STUDY QUESTIONS FOR M.A. EXAMINATION IN PHILOSOPHY OF SCIENCE                  
                                                                                
 JANUARY 2005                                                                   
                                                                                
1. One often speaks of the laws of physics, but less often of the laws of
geology or physiology.  What role do laws play in the various sciences?  Are
the so-called laws of economics or psychology the same sort of thing as the
laws of physics?

The laws of economics (supply and demand) are not the same sort of things as
the laws of physics, nor even are the laws of geology, physiology, genetics,
and biology.  Laws (or something similar) play a role (to some degree) in
explanation and prediction in all of these fields.  Laws are more frequently
discussed in the context of, and seem to play the greatest role in, physics.

Laws (at least in physics), point to universal generalizations, pick out
regularities, and allow us to make reliable predictions/explanations. 
Usually (not always, see epistemic regularity theory) laws are taken to be
generalizations plus some other characteristic.  Laws are distinguished from
other generalizations because they are nomological, explanatory as well as
predictive, capture modality and physical necessity, support counterfactuals
(subjunctive conditionals), unify, point to causal underlying mechanisms. 
Further, laws in physics usually specify precise quantitative relationship.

Whether a field has laws or not is sometimes taken as, or as part of, the
demarcation criteria as to whether or not it is a science.  As one example,
reductionist explanations rely heavily on laws.  One motivation for saying
that psychology (social science) has laws is that if these laws of
psychology can be reduced to the laws of biology (science), then we have
established the psychology is a science.  But surely this isn't the only way
to establish that something is a science.  After all, we don't need to
reduce biology to physics for it to be a science.

Sometimes the existence of laws in a field, rather than their role in
reductive explanations, is used as the demarcation criteria.  But do we
really want to say that any field which has laws is a science?  I don't
think so, as science is a discipline, and being a science should have
something to do with a method/practice.

Though there are many who support the claim that there are laws in economics
(Rosenberg and Kincaid), economics is sometimes barred from science on the
grounds that the laws of economics are hedged (ceteris paribus) laws.  But
this can't be right. Smart and Kitcher hold to hold that there are no laws
in biology/genetics (or if they are, they are very different than what we
usually mean by laws).  

Cartwright takes this even further, saying that if biology has no genuine
laws (because the so-called laws in biology are spatio-temporally restricted
and refer to specific entities), then physics as well is law-less, as laws
in physics both have exceptions and are at best approximately true rather
than strictly true (the same seems to be true in economics where
idealizations also play a role).  Then we'd have to say that physics is not
a science, at which point it seems there is no science.  So having hedged
laws cannot meaningfully exclude a field from being considered a science.

Despite the above similarities, I don't believe the laws in physics are the
same kind of things as the laws in biology or economics.  One thing they
have in common though is that, within certain conditions, all laws are
generalizations that support counterfactuals.  But what kind of
counterfactuals, or rather, what type of necessity?  The laws in biology do
not capture physically necessity, nor should we want them to.  Instead, the
laws of biology capture biological necessityand the laws of economics
economic necessity.  We should expect laws in different sciences to be not
of the same kind, but for there to be an analogous relationship between the
laws in various special sciences and those sciences themselves.

About the only thing which becomes clear from all of this, is that the
notion of laws, even within science, is geared specifically to physics and
even still is quite muddled.  That it is difficult to find or explain laws
in other sciences,then, is no surprise.  It may be that in asking whether
there are laws, we are asking the wrong question.  Instead, we should be
asking for a demarcation criteria.

Question: Are there any laws in psychology.

[See Roberts and Kincaid in debates; Kitcher's 1953 in C&C for no laws in
classical genetics; biology and social science in Companion?]

*****
                                                                                
2. Pierre Duhem provides a classic statement of a widely held thesis of
holism in the philosophy of science: "hypotheses shall be chosen in such a
manner that from them taken as a whole mathematical deduction may draw
consequences representing with a sufficient degree of approximation the
totality of experimental laws." What is to be said in favor of this thesis? 
What is to be said against it?  Why is it important to the philosophy of
science?

Full quote (Duhem p220, with original italics):  "hypotheses shall be chosen
in such a manner that from them *taken as a whole* mathematical deduction may
draw consequences representing with a sufficient degree of approximation the
*totality* of experimental laws.  In fact, the proper aim of physical theory
is the schematic representation by means of mathematical symbols of the laws
estaablished by the experimenter; any theory one of whose consequences is in
plain contradiction with an observed law should be mercilessly rejected. 
But it is not possible to compare an isolated consequence of theory with an
isolated experimental law.  The two systems must be taken in their
integrity: the entire system of theoretical representations on the one hand,
and the entire system of observed data on the other.  As such they are to be
compared to each other and their resemblance judged."

For Duhem, the observational agreement of a theory's consequences with
experimental results is key for confirmation.  "Agreement with experiment is
the sole criterion of truth for a physical theory." How is a theory
confirmed?  Hypothetico-deductively, that is, by looking at the
observational consequences that can be derived from the theory.

Duhem claims that individual hypotheses cannot be compared (confirmed) with
the results of experiment. However, even though hypotheses cannot meet
experiment (Foucault's experiment) individually, they can be compared
collectively as a theory (Newton's or Huygens's optics) with the totality of
experimental laws.  This is Duhem's holism.  From the results of such a
collective comparison, we can decide to accept or give up a whole theory
(collection of hypotheses) and when we do so in favor of another theory, we
effectively give up those individual hypotheses of the older theory which
are not also part of the newer theory.

Import of Duhem's holistic thesis: 1) used to argue against that Poincare's
conventionalism that all physical theories and postulates are testable and
falsifiable, 2) used to conclude that crucial experiments are impossible
in physics, 3) used to conclude a statement of local underdetermination.

Unsorted ramblings:

Duhem "agrees with Poincare that individual postulates [Newton's first law
of motion] cannot be tested directly by any experiment.  But, contrary to
Poincare, Duhem does not take this to imply that these and other laws are
irrefutable or that they are really defintions in disguise.  [They] can be
tested indirectly ... by being part of a system of hypotheses that make
testable predictions.... an individual postulate can be refuted if the
system to which it belongs fails to represent reality as well as some rival
systems of hypotheses that contains a different postulate.  When the new
system replaces the old, the postulate is, in effect, refuted." (C&C, 364-5)

We cannot give experimental meaning to a hypothesis (eg principle of
inertia) without making other assumptions.  Similarly, we cannot test a
single hypothesis in isolation (it might be true when taken with one set of
auxillary hypotheses/circumstances but false with another).

We cannot subject certain principles to direct experimental test, but Duhem
does not think this means that hypotheses are beyond the reach of
experimental refutation.  Taken in isolation, hypotheses have no
experimental meaning, but collectively they form theories which aim to
represent experimental laws.  We can compare the conequences of theory with
the whole of experimental laws.

Duhem's statement is important as an argument against Poincare's
conventionalism.  Duhem argues that our principles (held hypotheses) are not
mere definitions or conventions.  We can give them up.  He agrees that we
cannot confirm or refute them in isolation, but believes we can contradict a
group of hypotheses.

When we contradict a group of hypotheses, we can make adjustments to any of
the hypotheses.  Duhem warns against being too dogmatic in our acceptance of
any hypothesis.  No hypotheses are immune from refutation, even Newton's
laws of gravity can be overthrown.  (Of course, we may choose to keep a
particular hypothesis at all costs and make revisions elsewhere, but
sometimes "good sense" concludes that it's time to give up the hypothesis we
would prefer to save.)

For import, tie in holism, empirical equivalence, underdetermination, no
crucial experiment.  "The physicist can never subject an isolated hypothesis
to experimental test, but only a whole group of hypotheses; when the
experiment is in disagreement with his predictions, what he learns is that
at least one of the hypotheses constituting this group is unacceptable and
ought to be modified; but the experiment does not designate which one should
be changed." Duhem, p187.

Possibly related quote from Duhem p19: A physical theory is not an
explanation.  It is a system of mathematical propositions, deduced from a
small number of principles, which aim to represent as simply, as completely,
and as exactly as possible a set of experimental laws."

Can I work in the idea of postulating an idealization and then comparing it
with experimental observations?  I'm not sure.

[See Duhem; holism and convention in companion; Gillies and Duhem commentary
in C&C]

*****

3. Quine says that "physical theory is under-determined even by all possible
observations.... Physical theories can be at odds with each other and yet
compatible with all possible data in the broadest sense.  In a word, they
can be logically incompatible and empirically equivalent." What is the
argument for this claim?  Of what significance is it to the philosophy of
science?

There are many different theses that go by the name of underdetermination.
Roughly, underdetermination is the claim that multiple theories fit
observation, or that it is possible to make changes in theory without there
being a corresponding change in observation.

Weak underdetermination of theories is the claim that two theories can both
fit equally well to the observational data available at a particular point
in time.  This is relatively uncontroversial, and the status of the
Ptolemaic, Tychonic, and Copernican world views as of 1600 is a standard
example.  Even when observational data cannot determine between two
theories, there can still be pragmatic reasons to prefer/accept one theory
over the over (as arguably can be seen with the Tychonic and Copernican
theories in 1633).  Weak underdetermination is often illustrated as a curve
fit; multiple curves fit a set of data points.  

The thesis of strong underdetermination is far more controversial:  Perhaps
theories are underdetermined not just by the observations available at a
given point in time, but by all possible observations.  Perhaps, even were
all possible observational data to be available, every theory would have at
least one alternative (inconsistent) theory which is observationally
(predictively) equivalent.  

Realists, of course, must deny strong underdetermination, or else we have no
grounds for thinking that we are adopting truer and truer theories. 
Instrumentalists often appeal to strong underdetermination to argue against
realism.  This is the real import of the thesis: if true, it presents a
decisive blow to realism (we can't get to the truth).  This leads to
skepticism and/or instrumentalism.  A second import of the thesis is that
SUT implies WUT, suggesting that we need to look beyond empirical
considerations for understanding theory choice (and whether theory choice is
rational, see Kuhn).

Strong underdetermination has had a number of proponents, including Duhem,
van Fraassen, and Quine (for a time).  Duhem's and Quine's
underdetermination (quite similar) theses are based on holism.  Duhem:
Experimental tests of theory also rely on auxillary background assumptions.
When experimental observation does not conform with the prediction given by
a theory, there is no way to determine which of the experiment's assumptions
should be rejected.  That is, falsification is vague and there can be no
crucial experiment.  Logic alone cannot tell scientists when to give up one
theory in favor of another (but logic and "good sense" combined can, Duhem
217).  Quine: Theories face the tribunal of observations as a whole. 
Adjustments/revisions can be made somewhere in the system to construct an
alternate system which, while logically icompatible with the first on a
theoretical level, is observationally equivalent.  Therefore,
observation/evidence do not determine/constrain theory choice.

Problems with strong underdetermination:  (1) Strong underdetermination
presupposes a dichotomy between the observable and the theoretical which
realists find unacceptable.  (2) Apparent cases of underdetermination may be
cases of equivocation, not genuine incompatibilities.  (3) There is no
threat of underdeterminism when one takes a less narrow view of empirical
evidence and allows pragmatic factors (e.g., simplicity) to play a role in
theory choice which is consistent with realism.  (4) There are no actual
cases in history of strong underdeterminism, so there is no reason to think
that it is true.  (5) Strong underdetermination relies on an incorrect
analogy with WUT and curve-fitting when having all points would uniquely
determine a curve. (6) Perhaps rivals can always be constructed by
modification (as Quine suggests), but these are easily recognizable as
artifices, not real theories. (7) Arguments for strong underdetermination
relying on holism may be question-beggingly circular.

Quine (Two Dogmas): "total experience is like a field of force whose
boundary conditions are experience.  A conflict with experience at the
periphery occasions readjustments in the interior of the field .... the
total field is so underdetermined by its boundary conditions, experience,
that there is much latitude of choice as to what statements to reevaluate in
the light of any single contrary experience .... it is misleading to speak
of the empirical content of an individual statement .... Any statement can
be held true come what may, if we make drastic enough adjustments elsewhere
in the system."
     
[See underdetermination, holism, Quine, and pragmatic in Companion;
commentary for Duhem, Quine, and Laudan in C&C.]

*****
                                                                           
4. According to the "unity of science" thesis, the various special sciences
can be reduced to more fundamental sciences, as biology has been reduced to
chemistry and chemistry to physics.  What does it mean to say that one
science has been reduced to another?  Has biology been reduced to chemistry,
or chemistry to physics?  Should reduction to biology be a demand placed on
psychology?

Discuss varieties of reduction.  It looks like the kind here is
theoretical-derivational.  For Nagel, the reduction of (the statements of)
one scientific theory (T) to another (T') is a deductive explanation of T
(and all of its observational consequences) in terms of T'.

Examples of reductions: reduction of Galileo and Kepler's laws to Newton's
gravitation theory; reduction of classical thermodynamics to statistical
mechanics; reduction of optics to electromagnetic theory; reduction of
classical genetics to molecular biology (controversial); reduction of
biology to chemistry to physics (maybe one day).

Laws play a key role in reduction.  To say that a T has been reduced to T'
is to say (at least) that the laws of T can be restated in terms of T'.  But
the laws of psychology, biology, etc, seem to be of a different sort (and
not as important to theory) as those of physics.  (Kitcher says strict
reduction of classical genetics to molecular biology fails for these
reasons.) So this might show that reduction/unity of science understood in
these terms is impossible.

We want unification because it is an epistemic virtue (Kitcher).  Part of
what makes Newton's theory of graviation so good is that it unifies Kepler
and Galileo.  Unification is also related to conceptual simplicity (fewer
postulates encompass more observations).

One reason to want unity of science is that reduction is considered a
demarcation criteria for science.  If something reduces to a simpler
science, than it is itself a science.  So the motivation for attempting to
reduce psychology to biology is that if a reduction is possible then
psychology is a science.  The question then is really "Is psychology a
science?"

(This doesn't seem right to me though because we don't think that biology is
a science because it can be reduced to physics.  We don't require that
something be reduced to another science to be a science.)
 
We want reduction because we want explanation.  To say theory A is reduced
to theory B would be to say that we can rephrase all the claims of theory A
in the terms of theory B, that if B is established then so too is A.  Do we
need reduction to get explanation?

I don't think biology has been reduced to chemistry.  I'm not sure it even
makes sense to talk about reducing an entire science to another.  Contrast
the noun/verb senses of science.  When we talk about reducing biology to
chemistry, do we mean the facts of these sciences?  Or do we also mean that
the methods/approaches in biology get reduced to chemistry. I do think it is
clear how some theories within biology relate to some theories in chemistry.

I don't think we can/should want to strictly speaking reduce psych to bio to
chem to physics, but we can relate them in reductionist ways, find
relationships between these sciences.  And I think this is all we need for
explanation.  Similarly for the demarcation criteria, I think reduction is
sufficient for demonstrating that something is a science, but it is not
necessary (not the only way).

[See unification, unity in Companion;  Kitcher, Nagel in C&C?; Silberstein,
Dennett?, Fodor in Metaphysics Companion?, Rosenberg p79]

*****

5. "The best kinds of evidence for the reality of a postulated or inferred
entity is that we can begin to measure it or otherwise understand its causal
powers.  The best evidence, in turn, that we have this kind of understanding
is that we can set out, from scratch, to build machines that will work
fairly reliably, taking advantage of this or that causal nexus.  Hence,
engineering, not theorizing, is the best proof of scientific realism about
entities." To what extent does this proposal of Ian Hacking's end the
dispute between realism and instrumentalism?

In contrast to realism -- the view that the world studied by science exists,
has the properties it does independently of our beliefs, and that the aim of
scientific theories is truth -- and instrumentalism -- a form of
anti-realism which says that scientific theories are to be understand as
tools for prediction, but not literally claims of truth or ontological
commitments -- Hacking offers a position which is both realist and
instrumentalist.

Hacking is an experimentalist realist -- a realist about entities but not
about theories.  That is, he is committed to the existence of unobservable
entities (electrons), but not to the truth of theories about such entities. 
This sounds like what most experimenters believe: that their postulated
unobservable entitites really exist and can be interfered/intervened with in
experiment, but they aim not at at truth, but rather utilty in their
theories/models/descriptions of them.

Obviously experimentalists don't believe in the existence of all postulated
unobservable entitites.  How do they determine which/when entitites entitle
commitments?  Unlike scientific realists, they don't infer existence from an
entity's role in successful theories.  Instead, Hacking suggests that we are
convinced of the reality of entitites (electrons) when we use
well-understood causal properties of the entities to build new devices to
interfere in other more hypothetical parts of nature.  This is analogous to
the way we come to believe in the existence of observables.  Our notions of
reality are formed from our ability to change the world.

It is not that electrons are supported by (inferred from) experimental
results, but that they are presupposed in the design/execution of
experiments.  It is not, as the realist supposes, that the entities save the
phenomena, but that they create new phenomena.  We rely on their causal
properties to make instruments.  The fact that our so-designed instruments
work provides strong evidence that the entities exist, for we have been able
to manipulate them to produce new phenomena and investigate other aspects
of nature.

Hacking's argument for realism from the success of tools differs from the
realist's typical argument from the success of science because the tools are
useful in a different area of science than that in which they were
developed.  We rely on the existence of photons to create a microscope, but
then we successfully use that microscope in, say, biology.

Hacking's view ends some of the dispute between realism and instrumentalism
by offering an alternate theory, which is even something of a compromise
between them: realism about entitites, but not about theories. 
Instrumentalists who objected only to the truth-of-theories part of
scientific realism may welcome experimental realism.  Experimental realism
saves the instrumentalists' intuitition that what we want out of science is
practical tools, something we can use, rather than the (abstract) truth.

The rest of this is just me saying things badly, or rather stringing
together quotes:

First, realism and instrumentalism.  Realism: The world studied by science
exists and has the properties it does independently of our beliefs.  The
unobservable entities postulated by science exist.  Theoretical claims about
unobservable entities are to be understand as being claims of literal truth. 
The aim of science is truth.  Theory acceptence involves belief in the truth
of the theory.  Instrumentalism: One form of anti-realism, scientific
theories are to be undertood as instruments, useful tools for prediction,
but not literal claims that are true.  Talk of postulated unobservables does
not entain ontological commitment.  

Scientific realism and instrumentalism aren't the only options available.
There's can Fraassen's anti-realist view of constructive empiricism:
unobservables may exist, but theories and talk about unobservables is not to
be taken as literally true; what matters for theories is empirical adequacy
-- everything they say about observables is true.  There are also the views
of Poincare, Fine, Cartwright, Hacking.

Hacking suggests a position which is realist about entities, but not about
theories.  That is, he is committed to the existence of unobservable
entities (electrons), but not to the truth of theories about such entities. 
This is an experimentalist (or experimenter's) realism, and in fact, it
sounds like what most experimenters believe: that their postulated
unobservable entities really exist and can be interfered/intervened with in
experiment, but their theories/models of them are aiming not at truth but
adequacy.

Obviously experimentalists don't believe in the existence of all
unobservable entities.  And unlike scientific realists, they don't infer the
existence of their entities from their role in a successful theory. 
Experimentalists believe in those entities which they can manipulate, those
which can be used to investigate something else.  "We are completely
convinced of the reality of electrons when we regularly set out to build --
and often enough succeed in building -- new kinds of devices that use
various well understood causal properties of electrons to interfere in other
more hypothetical parts of nature" (Hacking, C&C 1158).

Just as we believe in the existence of observables when we can causally
interact with them, so too we believe in the existence of unobservables. 
"Reality has to do with causation and our notions of reality are formed from
our ability to change the world.... We shall count as real what we can use
to intervene in the world to affect something else, or what the world can
use to affect us." (Hacking, Representing and Intervening, C&C commentary)

We do not infer electrons from our experimental results, but rather
presuppose them when we design/execute experiments.  The experimentalist
does not believe in electrons because they save the phenomena, but because
they create new phenomena (Hacking, C&C 1165).  We rely on the causal
properties of electrons to develop instruments.

"The 'direct' proof of electrons and the like is our ability to manipulate
them using well understood low-level causal properties.  I do not of course
claim that 'reality' is constituted by human manipulability.  We can,
however, call something real, in the sense in which it matters to scientific
realism, only when we understand quite well what its causal properties are"
(Hacking, C&C 1153).

"Experimental physics provides the strongest evidence for scientific
realism.  Entities that in principle cannot be observed are regularly
manipulated to produce new phenomena and to investigate other aspeects of
nature.  They are tools, instruments not for thinking but for doing.  The
philosopher's standard 'theoretical entity' is the electron.  I shall
illustrate how electrons have become experimental entities, or
experimenter's entities.... if we come to understand some of its causal
powers and to use it to build devices that achieve well understood effects
in other parts of nature, then [it is not merely a hypothetical entity]....
The experimentalist need only be a realism about the entities used as tools
[not about the truth of theories]" (Hacking, "Experimentation and Scientific
Realism, C&C 1153).

The way in which this differs from the argument from success is that
instruments are useful in a different area ("other aspects") of science. 
Examples: We rely on the existence of photons to create a microscope, and
then we use that microscope in biology; we rely on the existence of
electrons to build transistors which we use for radios.  That microscopes
and transistors work (and work reliably) provides justification for belief
in the existence of photons and electrons.

Hacking's view ends (some of) the dispute between realism and
instrumentalism in part by offering a compromise between them: realism about
entities, but not about theories.  Hacking shows that we need not be
realists about theories (we need not even consider theories, theory-free
realism, consider experiments -- interactions with the world rather than
representations of the world) to be realists about entities. 
Instrumentalists who only took qualms with the truth-of-theories part of
scientific realism can happily be experimental realists.  Hacking's position
saves the instrumentalists' intuition that what we want out of science is
practical tools, something we can use, rather than (the more abstract)
truth.

Realism says science aims at truth, instrumentalism says science aims at
usefulness, perhaps Hacking says usefulness demonstrates truth?

"Hacking urges that we sever the view that theories are true from a realist
commitment to the existence of unobservables.  According to Hacking's
experimental realism, our justification for believing that unobservables
exist rests on our ability to manipulate them in scientific experiments"
(C&C, p 1051).

Ian Hacking argues that scientists need not be realists concerning at least
some elements of theory - but experimentalists become realists once they
learn to manipulate and use entities (e.g., electrons), especially if these
are used to learn something about other entities (e.g., the weak neutral
current). (http://www.drury.edu/ess/philsci/philsciov.html)

"Hacking develops a form of realism that involves no endorsement of theory
.... It is theoretical entities that we wish to be realist about, not the
theories that posit such entities."  Entities exist, but not theories about
them are not true.... Hacking appeals to [the entities'] technological use
in investigating yet conjectural aspects of nature." Companion, 400-401.

Can I work this in? Is Hacking really any different from the argument from
suscess?  Resnik thinks not (both rely on a similar abductive inference). 
One of the main arguments for scientific realism is the argument from the
success of science (Sellars, Smart, Harman, Putnam): scientific realism is
"the only view that doesn't make the success of science a miracle". 
However, this argument has been pretty well defeated on the grounds that: it
relies on a suspect abductive inference (Fine), past theories have been
susccessful but still false (Laudan), theory is underdetermined by
observation (Quine), and alternate accounts for the success are available
(Van Fraassen).


[See experiment, realism in Companion; Hacking & Resnik in C&C and
commentary; Griffiths 252-254 in Blackwell Guide?]
    
*****
                                                                                  
6. Carl Hempel once remarked, "The establishment of a general theory of
confirmation may well be regarded as one of the most urgent desiderata of
the present methodology of empirical science.  Indeed, it seems that a
precise analysis of the concept of confirmation is a necessary condition for
an adequate solution of various fundamental problems concerning the logical
structure of scientific procedure." What is the task of a theory of
confirmation, and why is it important?  Explain why the general approach
that ought to be taken to this task is still very much a matter of dispute,
in the process sketching at least two of the widely adopted general
approaches and indicating how they fall short.
 
Deciphering the question: What does it mean to have a theory of
confirmation?  Why would we want one?  What theories have been proposed?
What's wrong with them?

Scientific theories are not established as certain, but rather as nearly
certain or highly-likely or well-justified.  Confirmation involves
increasing the degree of justification of a theory, it is a matter of
proability and statistics.  To say that a body of information is evidence in
favor of a hypothesis is to say that the hypothesis receives some degree of
support or confirmatino from that information.

We want a theory of confirmation because we want to know which
hypotheses/theories are worth accepting and be able to evaluate how much
certainty we should place in them (or how tentative we need to be in
accepting them).  A theory of confirmation provides us with an approach for
comparing rival hypotheses.

What sort of information confirms a hypothesis has long been controversial.
There are many different theories of confirmation including:
hypothetical-deductivist, inductivist, falsificationist, predictionist,
simplicity, and a variety of Bayesian/probabilistic approaches.

Hypothetical-deductivist: Hypotheses are confirmed by the agreement of
evidence with their true deductive consequences.  (Problem: the raven
paradox.  Consider the hypothesis "all ravens are black".  A deductive
consequence is "all non-black things are non-ravens".  So an instance of a
white swan or a blue book would provide some degree of confirmation to "all
ravens are black", which is absurd.  Also, the grue paradox presents a
problem as alternative equally-well support theories can be constructed.)

Inductivist:  As more positive evidence for a theory is amassed, it is taken
to be better confirmed, more justified.  (Problem: Induction is problematic,
the future/unobserved may not be like the past/observed.  Once you are
looking for confirming instances, you may be able to find them everywhere. 
Could be systematically overlooking negative evidence by repeating the same
kind of experiment/observation.  Some positive instances provide more
confirmation than others.  Law of diminishing returns.)

Falsificationist: Hempel-modified Popper.  Popper argues against inductive
confirmation and induction more generally, saying that the scientific method
uses/can use falsification instead.  True theories are never inductively
confirmed; instead, false theories are falsified.  This suggests an
alternate criteria for confirmation (though Popper himself thought of it as
corroboration rather than confirmation).  The scientific method involves
proposing conjectures and attempting to falsify them.  Those theories which
stand up to genuine falsification attempts are more likely to be true in
virtue of the fact that they have survived.  (Problem: By giving up
induction, Popper gives up the ability to use prediction for testing.  Is
Popper really giving up induction?)

Predictionist: To confirm a theory, see what predictions follow from the
theory and test whether they obtain.  The more surprising/risky the
predicted outcome, the better confirmed the theory.  Newton offers
predictions about the figure of the earth, the variation of surface gravity
with latitude, etc to confirm universal gravitation. Note: A related view is
explanationism, the view that only the explanation of previously knkown
results can confirm a theory.  (Problem: Hempel: A result that strongly
confirms a result in one context, when the result is new, would not confirm
it very much in another context, when the result was already known.  Why
should observation-agreement count only for new observations?)

Simplicity: Variation on predictionist: Multiple theories can be fitted to
the same predictions, so it's not the predictive success in and of itself
that provides confirmation.  It's the fact that the theory which has
predictive success is also simple (not ad hoc).  (Problem: How do we connect
simplicity with truth?)

Bayesian (subjective Bayesian): Ramsey, Carnap, Reichenbach, Salmon.  Covers
a whole slew of confirmation theories which use probabilities (quantitative
rather than qualitative, seems to solve raven and grue paradoxes).  A
hypothesis is confirmed by evidence if the probability of the hypothesis
given the evidence is greater than the probability of the hypothesis
(without the evidence).  The probabilities (likelihood, prior probability)
can be interpreted as a priori logical probabilities, empirical frequencies,
expected values, or subjective personal probabilities (C&C 640).  Bayes'
theorem: P(H/E)=(P(E/H)P(H))/P(E). (Problem: Prior probabilities are
subjective.)

[See confirmation, evidence, probability, simplicity in Companion; Hempel p
445+, commentary, Popper p 7, Bayesian in C&C]
   
*****
                                                                                  
7. Marx's reaction to Darwin's Origin of Species was: "Not only is a death
blow dealt here for the first time to 'Teleology' in the natural sciences,
but their rational meaning is empirically explained." What is "teleology",
and what place does it have (if any) in science?  Is Marx right about
Darwin's contribution?

Teleology has no place in science. Explaining away teleology, like
explaining away ghosts, is no loss.

Teleological explanations are those that answer "for what purpose?"  Some
philosophers think such explanations have a place in science, others don't.
Teleological explanations include explanations of: 1. goal-directed
behavior, 2. functional tools, 3. organs of living organisms, and 4. social
functions within societies.  Teleological explanations appear frequently in
biology, as in "the organ x is present because it has/had function f."

Many opponents to the theory of evolution by natural selection worry that   
the theory explains away purpose, design, and intentionality.  "Darwin's
dangerous idea is that Design can emerge from mere Order via an algorithmic 
process that makes no use of pre-existing Mind" (DDI, 83).  Before Darwin's
theory, one might answer the question "Why are our bodies symmetrical" with
something like "Because God chose us to be in this aesthetically pleasing
way." Similarly, the question "Why are we here?" might be answered "For
God's purpose, to help others, ...." If we are designed by an intentional
designer (God), then we have purpose, the purpose for which we were
deisgned.  Instead, Darwin's theory tells us that there is no intentional
design.  We are not the products of some wise creator, but rather the
products of an intentionless mechanical process.

Natural selection is actual an explanation for the rational basis of
teleology.  It is possible for the apparent wonders of design (namely, us)
to be the result of a mindless mechanical process.  Purposive design can
emerge from mere order.

[See social science p456, teleology in Companion; Dennett,
http://pubs.socialistreviewindex.org.uk/isj71/darwin.htm]

*****     
                                                                                 
8. Thomas Kuhn remarked, "An apparently arbitrary element, compounded of
personal and historical accident, is always a formative ingredient of the
beliefs espoused by a given scientific community at a given time." Must
every scientific theory include elements that are in some sense or other
arbitrary?  Explain your answer and its bearing on the currently widely held
view that scientific truth is a "social construct".

Every scientific theory includes elements that are arbitrary, if by that we
mean not logical.  It would be foolish to deny that, in practice, scientists
are influenced by a variety of social factors.  The charge is that the
influence of these arbitrary elements cannot lead to objective progress. 
Scientific theory change, thought to be progressive, is really just change.

Social constructionism (which I know hardly anything about): Scientific truth
is a social construct; scientists' search for truth is motivated at least
partly by social concerns (personal ambition, politics, etc).

Strong social constructionism: Science is a set of conventions of one
particular culture in the circumstances of one particular historical period. 
It is a discource of a specialized interpretive community.  (Does this mean
I should be talking about conventionalism?)

Bloor's strong programme is a movement in sociology of science that studies
the psychological and sociological causes of scientific beliefs and
theories and insists that all scientific beliefs (whether true or false)
should be explained in the same sort of way in terms of social and cultural
factors.  Sociological factors are pre-eminent in deciding whether a theory
is accepted as being true or not.

Many have misread Kuhn as saying that there are no good reasons for
paradigm acceptance, that theory choice is arbitrary, that the picture of
science is not that of progress and success, etc.

However, I do not believe that Kuhn's "arbitrary elements" lead to social
constructivism.  The arbitrary elements need not strip science of its claims
to objectivity and progress.  Kuhn himself seems to hold this view in SSR
and more forcefully in his later Essential Tension.  The role of endorsement
of theory by the scientific community, with its shared training, standards,
and epistemic values, may serve to explain how the influence of arbitrary
elements on individual scientists is compatible with objectivity.  "What
better criterion than the decision of the scientific group could there be?"
(Kuhn, 170)

What sets science apart from other fields is the commonality and insularity
of the scientific community.  Scientists write for their colleagues, they
share a common training and a set of values and beliefs.  Kuhn lists five
shared epistemic values: accuracy (predictive and explanatory), consistency
(internal and external), scope (broader scope than design intended),
simplicity (unifying power and/or parsimony), and fruitfulness (future
promise for new discoveries).  Scientists may agree on these values, but
come to different conclusions when using them to judge between rival
theories.  Thus, there can be variety and subjective input to judgments, but
the judgments themselves can be undertood rationally, rather than in terms
of subjective factors.

A related and stronger claim is that these personal arbitrary elements, by
creating variations among scientists and scientific beliefs, are precisely
what allows science to progress (at least in some sense of the word
progress).  Note: Sounds like a Darwinian fitness landscape.  When anomalies
first begin to appear, some scientists focus on the area of anomaly and
(radically) adjust the paradigm. Personal arbitrary elements allow for the
creation of a plurality of diverse theory-attempts by different scientists. 
Those which are promising, gain supporters, but supporters are not gained
all at once; the shift is gradual, ending only when the last of the clingers
die.  Some scientists remain within the normal tradition.  Neither group of
scientists is irrational in their behaviors, although clearly there is
something subjective (non-logical) going on in their choice.  Both having a
core group who are resistant to change and having individuals who are
willing to test out "extravagantly" unchartered waters are necessary to the
process of science.  Without the former, there would never be the stability
of normal science; without the latter, there would be dogmatic commitments
and no change.

[See Kuhn especially p2+, postscript 5, chapters 9,12,13, and Objectivity;
Kuhn, social factors, values, pragmatic, social science in Companion; last
section of Routledge?, Bloor?]

*****    
                                                                                  
9. The phrase 'experimentum crucis' -- literally, "experiment of the cross"
-- was introduced by Robert Hooke in the 1660s, providing a Latin phrase to
designate Francis Bacon's idea of experiments that act like sign-posts at a
crossroads. It was deployed famously by Isaac Newton in his work in optics
in the 1670s.  This pedigree notwithstanding, some philosophers of science
-- e.g.  Duhem -- argue that there can be no such thing as a crucial
experiment.  Yet scientists continue to talk of crucial experiments.  What
requirements would an experiment have to meet to be "crucial"?  To what
extent are crucial experiments possible?

Philosophical question: Are crucial experiments possible?
Historical question: Are those experiments we call "crucial" actually
crucial?  What role did they really play?

Following Duhem, crucial (crossroads) experiments in science seem to be
impossible in two replated respects, both of which follow from Duhem's (and
Quine's) holism.

First, no single experiment which falsifies some theory
simultaneously/immediately determines another.  In geometry the
falsification of one statement simultaneously establishes another statement,
it's negation.  If there are two statements to decided between, the
falsification of one would determine the other.  But this is not the case in
science; in science, there are never just two theories to decide between. 
An experiment which falsifies the wave theory of light does not establish
the particle theory of light (nor vice versa), as there are other
alternatives.  If we could enumerate all of the alternatives, we could
establish one by falsifying all of the others.  But in science we cannot
enumerate all of the possibilities.

Second, it is impossible even to conclusively falsify a theory/hypothesis,
because an isolated hypothesis cannot be subject to experimental test. 
Instead, groups of hypotheses (hypothesis plus auxillary background
assumptions) meet experiment together as a whole.  If the outcome of the
experiment is not in agreement with the prediction, this shows that there is
a problem somewhere in the system, but it does not show where.  Further
(Quine), auxillary assumptions can always be revised/modified so as to avoid
falsification.  (As in the addition of epicycles to the Ptolemaic system to
keep it observationally equivalent with the Copernican/Tychonic.)

While a crucial experiment in the logically decisive (deductive) sense
considered above is impossible, we do talk of crucial experiments in several
important senses.  Scientists have, in fact, proposed experiements which
they thought would be crucial to establishing their theories.  Many
observational consequences follow from Newton's theory of universal
gravitation, some of which would be quite surprising without Newton's
theory.  Newton made several predictions (the figure of the earth, variation
in surface gravity with latitudes), the successful outcome of which, he
rightly believed would be crucial in confirming his theory.  The idea here
is that of the predictionist theory of confirmation: successful predictions
count as evidence.  Experiments relying on this are crucial not to
establishing the theory with certainty, but to confirmation.

Related to the idea of predictionist confirmation, is the idea that such
so-called crucial experiments are crucial in convincing other scientists not
to accept the theory as certain, but to accept the theory as promising, as
worth working with.  Or, at least, they tell a scientist when to abandon a
theory; then working with the best available alternate is the best thing to
do.  Duhem says logic alone cannot tell a scientists when to abandon one
theory in favor of another, but a sort of "good sense" can.  Presumably this
good sense involves assessing the degree of confirmation and the promise of
fertility of a theory.  Duhem, Kuhn, and Lakatos all note that theres's no
single point at which it becomes irrational to hold an older theory in light
of the increasing evidence in favor of a newer one.  However, over time,
individual scientists make the switch.  Certain experiments/results are
crucial not to the whole community, but (in a weaker sense) to individual
scientists.  An experiment is crucial to the history of science if it
convinces others to accept the theory for working purposes.  Only when
enough people have done so will the theory have been embraced to the extent
that it is established as part of normal science.  Lakatos notes that
crucial experiments can be recognized as such only with hindsight.

Roughly a repetition of what I've said above, but more quote-based.

Popper: A theory is 'scientific' if one is prepared to specify in  advance a
crucial experiment (or observation) which can  falsify it, and it is
pseudoscientific if one refuses to  specify such a 'potential falsifer'.

Duhem: No single experiment can simultaneous falsify one theory and
determine another.  An experiment may falsify, but that does not establish
an alternate theory.

Duhem: Experimental tests of theory also rely on auxillary background
assumptions. When experimental observation does not conform with the
prediction given by a theory, there is no way to determine which of the
experiment's assumptions should be rejected.  That is, falsification is
vague and there can be no crucial experiment.  Logic alone cannot tell
scientists when to give up one theory in favor of another.

"The physicist can never subject an isolated hypothesis to experimental
test, but only a whole group of hypotheses; when the experiment is in
disagreement with his predictions, what he learns is that at least one of
the hypotheses constituting this group is unacceptable and ought to be
modified; but the experiment does not designate which one should be
changed." Duhem, p187

Note: contradictory theories thought to determine between light as waves vs
particles.

Tie in underdetermination: Multiple theories can fit observation.  A theory
(including aux assumptions) can be revised to fit observation, so even
falsification my not be possible.

There are experiments which can be considered crucial, those where the
theory makes surprising hypotheses (the more surprising the better, Popper)
which are then confirmed by evidence.  Newton suggest several such
experiments (trajectory of the commets, shape of the earth, variation of
surface gravity with latitude, perturbation in the motion of Saturn)

"Newton preferred to have what he called, following Hooke, an experimentum
crucis or crossroads experiment to select among alternative physical
possibilities.... surely felt he needed an experimentum crucis separating
universal gravity from inverse-square celestial gravity." Smith, "How did
Newton discover universal gravity?"

Someone (Kuhn or Duhem?) also says that so-called crucial experiments come
long after a theory has been established; this seems to be the case to some
extent with Newton.

Lakatos says there are two kinds of crucial experiments, those that decided
between theories within a programme and those that decide between
programmes.  The latter are seldom recognized as crucial at the time.

There are also crucial experiments in another important sense.  Decide between
theories by appeal to logic, simplicity, promise, pragmatic factors. 
Crucial in convincing scientists to switch the focus of their studies, not
crucial in convincing them that the theories have been definitively decided
between.

[See Duhem p183-190, Duhem commentary in C&C Kuhn 158ish; experiment,
pragmatic in Companion]

*****

10. Over the last few decades increasing effort has been put into computer
simulations and experiments not only in such fields as "artificial life" and
"artificial intelligence," but also in areas of research in physical science
into which theory has not penetrated very far, such as global warming and
the immediate aftermath of the "Big Bang." Stephen Wolfram has gone so far
in A New Kind of Science to suggest that computer simulations and
experiments will likely end the need for theories, in the classical sense,
in future science.  What exactly are computer simulations and experiments? 
In what respects are they, and are they not, replacements for theoretical
investigations?
   
I'm not sure what this is asking.

Talk about the difference between theories and models.

Curve-fitting and simplicity.  We know how to fit curves for a set of data
points, but while doing so may lead to predictive success, it doesn't really
explain anything.  Computer simulations give us a what-answer but not a
why-answer.

[See simplicity, models, theories in Companion?; Blackwell Guide?]
 
*****
                                                                                  
11. In the Preface to his Treatise on Light, published in 1690, Christiaan
Huygens remarks, "One finds in this subject a kind of demonstration which
does not carry with it so high a degree of certainty as that employed in
geometry; and which differs distinctly from the method employed by geometers
in that they prove their propositions by well-established and
incontrovertible principles, while here principles are tested by the
inferences which are derivable from them.  The nature of the subject permits
no other treatment.  It is possible, however, in this way to establish a
probability which is little short of certainty.  This is the case when the
consequences of the assumed principles are in perfect accord with the
observed phenomena, and especially when these verifications are numerous;
but above all when one employs the hypothesis to predict new phenomena and
finds his expectations realized." Assess the adequacy of this description of
the nature of theoretical knowledge in science and of how this knowledge is
achieved.
 
Huygens' description of theory confirmation makes a number of interesting
claims, which I will attempt to separate out.

First, Huygens compares the nature of knowledge in science with that in
math/geometry.  In geometry, principles are certain.  One arrives at the
certainty of a principle by deduction from postulates.  In science,
principles are never established as certain.  No matter how well-established
a principle in science seems to be, it is fallible (may fail).

Second, a principle in science is supported on the basis of the principles
it is derived from, but on the basis of support for other principles which
are deduced from it.  This is the hypothetico-deductive method: Testing a
theory consists of deriving from it consequences that can then be compared
with observations and experimental results.

Third, although we cannot get certainty in science, this
hypothetico-deductive method gets us near certainty.  Huygens explains how:
Principles are assessed according to the H-D method and get degrees of
support.  When the consequences of the principle agree with observation,
they are confirmed.  When they agree with more observations, the principle
is better confirmed.

Finally, Huygens points to a specific kind of deduction: a prediction of new
phenomena.  The agreement of a principle with already-known phenomena is to
be expected, as that might be part of the consideration in devising the
principle.  Multiple principles could be formulated (merely curve fit) for
observational agreement.  (This leads into the problem of local
underdetermination, stated by Duhem and Quine.)

But the agreement of the principle with new, previously unknown, phenomena,
suggests that the principle really has discovered a truth in nature.  This
is a predictionist or accomodationist view: the success of novel predictions
carries the only confirmation or greater confirmation than agreement with
old evidence.  Popper and others have remarked that the "riskier" the
prediction, meaning the less likely the prediction would be if the principle
were not true, the higher the degree of confirmation.  (One reason for this
stems from an understanding that principles can be falsified.  In this way,
an experiment can be crucial in falsifying.  Those that stand up to tests
and are not falsified, seem to be in some way more certain.)

The idea of inferring confirmation or justification of a principle from the
success of its predictions is an interesting one.  It does seem to describe
what scientists do in many instances.  Newton in particular seems to have
agreed with and relied on this principle.  Newton makes several predictions
from his theory of universal gravity (the figure of the earth, variation of
surface gravity with latitude, the trajectory of the comets, etc) and
suggests that observation and experiment will confirm his theory in showing
that his predictions are successful.

Predictionism makes sense to me.  But does it really get us anything near
certainty?  We've already noted that multiple principles can fit the
observation.  Further, the argument for the truth of the principle on the
grounds of the success of its predictions seems to rest on an abductive
inference, something like an explanation to the likeliness of the principle
on the grounds that it is the best explanation for its predictive accuracy.

And now I've run out of things to say and need a conclusion.

The same thing said less coherently:

The first part seems to say that math (deductive) isn't science (inductive). 
This is pretty obvious.  There is no certainty; all theories are fallible.

The second part seems to say that science gets us good enough accuracy
(probability, pragmatic).

The third part seems to say certainty comes (in part) from agreement with
observation (empirical adequacy), the closer the agreement the better.  When
there is more positive evidence (verification) in accord with the
hypothesis, it is more likely or better confirmed.  

Finally, Huygens suggests a higher form of evidence based on predictions. 
This sounds remarkably similar to Newton who suggested that the figure of
the earth and the variation of surface gravity with latitude would better
confirm the theory of universal gravity.  (Should I attempt to tie in a
discussion of crucial experiments?)

Huygens' position sounds a bit like the hypothetical-deductive theory of
confirmation (confirmation of the deductive consequences of a theory is
immediately confirmation for the theory itself).  It also sounds like
predictionism, which places the emphasis for confirmation not on all
deductive consequences, but on novel predictions: A hypothesis is better
confirmed when it predicts new phenomena and tests confirm the prediction. 
(Note Popper's belief that "confirmations should count only if they are the
result of risky predictions".)

Note also that support for predictionism may come from an abductive
inference or inference to the best explanation. The best explanation for the
agreement of the observation with the prediction is that the predicting
theory is true or nearly so.  This is very similar to the argument for
scientific realism from the success of science.  (While intriguing, there
are several problems with the argument, including Fine's objection to the
abductive inference, Laudan's pessimistic meta-induction, and Quine's
underdetermination thesis.  Even though the argument for realism from the
success of science seems to fail, an argument for empirical adeqauacy and
usefulness does not.)


[see evidence in Companion; Popper, Hempel in C&C]

*****

12. The concept of normal science is part of a four-fold distinction Kuhn
introduced: immature science, normal science, science in crisis, and
scientific revolutions.  David Bloor has argued that, even though
comparatively little attention is given to it in The Structure of Scientific
Revolutions, the concept of normal science is the most important idea put
forward in the book.  By contrast, various followers of Karl Popper have
argued that there is no such thing as normal science -- or at least there
ought not to be.  Is there such a thing as normal science, as Kuhn
demarcates it?  Of what importance, if any, is the distinction between
normal science and extraordinary science to the philosophy of science?  Of
what importance is it to historians of science?

I don't know what I'm talking about here.

For Kuhn normal science is what scientists do most of time, working within
an accepted paradigm.  Normal science involves the actualization of the
promise of a paradigm and of the paradigm itself.  Normal science is puzzle
solving; puzzles are solved within the accepted paradigm.  Normal science
involves the expected and tying up loose ends, not unexpected novelty.
Sometimes though scientists are overwhelmed with anomalies and look to
develop alternate paradigms.  This is the crisis that leads to revolution.

Kuhn -- normal science may be the strength of our science; it means that
scientists can concentrate directly on puzzle solving, pushing theory as
far as it can go, without jumping from one theory to another like fads.  "By
focusing attention upon a small range of relatively esoteric problems, the
paradigm forces scientists to investigate some part of nature in a detail
and depth that would otherwise be unimaginable" (Kuhn 24).

Popper -- The best scientists (Bohr, Einsteing) realize the tentative nature
of their conjectures an dexpect that they will be superseded in time. 
Kuhn's verison of normal science would be the end of science because
scientists would accept the theory rather than attempt to falsify it.

Normal science is dogmatic science, when scientists work within a paradigm
and see all puzzles as solvable within the paradigm rather than questioning
the paradigm itself.  Perhaps Popper's concern (and maybe Duhem against
conventionalism as well?) is that scientists should always be willing to
give up their paradigmatic commitments.  The key for Popper is
falsification.  We should always be trying to falsify our theories, even as
we accept them.  We should be actively looking for anomalies.  This is what
happens in extraordinary science.

I don't know anything about Bloor, but maybe he's saying that we always
take various theories for granted, as conventions, and don't try to question
them.

[See Kuhn, Bloor]