Transcript PowerPoint - University of Pittsburgh

Causation as Folk Science
John D. Norton
Center for Philosophy of science
Department of History and Philosophy of Science
University of Pittsburgh
1
Principal Claims
Negative Thesis
Denial of causal fundamentalism, which asserts that
the world is governed by a principle of causality;
based on a dilemma.
Positive Thesis
Sciences restricted to hospitable domains adopt
forms of various folk sciences of causation, which
explains the prevalence and utility of causal notions.
Modern Physics
Causal anti-fundamentalism is compatible with the
widespread presence of causal terms and principles
in modern physics.
2
Guiding Intuitions
Negative
Causation is about the physical
connectedness of things in the
world. Its analysis is the province of
science and not a priori postulation.
Positive
Causal notions are pervasive
because, whatever the domain, we
are plastic enough and inventive
enough to find relations we are
comfortable to label “causal.”
Modern philosophical literature
in causation has lost its way.
It is devoted to finding out
precisely what we mean when we
say C causes E; and mistakes that
for finding out deep truths about
the connectedness of things in
the world.
3
Negative Thesis
Causal Anti-fundamentalism
4
Causal Fundamentalism: the doctrine
Nature is governed by cause and effect;
and the burden of individual sciences is to
find the particular expressions of the general
notion in the realm of their specialized
subject matter.
There is a universally applicable
principle of causality.
Causation is about the connectedness
of things in the world.
Unity of meaning of causal
talk in different domains.
5
Causal fundamentalist's
dilemma
Conforming a science
to cause and effect…
1.
EITHER (first horn)
…places a restriction
on the factual
content of a science
We must find some
restriction that can be
properly applied to all
sciences.
No appropriate
restriction; no enduring
principle of causality.
OR (second horn)
2.
…it does not place a
restriction on the
factual content of a
science.
The imposition of the causal
framework makes no difference
to the factual content of the
sciences.
It is an empty
honorific.
6
First horn
1. How might causation restrict the factual content of a science?
4th c. BC. Aristotle’s four causes.
Material, formal, efficient, final.
17th c. mechanical philosophy. No
place for final causes.
17th c. Newton. Action at a distance?
18th, 19th c. No mechanism found;
no finite velocity measured; no
shielding. Gravity is action at a
distance.
“…so great an absurdity that no man, who has
in philosophical matters a competent faculty of
thinking, can ever fall into it.”
19th century purification of causation.
No agents/patients, continued existence of
cause. Causation is determinism. Causes
guarantee effect.
20th c. Quantum theory. Physics
supplies probabilities only of
effects.
20th c. Principle of common cause
Fails for entangled states in quantum theory
20th c. Probabilistic account of
causation.
Virtually all physical theories
indeterministic. No probabilities for
undetermined outcomes. Even
7
Newtonian physics!
2. What if causation places no factual restriction on science?
Second horn
Any possible science, no matter odd, may be
conformed to causation.
…but it is no way restricted by it.
“Nature is governed by cause and effect;
and the burden of individual sciences is to
find the particular expressions of the
general notion in the realm of their
specialized subject matter.”
…so the burden consists solely in
assigning honorific, causal labels.
Causation as indispensable to
science. (Kant, Nagel, …)
Principle of causality
“…is an analytic consequence of what is
commonly meant by ‘theoretical science.’”
“…maxim for inquiry rather than a
statement with definite empirical content.”
Nagel
…so it does not tell us about the
world, but about our definitions and
or own psychology.
8
Russell on the sun as a cause of gravity.
ABC of Relativity
“The language of cause
and effect adds only a
number of quite
irrelevant imaginings,
connected with will,
muscular tension, and
such matters.”
JDN irrelevant  Pragmatic
ally useful
9
Interventionism
The definition delineates
which relations in the
world can bear the title
“causal.”
There is no assurance
that the world must host
such relations.
It is very convenient
if they do, if our interest
is to manipulate things
in the world.
From J. Woodward, “Causation and Manipulability,” Stanford
Encyclopedia of Philosophy
10
Positive Thesis
Causation as Folk Science
11
Generative capacity of reduction relations
General relativity:
fundamentally
gravity is not a force, but a
manifestation of the
curvature of spacetime.
Statistical physics:
fundamentally
Heat is just a disorderly
distribution of energy.
Gravity behaves just like
the force for which
Newton found powerful
evidence.
restrict to weak fields,
non-cosmic scales
restrict to systems that
do not exchange heat
and work
Gravitational forces, heat as caloric
can be pervasive and useful notions
without being fundamental.
Heat behaves just like the
conserved fluid “caloric”
of Lavoisier and Carnot.
…we require the same
of causal relations.
12
Generating causation
Greater science
not fundamentally
causal
restrict to
vacua
surrounded
by fluids
restrict to ordinary Newtonian
systems of finitely many
degrees of freedom
Determinism
Present state
causes future via
forces
restrict to
dissipative
systems
Vacua have active
powers. They suck.
States of lowest
energy, highest
entropy are final
causes
Sense of causation recovered
in each domain is different.
13
Folk theory of weather
High pressure zones,
low pressure zones,
cyclones,
anticyclones,
warm fronts,
cold fronts,
….
Rough and ready ways to
make the real weather
processes of the motion
of moist air more
intelligible.
14
Modern Physics and
Causal Anti-fundamentalism
15
Is there a Contradiction…?
1. Talk of causal relations and
causal principles permeates the
fundamentals of modern physics;
2. Causal anti-fundamentalists do not
believe that the world is governed by
a principle of causality.
There is not. The causal talk pertains to
contingent features of modern physical theories
that are not a manifestation of a deeper causal
necessity.
16
The causal notions and causality conditions of modern
physics express:
1. The existence of a finite, invariant velocity in spacetime
= spacetime has a light cone structure
2. Propagation of matter in spacetime must conform to the
lightcone structure. “No propagation outside the light
cone.”
These are contingent facts about the world. Physical theories
without them can be quite cogent. They are not metaphysical necessities.
17
Illustrations in Relativity Theory
Special
relativity
Causally connectible = timelike or lightlike connected
General
relativity
“local causality” = Laws governing matter fields are such that there is
no propagation outside the light cone. The condition is formulated more exactly
(+ numerous variant forms)
in terms of a Cauchy problem for the differential equations governing the matter
fields. (p. 60)
Global extensions
“causality condition” holds holds if there are no closed nonspacelike curves. (p. 190)
"The strong causality condition is said to hold at p if every
neighbourhood of p contains a neighborhood of p [sic] which no nonspacelike curve intersects more than once." (p. 192)
"Stable causality condition ... [informally] one can expand the light cones
slightly at everypoint without introducing timelike curves." (p. 198)
S. Hawking and G. F. R. Ellis. The Large Scale Structure of Spacetime.
18
Illustrations in EPR/Quantum Theory
"Locality Principle (L)
Elements of reality pertaining to one system cannot be affected by measurements performed 'at
a distance' on another system.”
"For Bell locality, 'at a distance' means in the absence of causal influences recognized by
current physical theories.”
“For Einstein locality, 'at a distance' means at a space-like separation between the space-time
locations where the element of reality pertaining to one system exists and the measurement on
the other system takes place.”
M. Redhead, Incompleteness, Nonlocality, Realism. p. 75
This principle is not offered a principle of quantum theory,
but as part of an expressed hope that quantum theory will
conform to it. It may fail to, as happened with the common
cause principle.
19
Illustrations from Quantum Field Theory
Wightman axioms
"E.
Causality.
The fields shall satisfy causal commutation relations of either the bosonic or fermionic type.
If the supports of the test functions f and h are spacelike to each other then either
[i(f), j(h)] = 0
or
[i(f), j(h]+ = 0”
(I.e. measurement on one field operator has no effect at events spacelike separated from it.)
"G. "Time-slice axiom." "Primitive
Causality"
"There should be a dynamical law which allows one to compute fields at an arbitrary time in
terms of the fields in a small time slice...[formula for time slice]”
(I.e. propagation of field operators akin to propagation of ordinary fields that admit well posed Cauchy problem, domains
dependence, etc.)
R. Haag, Local Quantum Physics: Fields, Particles, Algebras. p.57
20
Reichenbach's common cause principle
It is not a principle that defines the nature of causation.
It is really a rule for detecting causal connections
without defining what that connection is.
It is a fallible rule and so not suitable for a definition.
A correlation between A and B can be screened off by
conditionalizing on C licenses the inference that C is the
common cause of A and B. But there remains a very small but
non-vanishing probability that A and B may be interacting
causally nonetheless.
21
Causal talk in modern physics…
… does not conform to the first horn (causal principle as universal factual restriction).
It does not supply a contingent, universal causal principle.
Otherwise all physical theories that do not explicitly invoke
conformity to a light cone structure would be deficient causally.
Newtonian mechanics was not causally deficient, just odd.
Sciences can invoke causation without making explicit use of
the lightcone structure of spacetime. (Geology)
… does conform to the second horn (causal talk as honorific).
The complaint that processes violate causality requirements in
modern physics is not a complaint that they violate some
overarching metaphysical principle.
It is an abbreviation for the objection that they do not respect the
light cone structure spacetime is known empirically to have.
22
Conclusion
23
Principal Claims
Negative Thesis
Denial of causal fundamentalism, which asserts that
the world is governed by a principle of causality.
Failure of any stable, factual principle of
causality to emerge from our science.
Positive Thesis
Sciences restricted to hospitable domains adopt
forms of various folk sciences of causation, which
explains the prevalence and utility of causal notions.
Folk sciences generated by same
mechanism that gives us gravitational
forces and caloric.
Modern Physics
Causal anti-fundamentalism is compatible with the
widespread presence of causal terms and principles
in modern physics.
Causal talk arises as part of assuring that
processes respect the light cone structure
24
of spacetime.
Finis
25
26
Analogies to reduction relations in science
Folk theories of causation are warranted as physical theories in so far as they
capture the relevant physical content of the embracing theory.
Folk theories of causation are attractive for their ease of comprehension and
ease of use.
Disanalogies to reduction relations in science
Different notion of causation recovered in different domains.
Folk theories tend to be less precise than the corresponding scientific
theories.
27
Are causes real?
Strong fictionalism.
Nothing is real unless it is in the
ontology of the final science.
Inscrutability. Infinite regress.
We will never know that we know what is
real.
Qualified prudent realism.
Entities of present theories are
real in so far as they are
structures licensed by further
science.
It is a real property of spacetime that its
curvature sometimes manifests as a force.
It is a real property of random energy
distributions that they sometimes manifest as
a conserved fluid.
Strong realism.
Every entity of a functioning
science should be construed
literally.
Heat is NOT a conserved fluid; it just behaves
like one sometimes.
Causes are as real as
gravitational forces and caloric.
28
We see faces and figures in clouds…
Fundamentally, there are no faces and figures there.
We all see them.
The shapes are real in the sense that the nose really is a
lobe of nose-shaped cloud
29
Appendices
30
Varieties of causal skepticism
Humean/Positivist skepticism
Anti-Fundamentalism
Anti-metaphysical.
Epistemically pessimistic.
Skeptical of content of science beyond
observation.
Causation just is constant conjunction,
functional dependence of facts.
Anti-apriori.
Epistemically optimistic.
Science makes discoveries beyond observation.
Causation is an attempt to preempt discovery.
(Weaker) Eliminativism
Russell: “…oddly enough, in advanced
sciences such as gravitational astronomy, the
word ‘cause’ never occurs…”
Mach: “…science of the future will discard
the idea of cause and effect as being formally
obscure.”
31
An Independent Principle of
Causality in Electromagnetic
Scattering?
32
Scattering from a dielectric
Incident plane
wave approaches
from left…
…excites dielectric
atom and generates
scattered field…
…and passes.
Total field
(incident plus
scattered)
Scattered field
(total minus
incident)
33
Frames from animation at http://www.met.rdg.ac.uk/~swrhgnrj/maxwell/circle3.html
Basic Physics of Scattering
Scattered field at
position x
is a linear sum of
incident fields at same
position x and other times

scattered(x,t) 
G(x,t)  incident (x,t  t')dt'


No influence
from future
(“NIFF”).
G(x,s)=0 for all s<0.
How do we
arrive at this
relation?
(“LIN”)
Hence incident
can only
contribute to
scattered for
t-t’>=0
i.e. t’<t.
34
… Use the Principle of Causality
NIFF… “is an accord with our fundamental idea of causality in
physical phenomena”
LIN+NIFF is… “the most general spatially local, linear, and
causal relation can be written between [scattered] and
[incoming] in a uniform isotropic medium. Its validity
transcends any specific model of [the dielectric].”
Mathias Frisch
Final results are “are of very general validity, following from
little more than that assumption of the causal connection [LIN]
betweed [scattered] and [incoming].”
J. D. Jackson, Classical Electrodynamics
…and similar remarks from others.
35
… Use the Principle of Causality
Frisch:
“…now we postulate an additional constraint on
all causally possible models that
an effect cannot temporally
precede its cause…”
“It might in fact be true that effects never precede
their causes. But I think that we can allow for the
possibility that a certain causal condition is not
true in general and nevertheless take it to be
physically well-founded.”
36
Not clear that Jackson does use an
independent principle of causality.
First, Jackson computes a simple model of the
dielectric by ordinary methods and finds it
conforms with NIFF.
(Time reversals of ordinary scattering precluded by choosing
the boundary condition of dielectric charges initially at rest.)
Then, Jackson finds more general cases too hard to
compute, but he expects nonetheless that the final result
would still conform with NIFF if only we do all the sums.
Awkward… so he
proclaims NIFF fits
with “causality.”
No precise principle of
causality is formulated or
applied to arrive at NIFF.
37
Not clear that Jackson could use an
independent principle of causality.
Principle: “An effect cannot temporally
precede its cause.” may hold only some
times. When?
If only for scattering,
then we are merely
restating NIFF.
Literature that admits
backwards causation in
physics:
time travel, tachyons
If more broadly, then to
which processes does it apply?
Why is this any better than “it
applies except when it
doesn’t”?
38
How could a principle of causality pick out the
“true” forward direction?
Forward
Reversed
Forward
Reversed
Time revesibility of
electrodynamics.
Any feature of
“forward” system has a
perfect correlate in the
“reversed” system.
Add a factual time
direction to spacetime?
… but both processes assembled from
pieces that are locally time reversible.
Allow them locally but prohibit assembly?
39
What precisely does the principle of causality say?
“An effect cannot temporally precede its cause.”
Which state is the effect
and which the cause, in any
Must be precise enough to be
applied in a computation in
mathematical physics.
The cause comes earlier? That makes the
principle true by definition.
process with states that evolve
over time?
What counts as a
cause? An effect?
What sorts of processes
are properly causal?
States over time? At an instant? Which
hypersurface of simultaneity?
“Causal” means later states depend on earlier?
What of Lagrange principles that pick out
motions by extremizing the entire history? 40