I have already mentioned that AOE does not merely
describe scientific practice; it carries implications as to
how scientific practice can be improved. One such implication
concerns scientific revolutions. Kuhn (1970) gives a
brilliant description of the way, during a scientific
revolution, there is a breakdown of rationality, competing
arguments for the rival paradigms being circular, each
presupposing what is being argued for. This is a feature of
actual science. Scientists do not know how to assess
competing theories objectively, when empirical considerations
are inconclusive. But all this can be seen to be a direct
consequence of trying to do science without persisting
metaphysical assumptions concerning the comprehensibility of
the universe, there thus being nothing available to constrain
acceptance of theories when empirical considerations are
inconclusive. Consider Kuhn's breakdown of rationality. A
substantial revolution will involve, not just two rival
paradigms or theories, T1 and T2, but two rival blueprints, B1
lurking behind T1, and B2 lurking behind B2. Granted B1, T1 is
far more acceptable than T2, but the reverse granted B2. But
B1 and B2, being untestable, metaphysical theses, are not
explicitly discussable, and objectively assessable, within
science: so they are more or less repressed, excluded from
discussion. Nevertheless, scientists do think in terms of B1
and B2. Kuhn's Gestalt switch, involved in switching
allegiance from T1 to T2, can be pin-pointed as the act of
abandoning the old blueprint and adopting the new one. Non-
empirical arguments in favour of T1 or T2 can only take the
form of an appeal to B1 or B2, in however a muffled a way (due
to the point that blueprints are not open to explicit
discussion). Such arguments will be circular, and entirely
unconvincing to the opposition, in just the way described by
Kuhn. Accept B1, and T1 becomes the only possible choice;
accept B2 and T2 is the only choice. Each side in the dispute
is convinced that the other side is wrong, even incoherent.
What needs to be done, and cannot be done, of course, is to
discuss the relative merits of B1 and B2. Just this can be
done, granted AOE. T1, B1, T2 and B2 can all be assessed from
the standpoint of adequacy in exemplifying physicalism. When
the scientific community adopts AOE, the Kuhnian irrationality
of revolutions will disappear from science.
It may be asked: How is it possible for AOE to be both
more Popperian than Popper, and more Kuhnian that Kuhn? The
answer is that AOE is more Popperian that Popper in making
explicit, and so criticizable, metaphysical theses which
falsificationism denies, thus leaving implicit and
uncriticizable within science. But AOE is also more Popperian
than Popper in insisting we need to exploit criticism
critically, so that it furthers, and does not sabotage, the
growth of knowledge. Criticism needs to be marshalled and
directed at that part of our conjectural knowledge which it
is, we conjecture, the most fruitful to criticize. This means
directing critical fire at level 2 theories and level 3
blueprints, it being less likely, though still possible, that
criticism of the level 4 thesis of physicalism will aid the
growth of empirical knowledge. Physicalism has played an
extraordinarily fruitful role in the advancement of scientific
knowledge; it should not be abandoned unless an even more
apparently fruitful idea is forthcoming, or unless the
empirical and explanatory success that physicalism appears to
have engendered turns out to be illusory.
8 Imre Lakatos
Lakatos sought to reconcile the very different views of
science held by Popper and Thomas Kuhn. According to Kuhn,
far from seeking falsifications of the best available theory,
as Popper held, scientists protect the accepted theory, or
"paradigm", from refutation for most of the time, the task
being to fit recalcitrant phenomena into the framework of the
paradigm. Only when refutations become overwhelming, does
crisis set in; a new paradigm is sought for and found, a
revolution occurs, and scientists return to doing "normal
science", to the task of reconciling recalcitrant phenomena
with the new paradigm. Lakatos sought to reconcile Popper and
Kuhn by arguing that science consists of competing fragments
of Kuhnian normal science, or "research programmes", to be
assessed, eventually, in terms of their relative empirical
success and failure. Instead of research programmes running
in series, one after the other, as Kuhn thought, research
programmes run in parallel, in competition, this doing justice
to Popper's demand that there should be competition between
theories (a point emphasized especially by Feyerabend).[13]
Lakatos became so impressed with the Kuhnian point that
theories always face refutations, the empirical successes of a
theory being a far more important guide to scientific progress
than refutation, that he finally came to the conclusion that
Popper's philosophy of science was untenable.
AOE has a number of features in common with Lakatos's
methodology of scientific research programmes. AOE makes
extensive use of the notion of scientific research programme.
Like Lakatos's view, AOE exploits the idea that such research
programmes can, sometimes, be compared with respect to how
empirically progressive they are. AOE, again like Lakatos's
view, sees the whole of science as a gigantic scientific
research programme. And like Lakatos's view, AOE can be
construed as synthesizing Popper's and Kuhn's views.
But there are also striking differences. There are
differences in the way scientific research programmes are
conceived, especially research programmes in fundamental
physics. For Lakatos, main components of a research programme
are the "hard core" (corresponding to Kuhn's "paradigm"), and
the "protective belt" of "auxiliary hypotheses", which
facilitate the application of the hard core to empirical
phenomena. The main business of a research programme is to
develop the protective belt, thus extending, and making more
accurate, the empirical predictions of the hard core. The
hard core is a testable theory rendered metaphysical by the
methodological decision not to allow it to be refuted,
refutations being directed at the protective belt rather than
the hard core.
According to AOE, by contrast, the metaphysical kernel of
a research programme is not a testable theory but rather a
thesis that is genuinely metaphysical (i.e. more or less
unspecific, and usually untestable) - a thesis such as the
corpuscular hypothesis, Boscovich's point-atom blueprint,
Einstein's unified field blueprint, and so on. The basic aim
of the programme is to turn the relatively unspecific
blueprint into a precise, testable (and true) physical theory.
The research programme thus consists of a succession of
theories, T1, T2,...Tn, which can be compared, not only with
respect to empirical success, but also with respect to how
adequately each theory encapsulates, or exemplifies, the
blueprint of the programme. (The latter is not possible
within a Lakatosian programme.) Whereas a Lakatosian
programme has a fixed basic theory (or hard core), and seeks
to improve auxiliary hypotheses (the protective belt), an AOE
programme strives to capture the blueprint more and more
adequately by means of testable physical theories.
Both Lakatos's view and AOE permit one to see natural
science as one gigantic research programme, but how this
programme is construed is very different. For Lakatos
"science as a whole can be regarded as a huge research
programme with Popper's supreme heuristic rule: 'devise
conjectures with more empirical content than their
predecessors'" (1970, 132). The huge research programme of
natural science has, for Lakatos, no hard core; to this
extent, Lakatos's view is a variant of Popper's.[14]
According to AOE, however, if natural science is viewed as one
gigantic research programme, then it does have something like
a hard core. First, there is physicalism at level 4, a
metaphysical but nevertheless substantial thesis about the
nature of the universe. And then there is the current
blueprint at level 3, an even more substantial metaphysical
thesis about the nature of the universe. These provide severe
constraints on what theories are acceptable that are not
straightforwardly empirical,[15] something that is not
possible given the views of Popper or Lakatos[16] (or even
Kuhn).
Lakatos and AOE have very different motivations for taking
scientific research programmes so seriously. For Lakatos, the
motivation comes from appreciating that a scientific theory,
T, cannot be decisively refuted at an instant, as it were,
partly because auxiliary hypotheses can always be invented to
salvage T from a refutation, partly because early applications
of a new theory, such as Newton's, may make simplifying
assumptions which may well lead to false predictions (the
fault lying with the simplifying, auxiliary hypotheses rather
than the basic theory). Only by looking at a series of
theories, a given T1 (the hard core) plus changing auxiliary
hypotheses (the protective belt), and comparing this with a
rival series based on a different hard core, T2, and comparing
the extent to which the two series are empirically progressive
or degenerating, can one assess the relative empirical merits
of T1 and T2. For AOE, the situation is very different. A
research programme in theoretical physics consists of a
blueprint, B, and a succession of theories, T1, T2...Tn (each
equivalent to a Lakatosian hard core), successive attempts to
capture B as a testable theory. If T1, T2...Tn are
increasingly empirically successful (in a roughly Popperian
sense) and also increasingly successful at capturing B, then
this means that B is empirically fruitful. A rival blueprint,
B*, might be such that the series T1, T2...Tn moves further and
further away from B*: this would mean that B* is empirically
sterile. A major part of the point of research programmes,
for AOE, is to assess the relative empirical fruitfulness of
rival metaphysical theses, at levels 3 and 4 (and above, if
necessary). Though mostly untestable, nevertheless
metaphysical theses can be assessed in a quasi-empirical way,
in terms of the empirical progressiveness or degeneracy of the
research programmes with which they are associated (or can be
regarded as being associated).[17] This is, according to AOE,
a key feature of scientific method, one which makes scientific
progress possible. It makes it possible for improving
theoretical knowledge to lead to a reassessment of what is the
best available blueprint, which in turn leads to a
reassessment of the best available non-empirical
methodological rules, such as symmetry principles. In other
words, it makes it possible for there to be positive feedback
between improving knowledge and improving aims-and-methods
(improving knowledge-about-how-to-improve-knowledge), a vital
feature of scientific rationality according to AOE.
The differences indicated enable AOE to overcome problems
which Lakatos's view cannot solve. Lakatos insists that there
is no such thing as instant rationality: however apparently
decisive the refutation of a theory may be, it is always
possible to salvage it from refutation in a content increasing
way by the invention of an appropriate auxiliary hypothesis.
It is this consideration which leads Lakatos to argue that
only series of theories, competing research programmes, can be
assessed rationally, in terms of relative empirical
progressiveness. But in practice in science there do seem to
be instant refutations. A famous example is the refutation of
parity. This is a symmetry which declares, roughly, that if a
process can occur, then so can its mirror image. This was
decisively refuted by Wu et al. (1957), by means of an
experiment which showed that electrons were emitted in a
preferential direction from cobalt nuclei undergoing
radioactive decay in a magnetic field. Parity conservation
implied that this would not occur. Strictly speaking, it was
not parity conservation on its own that was refuted, but
parity plus quantum theory plus the theory of weak
interactions plus the theory of nuclear structure plus a
highly theoretical description of the experiment. One would
think there was plenty of scope, here, for auxiliary
hypotheses to be invented to salvage parity from refutation.
No such hypothesis was forthcoming; the refutation of parity
conservation was accepted immediately by the physics
community, despite strong resistance to accepting such a
conclusion (because of the implausibility of supposing that
nature distinguishes between left-handedness and right-
handedness at the level of fundamental physical theory).
Allan Franklin, who has produced what is probably the best
account of the downfall of parity conservation, has put the
matter like this: "It is fair to say that as soon as any
physicist saw the experimental result they were convinced that
parity was not conserved in the weak interactions" (Franklin,
1990, 66).[18] Scientific practice seems almost to refute
Lakatos's view.
But it does not refute AOE. According to Lakatos, in the
end only empirical considerations, plus considerations of
empirical content, restrict choice of theory; few restrictions
are placed on how a body of theory may be modified to salvage
it from refutation. AOE places much more severe restrictions
on choice of theory. In addition to those that it has in
common with Lakatos's view, AOE demands of a fundamental
physical theory that it, together with other such theories,
exemplifies physicalism, to a sufficient degree. This makes
it very much more difficult to modify a body of theory so as
to salvage it from refutation. Instant refutation is not
surprising, granted AOE.
Lakatos's view requires that science consists of
competing research programmes. Unquestionably, the history of
science reveals that competing research programmes have, on
occasions existed. But it is not clear that all science has
this character, as Lakatos's view would seem to require.
After Heisenberg and Schr"dinger had developed quantum theory
in the mid 1920's, there continued to be debate about how the
new theory should be interpreted, and whether the new theory,
interpreted along the orthodox lines advocated by Bohr,
Heisenberg and others, was ultimately acceptable. But there
was nothing like a competing research programme. Viewed from
the perspective of AOE, all this makes perfect sense. There
were indeed serious grounds for regarding the new theory as
unsatisfactory (see Maxwell, 1998, chapter 7). But the new
theory had achieved such striking successes, it was rational
to conjecture that progress lay in developing the new theory,
applying it to new phenomena, reconciling it with special
relativity - in doing something like Kuhnian normal science,
in other words - rather than in trying to develop a rival
theory, a rival research programme. (To say this is not to
say that serious attention should not have been given to the
theoretical defects of orthodox quantum theory.) Not only
does the history of science fail to reveal that there are
always competing research programmes; whenever a new theory
arrives on the scene that meets with extraordinary empirical
success and no refutation, no good rationale may exist for
inventing a rival research programme. (As we have seen,
unlike Popper's falsificationism and Lakatos's research
programme view, AOE holds that something like Kuhn's normal
science may well be rational, as long as it is accompanied by
some sustained tackling of problems associated with the
currently accepted blueprint. This may, eventually, but not
immediately, lead to the development of a new fundamental
theory, a new research programme.)
There are other, much more decisive ways in which AOE is
an improvement over Lakatos's view. Lakatos's methodology of
research programmes inherits a number of unsolved problems
from its two sources, Popper and Kuhn. Like Popper and Kuhn,
Lakatos has no solution to the problem of what the simplicity,
unity or explanatory character of a theory, or hard core, is;
AOE, as I have indicated briefly above, solves the problem
without difficulty. In failing to say what simplicity is,
Lakatos also fails to articulate with any precision that part
of scientific method concerned with simplicity; AOE faces no
difficulty here either. Like Popper and Kuhn, Lakatos can say
nothing useful about how new theories, new hard cores, are
created or discovered; AOE, as a result of including levels 3
and 4 within the domain of scientific knowledge, is able to
specify a rational, if fallible and non-mechanical, method for
the creation of new theories, even new fundamental theories of
physics. Finally, Lakatos's view fails to solve the problem
of verisimilitude, a problem which can be readily solved
granted AOE.
Popper, Kuhn and Lakatos, despite their differences, have
one big failure in common (the source of almost all the
others). All three take it for granted:
(A) In science no untestable but nevertheless substantial
thesis about the world can be accepted as a part of scientific
knowledge in such a firm way that theories which clash with
it, even if highly successful empirically, are nevertheless
rejected.
Popper accepts (A) in that, for him, untestable theses
are metaphysical, and therefore not a part of scientific
knowledge. Kuhn holds it, because, for Kuhn, nothing
theoretical survives a revolution. Kuhn's acceptance of (A)
is also apparent in his whole treatment of revolutions:
precisely because Kuhn accepts (A), Kuhn cannot invoke
anything like the level 4 thesis of physicalism to assess
rival paradigms during a revolution, when empirical
considerations are inconclusive. The Kuhnian irrationality of
revolutions is a consequence of scientists accepting (A); and
in so far as Kuhn thinks this irrationality is inevitable,
Kuhn accepts (A) as well.
A case could be made out for saying that Lakatos came
near to rejecting (A) in arguing for the need for science to
adopt a conjectural metaphysical inductive principle which, if
true, would more or less guarantee that Popperian, or rather
Lakatosian, methods deliver authentic theoretical knowledge.
But Lakatos here missed the fundamental point, central to
AOE, and highly Popperian in spirit, that our current methods
are all too likely to be more or less the wrong methods to
adopt, the metaphysics implicit in these methods being false,
there thus being a vital need, for scientific progress, to
make the metaphysics explicit so that it can be criticized, so
that alternatives can be developed and considered, leading to
improved metaphysics and methods, this in turn requiring the
development of a hierarchy of metaphysical theses to form a
framework of relatively unproblematic theses within which more
specific problematic theses may be developed and assessed.
Interestingly enough, Lakatos himself was aware of this
deficiency in his "plea to Popper for a whiff of
'inductivism'" (1978, 159). Discussing his proposal that one
should appeal to a metaphysical inductive principle as a
conjecture as a part of the solution to the problem of
induction, Lakatos says:
"Alas, a solution is interesting only if it is embedded in,
or leads to, a major research programme; if it creates new
problems - and solutions - in turn. But this would be
the case only if such an inductive principle could be
sufficiently richly formulated so that one may, say,
criticize our scientific game from its point of view. My
inductive principle tries to explain why we 'play' the game
of science. But it does so in an ad hoc, not in a 'fact-
correcting (or, if you wish, 'basic value judgment
correcting') way" (Lakatos, 1978, 164).
Lakatos highlights, here, the difference between his own
position and that of AOE. The (revisable) AOE thesis of
physicalism is indeed "sufficiently richly formulated so that
one may...criticize our scientific game from its point of
view". AOE not only offers a new research programme for the
philosophy of science; it modifies the research programme of
science, one modification being that the philosophy of science
becomes an integral part of science itself. The passage above
makes me wonder whether Lakatos might not have gone on to
develop or endorse AOE if he had lived.
References
Cartwright, N. (1999) The Dappled World, Cambridge University
Press, Cambridge.
Dupr, J., (1995) The Disorder of Things, Harvard University
Press, Cambridge, Mass.
Feyerabend, P., (1965) Problems of Empiricism I, in C.
Colodny, ed., Beyond the Edge of Certainty, University of
Pittsburgh Series in the Philosophy of Science, vol. 2,
Prentice Hall, Englewood Cliffs, NJ, 145-260.
________ (1975) Against Method, Verso, London.
Franklin, A., (1986) The Neglect of Experiment, Cambridge
University Press, Cambridge.
________ (1990) Experiment: Right or Wrong, Cambridge
University Press, Cambridge.
Glymour, C., (1980) Theory and Evidence, Princeton University
Press, Princeton.
Holton, G., (1973) Thematic Origins of Modern Science, Harvard
University Press, Cambridge, Mass.
Hooker, G. A. (1987) A Realistic Theory of Science, State
University of New York Press, Albany.
Howson, C. and Urbach, P. (1993) Scientific Reasoning, Open
Court, La Salle.
Hull, D. L., (1988) Science as a Process, University of
Chicago Press, Chicago.
Kitcher, P. (1993) The Advancement of Science, Oxford
University Press, New York.
Kuhn, T. S., (1970) The Structure of Scientific Revolutions,
University of Chicago Press, Chicago.
Lakatos, I., (1970) Falsification and the Methodology of
Scientific Research Programmes, in Lakatos and Musgrave
(1970), 91-195.
________ (1978) The Methodology of Scientific Research
Programmes, ed. J. Worrall and G. Currie, Cambridge
University Press, Cambridge.
Lakatos, I. and Musgrave, A., eds., (1970) Criticism and the
Growth of Knowledge, Cambridge University Press, London.
Laudan, L. (1984) Science and Values, University of California
Press, Berkeley.
Maxwell, N. (1972) A Critique of Popper's Views of Scientific
Method, Philosophy of Science 39, 131-152.
________ (1974) The Rationality of Scientific Discovery,
Philosophy of Science 41, 123-153 and 247-295.
________ (1979) Induction, Simplicity and Scientific Progress,
Scientia 114, 629-653.
________ (1984) From Knowledge to Wisdom, Blackwell, Oxford.
________ (1993) Induction and Scientific Realism: Einstein
versus van Fraassen, The British Journal for the Philosophy
of Science 44, 61-79, 81-101, and 275-305.
________ (1997) Must Science Make Cosmological Assumptions if
it is to be Rational, in The Philosophy of Science:
Proceedings of the Irish Philosophical Society Spring
Conference, ed., T. Kelly, Irish Philosophical Society,
Maynooth, 98-146.
_________ (1998) The Comprehensibility of the Universe, Oxford University Press, Oxford.
________ (1999) Has Science Established that the Universe is
Comprehensible?, Cogito 13, 139-145.
______ (2000a) Can Humanity Learn to become Civilized? The
Crisis of Science without Civilization, Journal of Applied Philosophy 17, 29-44.
______ (2000b) A new conception of science, Physics World 13,
17-18.
______ (2001) The Human World in the Physical Universe,
Rowman and Littlefield, Lanham, Maryland.
______ (2002a) Is Science Neurotic? Metaphilosophy 33, 259-
299.
______ (2002b) The Need for a Revolution in the Philosophy of
Science, Journal for General Philosophy of Science 33,
1-28.
______ (2002c) Karl Raimund Popper, in BRITISH PHILOSOPHERS,
1800-2000, edited by P. Dematteis et al., Bruccoli
Clark Layman, Columbia.
McAllister, J. W. (1996) Beauty and Revolution in Science,
Cornell University Press, Ithaca.
Musgrave, A. (1993) Common Sense, Science and Scepticism,
Cambridge University Press, Cambridge.
North, J. (1965) The Measure of the Universe, Clarendon
Press.
Popper, K. (1959) The Logic of Scientific Discovery,
Hutchinson, London.
________ (1961) The Poverty of Historicism, Routledge and
Kegan Paul, London.
________ (1963) Conjectures and Refutations, Routledge and
Kegan Paul, London.
________ (1970) Normal Science and its Dangers, in Lakatos and
Musgrave (1970), 51-58.
________ (1972) Objective Knowledge, Oxford University Press,
London.
________ (1974) Replies to Critics, in P. A. Schilpp, ed., The
Philosophy of Karl Popper, Open Court, La Salle, Illinois,
vol. 2, 961-1197.
________ (1982) Quantum Theory and the Schism in Physics,
Hutchinson, London.
________ (1983) Realism and the Aim of Science, Hutchinson,
London.
van Fraassen, B. (1980) The Scientific Image, Clarendon Press,
Oxford.
Watkins, J. (1984) Science and Scepticism, Princeton
University Press, Princeton.
Wu, C. S. et al, (1957) Physical Review 105, 1413.
Notes
[1] See Maxwell (1998). This improves considerably on earlier
versions of the view expounded in Maxwell (1972, 1974, 1979,
1984, 1993 and 1997). For summaries of (1998) see Maxwell
(1999, 2000).
[2] See Popper (1959, 1963, 1982).
[3] See Lakatos (1970, 1978).
[4] For Popper's replies to such criticisms: see Popper
(1972), chapter 1; (1974), sections II and III; and (1982),
Introduction and chapter 1.
[5] Popper discusses such "silly" rival theories in Popper
(1983, 67-71). He argues that they deserve to be rejected on
the grounds that they create more problems than they solve,
problems of explanation. This is a relevant consideration
granted dressed falsificationism, but not granted bare
falsificationism. He also argues that it does not matter if
such "silly" theories become potential rivals, since it can be
left to scientists themselves to criticize them. But what
this ignores is that it is precisely Popper's methodology
which should be providing guidelines for such criticism. Far
from condemning such a "silly" theory as worthy of rejecting,
bare falsificationism holds such a theory to be better than
the accepted theory (if it has greater empirical content, is
not falsified where the accepted theory appears to be, and
some of the excess content of the "silly" theory is
corroborated). Popper fails to appreciate that it his
methodology, not he himself, which needs to declare that silly
theories are indeed "silly". The fact that his methodology
declares these silly theories to be highly acceptable is a
devastating indictment of his methodology. To argue that
these silly theories, refuting instances of his methodology,
do not matter and can be discounted, is all too close to a
scientist arguing that evidence, that refutes his theory,
should be discounted, something which Popper resoundingly
condemns. The falsificationist stricture that scientists
should not discount falsifying instances, ought to apply to
methodologists as well!
[6] In fact even the methodological rules of bare
falsificationism are such that persistent application of these
rules commits one to making implicit metaphysical assumptions
(which may be false). Bare falsificationism, as formulated by
Popper, requires of an acceptable theory that it is strictly
universal in that it makes no reference to any specific time,
place or object. This makes it impossible for science to
discover that the laws of nature just are different within
specific space-time regions, or that there is a specific
object with unique dynamical properties. There is no scope,
within bare falsificationism, for the rejection of these
metaphysical theses, even though circumstances could
conceivably arise such that progress in knowledge would
require this. (AOE, by contrast, allows for this remote
possibility: that which is dogmatically upheld by bare
falsificationism becomes criticizable granted AOE.) Popper
recognizes that the methodological rule requiring any theory
to be strictly universal does have a metaphysical counterpart
(1959, sections 11 and 79), but fails to appreciate how
damaging this is for falsificationism.
[7] In holding that metaphysical theses and philosophies of
science are an integral part of science itself, AOE implies
that Popper's principle of demarcation (Popper, 1963, chapter
11) is to be rejected. Popper's demarcation proposal, apart
from being untenable, is in any case too simplistic, in that
it reduces to one a number of distinct demarcation issues.
Popper rolls into one the distinct tasks of demarcating (a)
good from bad science, (b) science from non-science, (c)
science from pseudo-science, (d) rational from irrational
inquiry, (e) knowledge from mere speculation, (f) knowledge
from dogma (or superstition, or prejudice, or popular belief),
(g) the empirical from the metaphysical, and (h) factual truth
from non-factual (analytic) truth. (a) to (d) involve
demarcating between disciplines, whereas (e) to (h) involve
demarcating between propositions.
[8] Dynamical theories are partially ordered with respect to
the extent that they exemplify physicalism, with respect to
their degree of unity, in other words. For further details
see Maxwell (1998), chapter 4.
[9] For a very much more detailed exposition of this solution
of the problem of simplicity, together with an account of the
way in which great unifying theories of physics illustrate the
solution, see Maxwell (1998), chapters 3 and 4.
[10] Corresponding to the level r thesis there is the level r
aim to turn the thesis into a precise, testable, true theory;
methods associated with this aim are those associated with the
corresponding thesis. Revising the thesis involves revising
the associated aim. As theses become increasingly unspecific
and unproblematic, as one ascends the hierarchy, so the
corresponding aims and methods become increasingly unspecific
and unproblematic.
[11] It may be objected that if T is assumed to be the true
unified theory of everything, no meaning can be given to the
idea that theoretical physics is making progress, by means of
a succession of false theories, to a more or less disunified
theory of everything. But T does not need to be assumed to be
unified; all that is required is that T is such that the
notion of "partial derivation" from T makes sense. For
further discussion of the inability of any standard empiricist
view such as falsificationism to solve the problem of
verisimilitude, and the ability to AOE to solve the problem,
see Maxwell (1998), 70-72, 211-217 and 226-227.
[12] For further discussion of the method of discovery
provided by AOE see Maxwell (1974), Part II; (1993), Part III;
and (1998), 159-163 and 219-223.
[13] See Lakatos (1970, 1978). For Feyerabend's argument that
severe testing requires the development of rival theories see
Feyerabend (1965).
[14] Granted Lakatos's overall view, the research programme of
science cannot have a hard core, for then, in order to ensure
Popperian severe testing, there would need to be a rival
research programme with a rival hard core - and that would
mean the original research programme was not the whole of
science. Actually, Lakatos is not quite consistent here;
after the sentence quoted in the text, Lakatos goes on "Such
methodological rules may be formulated, as Popper has pointed
out, as metaphysical principles. For instance, the universal
anti-conventionalist rule against exception-barring may be
stated as the metaphysical principle: 'Nature does not allow
exceptions'" (1970, 132). That this admission is damaging for
Popper's bare falsificationism was pointed out in footnote 6;
it is equally damaging for Lakatos's version of Popperianism.
[15] I say "not straightforwardly empirical" because both
physicalism and the best available blueprint are themselves
accepted on the grounds that they support a more empirically
progressive research programme than any rival theses. Long-
term empirical considerations influence choice of theses at
levels 3 and 4, while at the same time these theses can lead
to the rejection of potentially empirically successful
theories that clash too severely with them (i.e. are too
severely ad hoc).
[16] The Popperian and Lakatosian demand that theories be
strictly universal places weak but rigid constraints on what
theories are acceptable; the demand of AOE that theories
accord, as far as possible, with physicalism and the best
available blueprint, places strong, but flexible and revisable
constraints on what theories are acceptable. (For further
discussion see Maxwell, 1998, 89-102, chapter 4, and 223-227.)
[17] For further details and discussion, see Maxwell (1998),
172-180.
[18] For an account of the discovery of parity non-
conservation, and of the decisive character of the experiments
refuting parity conservation, see Franklin (1990), 63-6 and
151-2. See also Franklin (1986)._
