PHIL2420 Lecture Notes - Lecture 4: K562 Cells, Experimentum Crucis, Falsifiability
PHIL 2420
Critical Thinking
March 22, 2018
WEEK 4
Falsification and Underdetermination
Demarcation Question
• Model is a particular type of analogy where you’re drawing relationship between
two things
• Draw inference from result of the analogy. There are different kinds of analogies.
• What makes something scientific or unscientific?
• What makes a theory scientific?
o Theory – collection of models that makes a statement about a particular
science
Popper (philosopher for science) – Is there a criterion for the specific character or
status of a theory?
Science and Pseudoscience
• Popper wanted to differentiate between the two
• What makes germ theory of disease different from astrology?
• Problem with pseudo-scientific theories is that they’re too good at explaining
things
Pseudoscience
1. Vagueness
a. Virgos are generally passive, but on occasion can be quite assertive
b. Can apply to anything at all
2. Ad hoc qualification – convoluted explanations of exception to the rule
a. Virgos are generally passive, but your rising sign is Aries. Whilst your
inner nature is passive, your outward expression is assertive…
b. Hour makes a difference, was there daylight savings?
c. Rising sign – fallback
Good Theories
• Predictive power
o If theory has power to make predictions, what about astrology?
o Auxiliary assumptions VS Ad hoc hypothesis
o AA – have to make to move from a theory to a test of that theory
• Mechanism
• Fruitfulness
o Does it lead to a research with positive yield?
• Coherence
find more resources at oneclass.com
find more resources at oneclass.com
o Internal – logically consistent theory
o External – compatible with other theories known to be true
▪ There could be theory in bio that’s incompatible with something with
physics. Most probably go for physics one. But it’s implications in
terms of biology doesn’t make it true.
• Simplicity – simpler theory is more correct
o Occam’s Razor
▪ 12th-13th century philosopher
▪ Simple explanation is better than a complex one
• Problem with ad hoc – keep tacking on extra rules to deal
with exceptions to the rules, theory becomes very
complicated
▪ Data must be equal to apply this theory
General Laws
• Science proceeds by the discovery of general laws
• EXAMPLE: If you have a deficit of serotonin, you will feel depressed
Confirmation – gathering evidence to support claim
• We build our case through gathering evidence that supports our claim
• Induction, confirmation, verification, corroboration
o We use these to support our theories
o More and more evidence to prove hypothesis they think is true
Problem of Induction
• No amount of data can confirm a (general) hypothesis
• General statement says something about everything / whole class of things
• EXAMPLE: Turkey
o Fed nicely, treated well
o Data is consistent with farmer being good guy, looking after turkey
o Then suddenly, completely inconsistent, farmer kills turkey on Christmas
• Generalizing from finite amount of data can be dangerous
• Unless you know underlying causal structure of situation (farmer eats turkey on
Christmas), all observations may point you in wrong direction
Falsification
• Popper – don’t generalize, try to falsify
• Even if you need infinitely many instances to CONFIRM a hypothesis, one
instance can prove a hypothesis FALSE
• Provides deductive solution to the problem of induction
• Formal Justification for Falsification
o EXAMPLE: Swans
▪ Inductive generalization
• A is a black swan
• B is a black swan
find more resources at oneclass.com
find more resources at oneclass.com