PHILOSOPHY OF SCIENCE AND ITS RELEVANCE TO RESEARCH DESIGN - PTJ American University - kittenboo.com
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
PHILOSOPHY OF SCIENCE AND ITS
RELEVANCE TO RESEARCH DESIGN
PTJ • American University
2021 • lecture two
WAGERS
provisional position on likely unresolvable issue
philosophical ontology: pertaining to mind-world hook-u
more fundamental than “epistemology
arti cially—instrumentally—dichotomized for clarit
goal is a more adequate lexicon
fi
”
s
y
p
DUALISM AND MONISM
dualism: mind separate from world, valid knowledge mirrors the
worl
monism: mind continuous with world, knowledge a perspectival
disclosing of the worl
dualists like testing; monists like explication
d
dLOGICAL POSITIVISM
unveri able statements are nonsense
veri able = observable implication
empirical evidence determines whether a statement is tru
logical form of the statement gives truth conditions
purest logical form can be probabilistic
fi
fi
s
eKARL POPPER
doctorate in psychology
(University of Vienna, 1928
Logic der Forschung (1934
The Open Society and its Enemies
(1945
philosophy not just about dis-
solving linguistic puzzles
)
)
)FALSIFIABILITY
inverts the logical positivist positio
all knowledge is conjectural; none is certai
empirical testing should be continua
metaphysics as a source of hypothese
observable implications remain central
n
l
s
nTHE PRACTICE OF SCIENCE
history of science is not a linear stor
falsi cation doesn’t always drive change
now-accepted theories often start out “falsi ed
shifts in background assumptions and technique
falsi cation thus falsi ed by the history of scienti c practice
…even though scienti c practice is successful
fi
fi
fi
fi
y
s
fi
”
s
fi
…KUHN AND LAKATOS
Kuhn: discontinuous jumps
Lakatos: …but retrospective
rational reconstruction shows
progres
both skeptical about the scienti c
status of the social sciences!
s
…
fiNEOPOSITIVISM
combination of falsi cation with the emphasis on precise logical
for
a preference for numbers as making precision easier to attai
testing of hypothetical general laws as the basic procedure
m
fi
nNEOPOSITIVIST COMPARISON
X1 X2 X3 Y
C1 yes 2.7 a yes
C2 yes 1.8 b yes
C3 yes 3.9 b no
C4 yes 2.7 aREGRESSION
X Variable 1 Line Fit Plot
7
6
5
4
Y
Y
Predicted Y
3 Linear (Predicted Y)
2
1
0
5.2 5.4 5.6 5.8 6 6.2 6.4
X Variable 1CORRELATION AND CAUSATION
neopositivists say that these aren’t the same
…but they have no alternative to correlation as the mark of
causatio
de nition of neopositivist “cause”
X ➜ Y (suf cient); ~X ➜ ~Y (necessary
“with probability p” does not make a difference
fi
n
fi
:
)
…CAUSAL POWERS
Rom Harré, student of J. L. Austi
not statistical tendencies, but
deeper proclivitie
can manifest as statistical
tendencies in particular
environments (laboratories)
s
nBEYOND THE EMPIRICAL
perception not exhaustiv
“unobservable” too imprecis
things an observer hasn’t experienced or perceive
things no one has ever experienced or perceived, at least not ye
things no one could experience or perceive, even in principle
e
e
d
tTYPES OF “UNOBSERVABLE”
I haven’t
unobserved Angor Wat go see it
observed
we can’t Neptune;
undectected build detector
observe (yet) Higgs boson
we can’t single quarks;
observe (in undetectable social ?
principle) structureTHEORETICAL OBJECTS
known by what they d
abductive inference from observed outcome
have to be isolated and vetted
in a laborator
via transcendental argument
y
o
:
sCAUSAL MECHANISMS
linked series of occurrences that
unfold in a similar way on
different occasion
need not always yield the same
observable outcome
“actually present
brokerage, or balancing
”
s
sREALIST EXPLANATIONS
causal powers/mechanisms in an open syste
can’t correlate using real-world dat
can’t isolate, except in a lab or conceptually…but have to be vetted
someho
show how they interact in a speci c cas
INUS condition
“complete” explanations
w
s
fi
a
e
mYou can also read
