[lit-ideas] Ayer on Popper
- From: Donal McEvoy <donalmcevoyuk@xxxxxxxxxxx>
- To: "lit-ideas@xxxxxxxxxxxxx" <lit-ideas@xxxxxxxxxxxxx>
- Date: Mon, 4 Jun 2012 23:45:45 +0100 (BST)
JLS' references to the kind of comments academics may make of each other, in
sometimes quite personal terms, may be of interest - but here I turn to what
they make of each other in terms of their work:-
Ayer’s Philosophy in
the Twentieth Century gave Ayer the opportunity to put Popper in his place
or give him his due. The result? Take a wild, non-inductive guess.
It consists, second of all, of a dismissive reference to P’s Objective
Knowledge on p.200: …“except
in the Platonic tradition which Sir Karl Popper has attempted to revive,
theories do not exist apart from those who hold them” – where the attempt “to
revive” might seem to express an underlying assumption that a theory of
‘objective
knowledge’ is something of a philosophical corpse. Though remarked en passant,
this is both glib and hardly
plausible when we consider that theories have characteristics as ‘objects of
thought’ that are logically independent of the thinking of those who hold them
[for example, contradiction between two theories may hold independent of
whether anyone realises they contradict]. Popper’s Objective Knowledge
emphasises that knowledge can be considered in
terms of its ‘objective’ characteristics, that are not reducible to subjective
individual psychology. But is it thus in the “Platonic tradition”? Popper’s
view of ‘objective knowledge’ is antithetical in several fundamental respects
to Platonic tradition. Does P hold theories arise and exist and may be
developed in ways that are entirely independent of “those who hold them”? Not
at all – P is a three-worlds interactionist, and the interaction of Worlds 123
involves
the existence of all three worlds: while World 1 could exist without Worlds 2
and 3, and World 2 could exist without a World 3 [but not without World 1],
World 3 depends on the existence of World 2 and World 1. In P’s account,
‘objective
knowledge’, in its World 3 sense, arises from World 2 activity – and in this
sense ‘objective knowledge’ depends for its development and growth on “those”
who produce and consume ‘objective knowledge’.
But as perhaps we find from Language, Truth and Logic onwards, we should not
rely on Ayer for a
logically accurate and incisive account of Popper’s work.
The other reference to P, and first of all, is the section
pp.131-4 Karl Popper on Induction.
This begins by positioning P within the framework of the Vienna Circle as P’s
Logik der Forschung first appeared in a
series edited by Schlick and Frank (as P later dryly commented, “The Circle was
a Circle of politics”). Ayer goes on to say, “The appearance of Popper’s book
in the series was notable for several reasons”, but stops short of explaining
that P’s book contains the chief arguments that led to Logical Positivism
falling from favour (though others might credit the later Wittgenstein here).
Nor
does Ayer emphasise how antithetical to Logical Positivism is P’s position.
Nevertheless the reasons given, as to why it is “notable”
that P was published ‘by the Circle’, turn out to be that P’s position takes
issue with the Circle’s central tenets. P did not view non-scientific
statements as meaningless or even valueless; P argued against distinguishing
science from metaphysics by a criterion of meaning; and P argued the criterion
lay in the falsifiability of statements by observation rather than their being
verifiable.
Having acknowledged as “valid and important” P’s
“distinction” between verifiability and falsifiability [i.e. that while, as
Ayer writes, “positive instances” do not “fully
establish a generalization”, “one negative instance can definitely refute it”],
Ayer contends this “distinction… in practice…is not so clear-cut as it might at
first appear.” Ayer does not amplify what specific thesis he is proposing here
nor whether it is a thesis that undermines P’s position: we might think Ayer
means this as a criticism, but we cannot say that Ayer says so. P’s position in
Logik der Forschung is that the
logical asymmetry between falsifiability and verifiability is “clear-cut” qua
logical asymmetry – it is as clear-cut as manys a logical distinction. What is
not “so clear-cut”, because it is not
simply a matter of logic, is how we decide whether a theory or statement is
falsified. This of course does not mean that the “distinction…is not so
clear-cut” but rather that applying a
falsificationist methodology “in practice..is not so clear-cut” (not “so
clear-cut” as the logical distinction between falsifiability and verifiability
on which that methodology is based), and this is because its application
involves more than clear-cut logic. Ayer
does not make clear whether he means anything more than this by his somewhat
opaque prose.
Nevertheless Ayer goes on to give an account of scientific
procedure. Unfortunately, Ayer is not entirely clear whether this account is
what Ayer claims to be P’s account or whether it is instead a correct account.
As it happens, what Ayer offers is neither P’s account nor a correct account.
Those interested in academic credentials might consider that
Ayer speaks with the authority of someone who for almost two decades was
Wykeham Professor of Logic in the University of Oxford. P was formerly
Professor of Logic and Scientific Method at the London School of Economics. We
might think their professorships would have trained them to be logically clear
and
correct. P claims that there is a logical assymetry between the falsifiability
and the verifiability of a “generalisation”, and indeed in logical terms this
asymmetry is clear-cut. But Ayer is suggesting, it would seem, that “in
practice” this “distinction” is “not so clear-cut.” How so?
What we find is that Ayer, having seemed to make this claim,
does not substantiate this claim at all. Instead Ayer continues by asserting,
“The first question that arises is what is to constitute falsification?” But if
the answer to this “first question” is to substantiate that the asymmetry
between falsification and verification is “not so clear-cut”, we might imagine
that is because what constitutes a “falsification” depends on something akin to
verification – for example, it depends on verification of the “negative
instance”. But Ayer does not make clear whether or not he agrees with this, but
proceeds along lines where it is very hard to extract from Ayer’s claims how
exactly – especially in logical terms – he thinks the asymmetry is “in
practice…not so clear-cut.” Instead Ayer proceeds to make several claims that
are simply false or misleading as to P’s account of “what is to constitute
falsification?”
Having raised the “first question”, Ayer claims “Popper’s
answer to this was that a theory or a hypothesis is falsifiable if it is
logically incompatible with some set of basic propositions.” This is true as
far as it goes, though it is only part of Popper’s account of “what is to
constitute falsification”, which extends to explaining how “basic propositions”
[or, in P’s terminology, “basic statements” or “test statements”] may be
accepted without being logically verified.
Ayer continues by claiming, “It [i.e. a theory] is falsified if in conjunction
with one or more
accepted basic propositions it entails the negation of an accepted basic
proposition”. Though a former Wykeham Professor of Logic, and though P’s Logik
der Forschung is a logical
analysis of scientific method in terms of different kinds of statement, Ayer
has here decided to depart from P’s own logical terminology for reasons that
are not explained (and which departure could hardly be discerned except by a
reader who knew the contents of LdF).
This departure is only apt to confuse the issues as P’s own terminology is
exemplary
in its logical clarity whereas Ayer’s mode of expression is not so logically
clear or correct.
In the key terminology of LdF, we may distinguish a ‘universal law’ [‘UL’],
‘initial conditions’ [‘ICs’]
and a ‘basic statement’ [‘BS’]. A UL could be a theory like “All swans are
white”. If we take this UL in
conjunction with some relevant ICs,
such as ‘Here is a swan’, then we may deduce that ‘Here is a white swan’. The
statement ‘Here is a white swan’ is a BS that may be checked by observation in
that it may be falsified by observation.
If that BS is falsified by
observation – if we observe instead a non-white swan here – then we may deduce
that the UL is false (indeed if we
observe a non-white swan, and thus accept a concomitant BS inconsistent with
the UL,
we falsify the UL without use of ICs).
Ayer’s form of expression is at best a clumsy approximation
of this: “It is falsified if in conjunction with one of more accepted basic
propositions it entails the negation of an accepted basic proposition”. Here
Ayer’s “basic propositions” seems to denote (and therefore conflate) both ICs
and BSs, even though ICs and BSs are logically distinct given their
role in scientific reasoning.
Ayer continues, “What, then, is a basic proposition, and
when is such a proposition to be accepted? The answer is that a basic
proposition is one that assigns some observable physical property to some
region of space-time and that is accepted by fiat.” This account of a “basic
proposition” [or, in P’s logically more exact terminology, BS] is largely
unobjectionable except for the unargued conclusion that
“that is accepted by fiat”. Exactly what “is accepted by fiat”? It seems
clear that Ayer does not mean merely that this definition of a BS must be
accepted by fiat but that our
acceptance of a BS – as being true –
is “by fiat”. For Ayer continues, “One’s acceptance will indeed be motivated by
some sense-experience…” so that it is a clear that it is the acceptance of the
truth of a BS, and not the definition
of a BS, that is being talked about.
The result is at once incredible, false, and not at all P’s position. P’s
position is that the acceptance of a BS depends on its being checked by
observation – if observation bears out the
truth of a BS it may be accepted and
if observation falsifies a BS it may
be rejected (though there is more to be said than this, this is nevertheless
essential and is why a BS is defined so that it must posit an observable state
of affairs). It is absolutely key to the empirical character of science that
the acceptance of BSs is not simply
“by fiat” but because their truth has been checked by observation. To say a BS
“is accepted by fiat” is to strip BSs of their scientific character, their
falsifiable character – for how could one ever falsify what “is accepted by
fiat”? If we are to take Ayer
seriously here, science is not based on what may be tested by observation
[contra P] but on what “is accepted by
fiat”, so that science appears to be simply a form of dogmatism without proper
critical and empirical underpinning.
Next is a paragraph that mentions the need for “auxiliary
hypotheses” to render “theories [containing]…terms for unobservable objects” so
that they are “logically inconsistent with some set of basic propositions.” The
paragraph further explains that if probabilistic theories are to be falsifiable
there must “be a convention that a statement of probability was to count as
being falsified if, at a certain stage, the recorded frequency differed from
the predicted frequency by more than an agreed amount.”
In the next paragraph, Ayer continues: “The upshot of these
various factors is that a theory can always be protected from falsification,
whether by one’s making a different assessment of probability, or by one’s
rejecting
some auxiliary hypothesis, or even by one’s refusing to accept some basic
proposition, however much favour it may have found with other judges. Popper
admits that this is so and has to fall back on saying that someone who so
cherishes his theories that he makes them immune from falsification is simply
not playing the scientific game.” This is not inaccurate but might be better
explained:
while P admits a falsification can always be evaded, P suggests that to learn
from observation we need to give observation a role where it may be decisive
against
a theory by way of its falsification – and insofar as we might sometimes
properly
‘evade’ a threatening falsification this should be consistent with a
falsificationist methodology, for example by specifying an explanation for the
apparent falsifying counterinstance that is itself falsifiable [as in the
famous case of explaining a planet’s deviation from its predicted orbit by
specifying a hitherto undiscovered but observable planet whose gravity would
explain the deviation].
This paragraph concludes, “There never comes a point where a
theory can be said to be true.” This should be amended – there never comes a
point where a theory can be said to be proved true (but it may be said to be
true, and be true, notwithstanding this absence
of proof by verification). “The most that one can claim for any theory is that
it has shared the successes of all its rivals and that it has passed at least
one test which they have failed.” This is not “the most that one can claim” –
for starters, it is to claim more than this that in addition the theory has
passed all known tests. And we may be able to claim that a theory has passed
all known tests.
Having tried to put Popper’s LdF in its place as an account of scientific
method, Ayer wants to
put it further in its place philosophically by showing that it does not solve
“the
problem of induction” – “A startling claim which Popper and his adherents made
for his account of scientific procedure is that it solves the problem of
induction.” What Ayer does not make clear is that P’s is not a positive
solution to “the problem of
induction” and if we are demanding a positive solution – i.e. a justification
of induction – then P’s is not what we might
regard as a solution. Popper’s is a negative solution to “the problem of
induction”, that shows how scientific procedure and
testing may be explained in terms of falsifiability and without a justification
of induction.
Yet it is clear that Ayer wants to suggest that something is
amiss here. What is amiss? Ayer writes that P “has no truck with any talk of
confirmation.
All the same, he is willing to say that a hypothesis is corroborated if it
passes a severe test. We are not, however, supposed to infer that its being
corroborated makes it any more credible.
“But this is very
strange. For what would be the point of our testing our hypothesis at all if
they earned no greater credibility by passing the tests? It is not just a
matter of our abiding by the rules of a game. We seek justification for our
beliefs, and the whole process of testing would be futile if it were not
thought capable of providing it.” So far, a non-justificationist might shrug
that what we have here is simply an appeal to the justificationist stance that
“[w]e
seek justification for our beliefs”, without Ayer showing there is
justification in an inductive sense or why there should be any more
‘justification’
than critical preference based on severe testing.
“Not only that,
but the whole pretence that we do not reason inductively becomes ridiculous
when we consider how much inductive theory is built into our ordinary ways of
speaking. My belief in the existence of the house in which I live, the clothes
I am wearing, the pen with which I am writing….involve the assumption that a
set of properties habitually go together; these assumptions are founded in my
past experience, and however difficult it may to devise a satisfactory account
of confirmation, there is surely no doubt that they are justified.” Ayer does
not say so, but P does not actually deny such beliefs are true or deny they are
preferable to their alternatives – what P denies is that such beliefs are
justified inductively. It is perfectly possible to defend such beliefs in P’s
terms, simply by way of critical discussion in the light of evidence and
arguments that consider such ‘beliefs’ against alternatives. Though Ayer stops
short of saying it, Ayer writes as if to give the impression that P would deny
such beliefs or would claim they are not
rational – whereas the issue is whether such beliefs have any greater
rationality than that provided, non-inductively,
by critical discussion. It is P’s contention that the rationality of science is
provided, non-inductively, by critical discussion and attendant severe tests.
The
rationality of the kinds of everyday ‘beliefs’ that Ayer here holds so dear, may
also be defended in terms of critical discussion – without induction and without
‘justification’ in an inductive sense being deployed. As a result, it is clear
that it is only inductivist make-believe “how much inductive theory is built
into our ordinary ways of speaking”, for our ordinary ways of speaking are
compatible with P’s non-inductivist account of such beliefs.
Ayer’s rejection of P’s negative solution to “the problem of
induction” amounts to no more than an appeal to the supposedly self-evident fact
that we think inductively and are justified in doing so. But we don’t and we’re
not. And if we don't and we're not, P's work perhaps merits more space than
Ayer grants it.
Donal
Stop me if you've heard this one..
Other related posts:
- » [lit-ideas] Ayer on Popper - Donal McEvoy
