Does science have a
method, or a collection of methods?  In
this book, Achinstein has culled from a variety of scientists and philosophers
various "rules for constructing and testing scientific theories"
(1).  The five parts into which the book
is divided each offer at least one scientific example — all from physics —
together with exegesis or criticism of the rules or mode of argument advocated
there.  The editor, who published a
large-scale analysis of evidence in 2001 (The Book of Evidence, Oxford
University Press), lets slip his own predilections in some of these case
studies, but in general stays above the fray.

The first part of the
book might seem somewhat out of place, a discredited rationalist museum-piece
in comparison with the others.  It is
concerned with Descartes’ "rules for the direction of the mind" in trying
to understand the world and with some of the conclusions Descartes arrived at
by following them, at least in part. 
Achinstein offers the first thirteen rules from the posthumously
published treatise, the ontological argument for the existence of God that
Descartes gave in the Fifth Meditation, and then Descartes’ formulation
of the laws of motion in the Principles of Philosophy.   By way of critical commentary, he then
gives us some material from Daniel Garber’s 1992 book, Descartes’
Metaphysical Physics. 

The material from
Descartes is interesting: one can see how far he could get by appeal, not to
the behaviour of things in the world, but to the perfection of an unchangeable
god and his own rational intuition of fundamental concepts.  As Achinstein notes (12), Descartes was able
to reject, on these grounds, the common-sense notion that moving objects need a
force to keep them moving, and to arrive at an "ancestor" of the
conservation of momentum.  One can also
see similarities between Descartes’ theologically loaded arguments and agnostic
appeals to simplicity or elegance that we continue to hear.  But the material also reveals the Achilles’
heel of rationalism: Descartes took one of the fundamental concepts, quantity
of motion, to be volume times speed, while Newton was luckier with his momentum,
mass times velocity.  You can intuit as
keenly as you can, but if you happen to pick the wrong, or less fruitful,
concept, your intuitions are so much waste of effort.  Reason seems not to guide you where it really matters.

While few would now try
to base their physics on intuitions about the perfection of god, the next four
parts of Achinstein’s collection deal with issues that are still current, at
least among philosophers of science. 
Proceeding in mainly chronological order, Achinstein next gives us
Newton: his four  "rules for the
study of natural philosophy" from the second and third editions of Principia,
the "deduction" of universal gravitation from various phenomena —
arguments that explicitly invoke the four rules — and parts of the General
Scholium, all from Book 3.  These are
commented on by I. Bernard Cohen.  This
part concludes with some of Whewell’s criticisms of the overemphasis on
observation he found embodied in Newton’s rules.  Achinstein recurs to Newton’s rules later in the book,
particularly since he wishes to endorse the idea in the fourth rule that we
ought not to trouble about the fact that evidence can be explained by
alternative hypotheses, at least when we have another theory that has been "deduced"
in Newton’s way from the phenomena to be explained.  It is only when we find recalcitrant phenomena that we need to
revise our theoretical commitments. 
Newton’s "deduction from the phenomena" is a matter of
inferring like causes from the observation of like effects and of the inductive
move from the observationally based truth that these A are B to all A
are B — as such, Newton’s terminology rings the wrong bells for us since
both kinds of inference are deductively invalid, but, as the fourth rule shows,
Newton recognized their fallibility. 
Achinstein suggests that when Newton himself invokes hypotheses (i.e.,
claims not so "deduced" from phenomena) he does not think their
successes can be taken as evidence of their truth, but merely of their fruitfulness.

The central and largest
part of the book deals with nineteenth century arguments in favor of the wave
theory of light, using one of Young’s 1807 lectures as the focal point.  Achinstein gives considerable space to a
slightly later debate between Whewell and J.S. Mill on the
hypothetico-deductive method, which saw wave theorists as committed to this
particular type of argument, inserting a small chunk of Popper’s more extreme
anti-inductivism between them (the Mill selection includes his four methods for
investigating causal relationships).  
This part concludes with Achinstein’s own evaluation of the bearing of
the Whewell-Mill debate on arguments for and against the wave theory.  Straw-man hypothetico-deductivism would say
that because hypothesis P deductively entails observations Q, we have evidence
that P is true.  In supporting Newton’s
fourth rule, Achinstein has already indicated a conspicuous flaw in such a line
of argument: any number of conflicting Ps can deductively entail Q, so to be
taken seriously hypothetico-deductivism has to say more.  Whewell offered three extra ideas: P should
not only entail the already known Q but also predict not yet known R; it should
entail true predictions of a different kind from Q; and it should be part of a
sequence of hypotheses that lead to more unified and simpler theories. Mill’s
fundamental objection is that these features are still not enough to argue for
the truth of P — some other hypothesis, perhaps not yet framed or even
framable given where we have reached, might do these jobs as well or
better.  As Achinstein observes, however
better the wave theory was doing by mid-century than the particle theory, it
turned out that Mill’s possibility was actualized by the development of quantum
mechanics. 

Given its continuing
popularity, it is also worth noting Mill’s unimpressed response to the argument
from the prediction of new kinds of effects:

If the laws of
the propagation of light accord with those of the vibrations of an elastic
fluid in as many respects as is necessary to make the hypothesis afford a
correct expression of all or most of the phenomena known at the time, it is
nothing strange that they should accord with each other in one respect more….
[I]t would not follow that the phenomena of light were results of the laws of
elastic fluids, but at most that they are governed by laws partially identical
with these; which, we may observe, is already certain, from the fact that the
hypothesis in question could be for a moment tenable.  (223-4)

Part IV, built around
one of Perrin’s papers summarizing work on atoms and molecules, focuses on the
dispute between scientific realism and anti-realism.  It begins with extracts in support of anti-realist views from
Duhem and Van Fraassen, then gives us some pages from Perrin’s Atoms,
followed by Wesley Salmon and Achinstein arguing that Perrin has found an
experimental argument for the reality of atoms.  One of Achinstein’s important caveats is that we do not have a
general argument here for realism about all theoretical entities (an argument
we ought not to want, since it would burden us with all the exploded
theoretical entities proposed so far), but an empirical argument for the
particular case of atoms and molecules. 

The last part of the
book uses the Tower argument, a passage from Galileo’s Dialogue Concerning
the Two Chief World Systems, to inform a contrast between Feyerabend’s
rejection of methodological rules (other than his own "anything
goes") and Achinstein’s own response. 
Achinstein carefully sets out what sensible positions might be covered
by Feyerabend’s calls to proliferate hypotheses inconsistent with what we take
ourselves to know, and in passing offers a different reconstruction of what
Galileo is doing in the Tower argument — suggesting that on both theories of
the universe we would expect to see stones falling straight down a tower, but
that the explanations given of this observation will be very
different.  Feyerabend had taken the
presuppositions of a stationary earth to have insinuated themselves much more
deeply in our natural interpretation of experience.  The same section of the book continues with a selection from
Kuhn, followed by another of Achinstein’s own pieces that sets out to delineate
various senses in which Kuhn might hold that scientific judgments of evidential
worth are subjective.

Overall, then,
Achinstein seems to have been guided somewhat more by philosophical than
scientific questions.  I’m not sure it
gives one any kind of definitive answer to the question I started with.  We do have Newton’s rules, and Mill’s
methods, and the modes of argument invoked by Young and Perrin — both of
which, Achinstein argues, are somewhat more complex than their philosophical
commentators have recognised — but the intricacies of what "induction"
should cover (debated at length by Whewell and Mill), or the issues of realism
versus anti-realism, or Feyerabend and Kuhn’s rejections of stable
methodological precepts, seem to be issues for philosophers rather than
practitioners.   

Achinstein’s volume
weighs in at 427 pages so one hesitates to suggest omissions, but I regretted
the absence of matter related to debates on the universality and scope of
methodological rules (for instance, that between Laudan and Worrall, some of
which occupied the pages of the British Journal for the Philosophy of
Science around 1988-1990).  The
question of scope connects with the feeling one might well get from looking at
the rules actually invoked here that these are so general as hardly to warrant
mention.  Neophyte scientists need to
learn fairly complex statistical methods for analyzing experimental results,
but these are at a level of detail far removed from Mill’s methods or Newton’s
rules.  Again, they need to learn that
in medical trials, but not in testing bridges, you need to use double-blind
procedures.  There is philosophical
point in noting that these actual ways of testing, if not constructing,
scientific theories can be seen as fitting into general logical patterns, but
the scientists themselves may not be too impressed by the general patterns,
since the empirical detail is also clearly salient.   Another topic that might have been covered — it certainly got a
lot of exposure in the Lakatosian school of philosophically sophisticated
history of science — is the nature of ad hoc modifications of theory,
an issue raised by Whewell’s third addition to the hypothetico-deductive
method.  But to return to what we do
have, Achinstein has given us a series of stimulating debates that, as he hopes,
should provoke new ways of presenting some of the standard topics in the
philosophy of science and encourage greater attention to the modes of argument
actually employed by practitioners.

© 2005 Ed Brandon

Ed Brandon is, by training, a philosopher, and now is working in a policy position in the University of the West Indies at its Cave Hill Campus in Barbados.

Categories: Philosophical