Confirmation Flashcards
basic idea bayesianism
S beliefs that p with degree g between 0 & 1 equal to some probability function
Three pillars of Bayesianism
- quantitative conception of belief
- rational beliefs at any fixed given moment in time
- how beliefs at different moments in time ought to be related
credence g=0
total disbelief, contradictions (absurdum)
credence g=1
certain belief, tautology (veritas)
probability function of g
probability function is always in between absurdum & veritas, so between 0 & 1
incompatibility of p & q
Pr(p or q) = pr (p) + pr (q)
or Pr (p and q) = 0
bayesian explanation of propositional knowledge
p is true
s believes that p to a degree of 0.90 or larger
s has reached 0.90 by bayesian updating
confirmation
confirmation (e,h) iff e is empirical evidence & f (e,h) > 1
Infirmation
infirmation (e,h) iff e is empirical evidence & f (e,h) < 1
Irrelevance
irrelevant (e,h) iff e is empirical evidence and f (e,h) = 1
verification
maximal confirmation
Pr1(h) = Pr0 (h|e) = 1
Falsification
maximal infirmation
Pr1 (h) = Pr0 (h|e) = 0
mu (e,h)
quantify the degree of confirmation or disconfirmation that evidence e provides for a hypothesis. 10 for verification, -10 for falsification, irrelevance = 0.
Diachronic coherence
change of beliefs over time, quantified by belief-change factor f(e,h)
Hempel’s raven paradox
things that confirm should not confirm and things that do not confirm should confirm. Therefore, background knowledge is needed.
Example: seeing colored things that are not ravens confirms hypothesis that all ravens are black.
Goodman’s new riddle of induction
even if we are able to see all emeralds in the world are green, we can still not promote hgreen to knowledge, since our observations also are able to confirm ‘magical’ hypotheses, like all emeralds are grue.
‘solution’: magical hypothesis will most likely not reach the epistemic level of >0.90 by bayesian updating