
• If the test does not show the implication to be
  false, go to step 2.
• The inane hypothesis problem
• My profs name is Jen

52
Dealing with the inane hypothesis problem

• Science should only investigate the truth value
  of sentences with high empirical content.
• According to Popper the more likely a statement
  is to be false, the higher its empirical content.
• Consider
• Its going to rain or be sunny or snow tomorrow
• Its going to hail for five hours tomorrow.

53
Empirical Content

• Popper made the intuitive concept of empirical
  content more rigorous and quantitative by
  introducing the idea of potential falsifiers.
• Potential falsifiers are observation statements
  that, if true, makes the statement false.
• The falsifiablity of a statement can be judged by
  its number of potential falsifiers.
• The more potential falsifiers a statement has,
  the higher its empirical content.
• (parallel with the logical positivists)

54
Dealing with the infinite loop problem

• Hypothesis Corroboration
• Consider hypothesis A
• Potential Falsifier 1
• Potential Falsifier 2
• ...
• Potential Falsifier n
• The more potential falsifiers are discarded, the
  more the hypothesis itself is corroborated
• Also- more or less severe tests

55
Trusted Hypotheses

• Eventually, the hypothesis will have passed so
  many tests that it will be accepted by the
  scientific community.
• They will treat the hypothesis as de facto true
  and move on to test other hypotheses.
• Even so, scientists must always keep in mind that
  it has never been proven to be true.

56
Add ons to Poppers Method

• Falsifier An observation statement, derived
  through deductive logic, that can prove a
  hypothesis false.
• Empirical Content Hypotheses with high empirical
  content are better than ones without it
• Potential Falsifiers The more potential
  falsifiers a hypothesis has, the higher the
  empirical content
• Corroboration Is like truth, but Popper doesn't
  want to say it is truth. Depends on the number
  and severity of the tests a hypothesis has been
  through

57
Compare Poppers Method, Inductive Methods (e.g.
Bacon)

• Need to make sure inductive logic is not hiding
  in Poppers method!
• Quick Group Question Do you think there is any
  inductive logic hiding in Poppers method?
• (Ill take a poll after the discussion)

58
Reviewing a few concepts

• Truth and Falsity Poppers theory rests on the
  idea that they are different.
• Consider
• This chair is orange.
• Some chairs are orange.
• Some chairs are not orange.
• No chairs are orange.
• All chairs are not orange.
• All chairs are orange.
• How would you prove these true? false?

59
Reviewing a few concepts

• Induction
• possibly true conclusions
• way of increasing the truth probability of a
  statement is to observe multiple instances that
  confirm (but do not prove) the truth of the
  statement
• Completely separate from induction is the issue
  of indirect testing (auxiliary hypotheses)

60
Reviewing a few concepts

• Indirect testing
• Indirect testing involves deriving (deductively
  valid) implications of the hypothesis and testing
  these implications.
• If the implied statement is true, then we can
  (deductively) conclude that the hypothesis is
  true.
• But the problem of auxiliary hypotheses means
  that we can never say for sure that the statement
  is true- We can never state all of the premises
  in our deductive argument.

61
Corroboration and Induction

• some similarities between Poppers scientific
  method and the Bacons
• Both employ the idea of deducing observable,
  testable consequences from statements.
• In some types of induction (e.g. universal
  generalization) multiple observations increase
  the probability that a statement is true.
• In Poppers method, there are multiple tests of
  (hypothesis, deduced potential falsifier) pairs
  (involving observation of test results.

62
Key Differences

• Observing the falsity of potential falsifier
  means that the hypothesis is not false
  (deductively).
• Inductively, finding observations consistent with
  the truth of the statement suggests that the
  statement is true.
• Popper did not want to equate corroboration with
  truth. (not false DOES NOT EQUAL is true)
• He intended corroboration as a way to choose
  between competing hypotheses, not show that a
  hypothesis is true.
• This is a fine line to walk

63
Value of Popper

• The idea that non-falsifiable statements cant be
  studied scientifically
• The distinction between truth and falsity (they
  are not exactly equivalent)
• A scientific method which at least seems less
  reliant on induction... maybe...
• At the very least, a scientific method which has
  been generally adopted by science

64
Review of Popper

• Some sample test questions
• Define Empirical Content, Severe Test,
  Corroboration. Give examples illustrating these
  concepts. What motivated Popper to invent his
  method? Was Popper a logical positivist? In what
  ways did his theory differ? In what ways was it
  the same?
• Pick a hypotheses and create three potential
  falsifiers. Which of these is a more severe test?
  
• Based on what you know of Bacons Method, how
  might Bacon criticize Popper?

65
Classic Tradition

• The 1950s and 1960s marked the high point of
  what is commonly referred to as the classical
  tradition in science.
• What is this classical tradition?
• It is a certain basic view of the world and of
  science which has been shared, to some extent by
  all of the philosophers of science we have
  considered so far.

66
Assumptions of the Classical Tradition

• There is an outside world, which exists
  independent from all observers.
• The ultimate goal of science is an accurate (as
  possible) description of this observer
  independent world.
• Scientists can observe the objective natural
  world, in some way, and learn about it.
• the universe has underlying regularities that,
  if discovered can explain the behaviour of the
  universe.

67
General Agreement that

• There is independent observation (if you had
  twenty scientists observing something, they would
  see it about the same).
• Observation is not dependent on the particular
  observer, and deductive logic is not dependent on
  the observer.
• Reality is also independent of any particular
  observer.

68
Requirements of Science

• The question is
• if you lose any of these assumptions, is it still
  possible to even do science?
• In other words, can you be a scientist if you
  dont believe in
• An objective reality
• A reality with causal universal laws
• An objective observer (or collectively, a group
  of objective observers) who are able to draw
  objective conclusions from objective observations
  of reality.

69
Two branches of Twentieth Century Philosophy of
Science

• Branch One Logical Positivists, Popper
• Branch Two Duhem-Quine Thesis, Kuhn
• Next Class Talking about Branch Two
• Core ideas Meaning is subjective, perception is
  subjective, science is subjective.