[Return to Dmitry Sepety’s Personal
Page]
Critical
Rationalism and the Problem of Reasonable Beliefs
by Dmitry
Sepety,
Which beliefs are
reasonable and which aren’t? How are we to distinguish? These are questions
which are traditionally paid much attention, considered as of highest
importance in philosophy of knowledge and rationality. Among reputed Karl
Popper’s followers, Alan Musgrave makes these questions central; and attempts
to construe Popper’s philosophy of critical rationalism (CR) and William
Bartley’s philosophy of comprehensively critical rationalism (CCR) into a
general theory of reasonable beliefs. In this article I argue that this
interpretation (I will abbreviate it as CRM) is mistaken; that it fails to meet
the problem for which CR was designed – the problem of infinite justificatory
regress (in Bartley’s description – the problem of limits of rationality); that
CR is not a general theory of reasonable beliefs; that a satisfactory general
theory of reasonable beliefs is impossible. Discussion of Alan Musgrave’s views
is based mostly on his article Critical
Rationalism (2007).
1. What was Popper’s meaning of 'ñritical rationalism'
Popper’s description of
CR in a Lecture read before the British Academe on
“I propose to assume …
that all 'sources' are liable to lead
us into error at times. …there are all kinds of sources of our knowledge; but none has authority… I propose to
replace, therefore, the question of the sources of our knowledge by the
entirely different question: "How
can we hope to detect and eliminate error?" … The proper answer … is,
I believe, "By criticizing the
theories or guesses of others and … by criticizing
our own theories or guesses". … This answer sums up a position which I
propose to call "critical rationalism"”(Popper 2008, 32-34)
Popper’s explanation in Intellectual
autobiography, 1974 (later reprinted as a book Unended Quest):
“In The Open Society I stressed that the critical method … can be
generalized into what I described as the critical or rational attitude. I
argued that one of the best senses of "reason" and
"reasonableness" was openness to criticism … I suggested that the
demand that we extend the critical attitude as far as possible might be called
"critical rationalism"…” (Popper 1974, 92)
Popper’s explanation
at Workshop in
“…what I called Critical
Rationalism is an attitude which I described only in roundabout way, namely I
said it is the attitude "I may be wrong, and you may be right, but let us
sit together and discuss matter critically, and in the end we may not agree but
we will both have learnt something". That attitude I called Critical
Rationalism. The attitude of rational discussion, perhaps with the hope that it
leads to agreement but with the clear realisation that it may not lead to
agreement.” (quot. by: Artigas 1999, 30)
Popper’s
explanation in Introduction to
collection The Myth of the Framework,
1994:
“…rationality and
rational criticism … is a way of thinking, and even a way of living: a
readiness to listen to critical arguments, to search for one’s own mistakes,
and to learn from them. It is, fundamentally, an attitude that I have tried to
formulate … in the following two lines:
'I
may be wrong and you may be right,
and by an effort, we may get
nearer to the truth'.
These
two lines in italics here quoted were first printed in 1945 in my Open Society… The view that they summed
up I called 'critical rationalism'…” (Popper 1994, xii)
2. What is CR’s and CCR’s problem
Popper have introduced a
name 'critical rationalism' in Open
Society to designate the view of rationality and rationalism which could repel
irrationalist charge that rationalism is self-defeating. That alleged
self-defeatingness of rationalism stems from a demand which is traditionally
considered as a definitive condition of rationality: “any assumption which cannot be supported either by
argument or by experience is to be discarded”. Later, Bartley has coined a designation for
this demand – 'justificationism'. Popper admits that this demand is both unfeasible,
for it generates infinite justificatory regress (“Since all argument must proceed from assumptions, it
is plainly impossible to demand that all assumptions should be based on
argument”), and self-defeating,
for it fails to meet its own demand (“since it cannot, in its turn, be supported by argument or by experience,
it implies that it should itself be discarded”). (Popper 1977, 230)
Bartley proposed comprehensively
critical rationalism (CCR) as a solution for a closely allied problem – the problem of limits of rationality.
The problem is posed by the contention (supported by many influential philosophers
of XX century) that to avoid justificatory infinite regress and vicious circularity
it is logically necessary to dogmatize some 'basic' positions, make them exempt
from criticism. CCR, Bartley’s solution to the problem of limits of rationality,
is refutation of the conclusion of this inference, explanation that it doesn’t
follow from the premises, and that another possibility is open.
3. How CR and CCR solve the problem of infinite justificatory regress vs irrationalism-dogmatism (the problem
of limits of rationality)
In Open Society Popper has proposed the solution to the problem of
infinite justificatory regress which he did call 'critical rationalism'. It
consists of three propositions.
(1)
to admit that justificationist demand is unfeasible and self-defeating;
(2) to identify rationality not with with
justificationist demand, but with critical attitude ("I may be wrong and you may be right, and by an effort, we may get
nearer to the truth");
(3) to admit that rational attitude is
based on "irrational decision", "irrational faith in
reason".
In The Retreat to Commitment Bartley harshly criticized both
traditional rationalism as 'justificationism' and critical rationalism as
described by Popper in Open Society.
He proposed his theory – comprehensively critical rationalism – in a sharp
opposition to both. It is important to understand what exactly he criticized
about CR. From three above propositions Bartley accepts (1) and (2); his
criticism is aimed at (3). Bartley did see in (3) fideism, the tendency "to
ground or justify rationalist position in personal or social commitment to
standards which are beyond challenge" (Bartley 1984, 97), remnants of justificationism.
Probably he wasn’t quite right, but there is an important robust core in his
criticism. The point of Bartley’s criticism of CR (or, rather, its description
in Open Society) is as follows: (3)
is description of CR in terms of justificationist conception of rationality; if
we reject justificationist demand (1) and identify rationality with critical
attitude (2), then holding CR needn’t be irrational in CR’s own terms. CR is held
rationally if it applies critical attitude to itself. And there are no logical
reasons why critical rationalist can’t apply critical attitude to itself and to
CR, i.e. to be rational toward CR in CR’s own terms. (To contrast with
justificationist conception of rationality which is irrational in its own
terms, for simple logical reasons.) And there are no logical reasons why critical
rationalist can’t apply critical attitude to all positions, i.e. be comprehensively critical, i.e. comprehensively rational in CR’s own
terms. This is, in a nutshell, Bartley’s conception of comprehensively critical
rationalism.
CCR is nothing but
refutation of the following "fideist" contention:
(F) As
any attempt to demonstrate truth of any position involves us in infinite
regress or vicious circle, we have no choice but to commit themselves to some
“basic” positions and principles, and then use them in justifications and
criticisms of other positions. Such commitment means that “basic” positions and
principles are accepted dogmatically, as such that 1) can’t be justified and
need no justification and 2) can’t be criticized (there can be nothing to
criticize them with, for they are basic); are exempt from criticism. This
contention is what Bartley calls "the retreat to commitment".
Now CCR shows that this contention
is false. CCR agrees with F that we can’t 'justify' (in the sense of demonstrating
truth, or even high probability) anything as demanded by justificationism, and,
so, must accept some (many) positions without justification. (Hence, all other
positions aren’t justified as justificationism demands; they may be 'justified'
only relative to some unjustified
positions.) But, contrary to F, CCR contends that this does not mean that we must dogmatize something, make some positions
exempt from criticism. We can hold
all our beliefs, however basic, open to (not exempt from) criticism. I.e. we
can consider all our beliefs as
tentative, admitting that possibly we are mistaken and that critical
discussion can reveal this mistake; and we can be willing to enter into such
discussions to find out our mistakes and eliminate them. In a nutshell, CCR is
a conjunction of three positions: 1) comprehensive fallibilism; 2) seeing truth
as the aim and critical discussion as the best means for it; 3) considering
deductive logics, especially the demand of non-contradiction, as "organon
of critical discussion" (Popper’s expression).
Compare this with
Bartley’s explanations:
“I was confronting the contention … that there
is an essential logical limitation to rationality: the rational defence and
examination of ideas must, for logical reasons, be terminated by an
arbitrary and irrational appeal to what can be called dogmas or absolute
presuppositions. …I argued: (1) that nothing of any interest can be
justified in the way required…; (2) criticism is nonetheless possible provided
one unfuses justification and criticism…;
(3) there are no limits to rationality in the sense that one must postulate dogmas or presuppositions
that must be held exempt from review…” (Bartley 1984, 221)
“What I did … was to show that no authorities or justifiers in this sense
were needed in criticism. I separated the notions of justification and
criticism (for the first time explicitly) and showed that criticism can be
carried out successfully and satisfactorily without … any resort to dogmas or
authorities. That is, when I declare that all statements are criticizable, I
mean that it is not necessary, in criticism, … to declare a dogma that cannot
be criticized…; I mean that it is not necessary to mark off a special class of
statements, the justifiers, which do
the justifying and criticizing but are not open to criticism…” (Bartley 1984,
223)
Though Popper denied correctness
of Bartley’s criticism of CR, he have adopted Bartley’s correction that rationality
as openness to critical discussion can
be comprehensive and self-subsumptive. So, in Addendum to 4-th edition of The Open Society Popper twice mentions
with approval CCR’s central contention – "the principle that everything is open to criticism (from which this principle itself is not exempt)" (Popper 1977, 408-409). He
also adopted Bartley’s term 'justificationism' to designate the view CR
opposed. Popper’s later descriptions of CR are in conformity with CCR. So, I
think that CR and CCR may be identified.
4. How Bartley Have Let in the "Troyan
Horse" of Justificationism
Bartley’s solution to
the problem of limits of rationality is often misunderstood. I think that one
reason of this wide misunderstanding is Bartley’s own misdescription. Probably
the most often quoted – equally by critics and supporters – Bartley’s statement
about CCR is:
“The position may be held rationally without
needing justification at all – provided
that it can be and is held open to criticism and survives severe examination”. (Bartley 1984, 119)
This statement is
misleading in two directions:
(1) It may be easily misunderstood
it the sense that some positions can be held open to criticism and some
positions can not. Hence, misinterpretation of CCR as a theory about
criticizability as a logical property of some positions (statements). For
example, most discussed criticisms of CCR – those of Post and Watkins – are
based on this misinterpretation. Musgrave also writes that "Sometimes
Bartley seemed to be saying that a belief is reasonable if what is believed is criticizable." And he adds:
"But this is a crazy thesis" (Musgrave 2007, 194)
Really, Bartley’s
statement does not mean anything of the kind. Phrase "provided that it can be open to criticism" is meant for quite
a different purpose – to separate "needing no justification" from
"can not be held open to criticism" – to contrast with the
commitment-contention F that as we can’t avoid accepting some positions without
justification (which, on Bartley’s view, is true), so these positions are
exempt from criticism (which, on Bartley’s view, is false).
I doubt if Bartley wrote
somewhere anything like that “a belief is reasonable if what believed is criticizable”. I can’t find such, or
similar formulation in “Retreat to Commitment”. Really, Bartley’s position is
that all beliefs are criticizable,
where 'criticizable' means 'can be hold open to criticism, needn’t be exempt
from criticism'. To qoute Bartley:
“… when I declare that all statements are
criticizable, I mean that it is not necessary, in criticism, … to
declare a dogma that cannot be criticized…” (Bartley 1984, 223)
“It is … one of the merits of
pancritical rationalism … that it presents a theory about people, not
statements. It is, quite explicitly, an account of the essense of being a rationalist. It is an account of how a
rationalist or critical person might behave. It is not an account … of rational
statements or of rational belief. … I doubt the merit of discussing the problem
of rationality in terms of rationality of statements… For statements are
intrinsically neither rational nor irrational. …rationality is not a property
of statement but is a matter of the way in which a statement is held, and also
… of the way in which the statement has been examined.” (Bartley 1984, 233-234)
(2) The real – and highly
misleading – misdescription of CCR is the concluding phrase "and survives severe examination”.
As far as I know, nobody
of critics or supporters of Bartley has noticed that the statement "For a position
to be held rationally it must survive severe examination" directly
contradicts Bartley’s further explanations and undoes his solution to the
problem of limits of rationality.
First: what is to be
counted as "severe examination"? What can be criteria of severity: if
examinations were severe enough to make holding the position rational?
Second: no examination,
no criticism, no critical discussion is possible without assuming some
positions as true. Really, to criticize we need some assumptions used as
'criticizers'. But there seems to be a problem: we never know if those
'criticizers' are true; their truth (or high probability) cannot be
demonstrated. If so, is our choice of 'criticizers' arbitrary? This would ruin
rational process; we could "successfully" "refute"
anything.
It wouldn’t do to say
that we accept as criticizers only those positions which "survive severe
examination" – for to make that examination we would need some
'criticizers' to start with; and so we get into infinite regress of examining
criticizers of criticizers of criticizers …
Now if, for position X
to be held rationally it is demanded that there was severe examination, then we
need some set of "criticisers" {C1}. But it is clear that
those criticisers can’t be anything whatever. They must themselves be
rationally held positions. But for them to be rationally held they must survive
severe examination, and so we need another set of criticisers {C2}
which must themselves be rationally held positions, i.e. survive severe
examination. So we need next set of criticisers {C3} which must
themselves be rationally held positions, i.e. survive severe examination. And
so ad infinitum.
Bartley’s explanations
provide a solution to this problem:
“…pancritical rationalist, like other people, holds
countless unexamined presuppositions and assumptions, many of which may
be false. His rationality consists in his willingness to submit these to
critical considerations when he discovers them or when they are pointed out to
him... When one belief is subjected to criticism, many others, of course, have
to be taken for granted – including those with which the criticism is being
carried out. The latter are used as the basis of criticism not because they are
themselves justified or beyond criticism, but because they are unproblematical at present. These are, in that sense alone and during that time
alone, beyond criticism. We stop criticizing – temporarily – not when we
reach uncriticizable authorities, but when we reach positions against which we
can find no criticisms.[1] If criticisms of these are raised
later, the critical process then continues.” (Bartley 1984, 121-122)
But now, if
"pancritical rationalist, like other people, holds countless unexamined
presuppositions and assumptions", "including those with which the
criticism is being carried out", does he hold them rationally? I am sure that
Bartley’s answer is "Yes" (surely, it is not Bartley’s view that
pancritical rationalist non-rationally holds countless unexamined
presuppositions and assumptions, and that holding a position becomes rational
after it is criticized with other positions which are hold non-rationally). But
this directly contradicts with the statement that for a position to be held
rationally it must survive severe examination.
The demand that a position
– to be rationally held – must
survive severe examination happens to be for CCR "Troyan horse" of justificationism.
Surely, I don’t mean that it may be rational to hold a position which failed to survive severe examination. (It
is not only not rational; it is impossible, as being a married bachelor! Really,
if a position failed to survive
examination, whether severe or not, this means that it is not held already. If
it is still held, this means that it has
survived!) I mean that it may be rational to hold a position if there was no
severe examination, and if we have no idea of how it can be examined or criticized.
I hold a position X. X
seems true to me. I believe that X is true. I have no idea how X can be
seriously criticized or examined. But I admit that possibly I am mistaken and
that possibly someone can produce criticism (or examination) of X which will
succeed to persuade me out of my belief that X is true, which will make me stop
believing it. I am willing to listen to criticisms of X, if produced. So, I
hold X open to criticism, i.e. rationally.
So, I think that we can
make sense of CCR if, and only if, we consider the phrase "and survives
severe examination" as mistaken slip (probably, remnants of earlier failed
attempt of stronger version of CCR). Or we can replace this phrase with:
"and, if examined (criticized),
survives examination (criticisms)". But then this phrase would be
redundant: if position is still held, this means that it has survived.
So, I think that
adequate statement of CCR’s condition of rationality is:
“The position is held
rationally if it is held open to criticism”.
5. Isn’t CCR, in my interpretation, too weak?
On the first glance,
such a conception of rationality can seem too weak. That is the view of
Musgrave:
"Much of the time Bartley seems to be
saying that a belief is reasonable if the believer takes a certain attitude or stance towards it, that of "not being irrationally committed
to it" or "holding it open to criticism" or "being willing
to listen to criticism and take it seriously". Now I know what it is to
hold a belief. But I am not sure that I know what it is to "hold a belief
open to criticism". It cannot be anything as strong as to accept any criticism once it is offered
and to reject the belief, since criticisms can themselves be criticized and
rejected. Anything weaker threatens to be too week." (Musgrave 2007, 195)
My reply is as follows.
First. Too week for
what? It is strong enough for solving that problem for which it was designed –
the problem of limits of rationality.
Second. Popper’s conception
of rationality in his descriptions of CR (see section 1) is just as weak: Popper,
just as Bartley, identifies rationality with an attitude of openness to
critical discussion.[2] As Popper and Bartley explain, there
are principally two main trends of meanings in which words 'reason',
'reasonablenes', 'rationality' are used: the stronger sense presupposes "that any assumption which cannot be supported either
by argument or by experience is to be discarded", identifies 'rationality' with demand of
absolute (conclusive or probabilistic) justification; the weaker sense identifies
'rationality' and 'reason' with critical discussion and openness to criticism.
The idea of 'rationality' in the stronger sense is indefensible,
self-defeating. If we admit this and still remain rationalists, then what we
are left with is the idea of rationality in the weaker sense.
Third. If Musgrave
proposes something stronger? And if what he proposes solves the problem for
which CR and CCR were designed?
6. What is CRM’s Initial Problem
and How Its Solution is Extended into a General Theory of Reasonable
Beliefs
In all his articles and
books Musgrave initially proposes his theory of reasonable beliefs as the solution
to Hume’s problem of induction, and then attempts to extend it outside the
proper domain of that problem. Musgrave’s theory of reasonable beliefs is
derivative from Popper’s theory – but not that which Popper called 'critical
rationalism'. It is derivative from more specific (applicable only to empirical domain) Popper’s theory – falsificationism.
This theory was designed as a solution to Hume’s problem of induction. Critical
rationalism, in Popper’s meaning, was designed to solve a different problem (and, accordingly, it is a different solution). This is a point of importance.
It is expedient to
emphasize here the difference between these two Popper’s theories:
– critical rationalism (CR) as non-justificationist conception of rationality designed as a
solution to the problem of infinite justificatory regress
and
– falsificationism as a theory of methodology of empirical sciences (application of hypothetico-deductivism) designed as a
solution to the problem of induction.
Falsificationism is about inferences "in inductive direction", from singular
to universal statements; (C)CR is
about inferences in whatever direction.
In terms of 'justification'
(in the meaning: logically valid demonstration of truth or high probability):
(C)CR is
about impossibility of absolute
justification, i.e. about justification
without unjustified assumptions;
falsificationism is about invalidity of inductive
justification, which is relative
justification – justification of universal statements (scientific theories)
relative to (on assumption of) singular statements (observational "basic
statements"), which themselves are "unjustified".
Falsificationism argues against induction (as invalid) and verification (as either
impossible, if understood in a strong sense, or useless, if understood in a
week sense), and advocates the view that scientific knowledge is developing in
the process of inventing hypotheses and their empirical testing – attempts of
falsification.
CR argues, with irrationalists, that justificationist
demand ("that any
assumption which cannot be supported either by argument or by experience is to
be discarded"),
which is usually identified with rationalism, is unfeasible and self-defeating.
I.e., absolute justification is impossible; no statement (except tautologies)
can be justified as demanded. CR also argues, against irrationalism, that it is not ruinous for rationality and
rationalism; that rationalism as faith in
reason and force of argumentation is defensible. CR is nothing but
proposition to renounce justificationist
demand and to identify 'rationality'
with critical discussion, critical attitude, openness to criticism.
Popper’s solution to the
problem of induction – falsificationism
– can be "translated" into terms of reasonable beliefs. Popper
himself explained that according with his theory it is reasonable to prefer those scientific theories (empirically
falsifiable universal hypotheses) which have best withstood attempts of empirical
falsification (refutation with observational singular statements) and other
criticisms. He also admitted that it is equivalent to saying that it is
reasonable to believe that such theories
are preferable. It is as well correct, that Popper’s theory gives an answer
to the question: when it is reasonable to
tentatively adopt a theory as true?
Musgrave transforms this
into a theory of when it is reasonable to believe a (non-perceptual)
hypothesis, and does not limit this theory to the case of empirically falsifiable universal statements, but proposes it as a general theory of reasonable beliefs. And
he identifies it with CR.
Hume’s problem of
induction and the problem of infinite justificatory regress are two different problems which need – and find
in philosophy of Popper and Bartley – two different
solutions. Musgrave overlooks most of the difference. He assumes that the
theory of rational beliefs about empirical
theories (universal empirically falsifiable statements) which he extracts
from Popper’s falsificationism is straightforwardly extendable on all other
matters so that we get a general theory of reasonable beliefs identical with
CR.
I argue that the
assumption about extendability, the extension proposed and its supposed
identity with CR are mistaken.
7. How CRM Fails to Solve the
Problem of Infinite Justificatory Regress
That is how Musgrave
formulates what he describes as "critical rationalism’s central contention"
(CRM1) "…if a
hypothesis has withstood our best efforts to show that it is false, then this
is a good reason to believe it but not a
good reason for the hypothesis itself." (Musgrave 2007, 181)
7.1. Do "Reasons for Believing a
Statement" Help?
This
formulation contains some technicality – distinction between 1) a good reason
for a hypothesis and 2) a good reason for believing a hypothesis (believing
that a hypothesis is true). Earlier Musgrave has explained that by 'a good
reason for a hypothesis' he means a logically valid demonstration that the
hypothesis is true or highly probable. So, (CRM1) can be
reformulated as follows:
(CRM1')
"…if a hypothesis has withstood
our best efforts to show that it is false, then this is a good reason to
believe it, but it is not logically valid
demonstration that it is true or highly probable."
Musgrave seems to
consider distinction between reasons for statements (for belief-contents) and
reasons for believing (adopting etc.) statements (for belief-acts) as decisive
for non-justificationism. He often describes-defines justificationism as "view
that a reason for a belief-act must be a reason for its content" (Musgrave
2009, 5), "the assumption … that a reason for believing something must
also be a reason for what is believed", "that a good reason for believing something must also be a good reason in the logical sense for
what is believed" (Musgrave 2009, 6)
As far as
I see it, this description-definition uses the term 'justificationism' in the
sense very different from that in which Bartley and Popper used it. Distinction
between reasons for statements (for belief-contents) and reasons for believing
(adopting etc.) statements (for belief-acts) have some merits, especially in
context of the problem of induction, but it helps nothing as far as the problem
of infinite justificatory regress (the problem of limits of rationality) is
concerned.
The
distinction has a merit for hypothetico-deductivism. It provides means to
answer in hypothetico-deductive manner, without reference to induction, the
question: why it is reasonable to adopt an empirical theory (empirically
falsifiable universal hypothesis). The argument is a deductive inference from
two premises:
(1) a principle defining a condition
of reasonable adoption ("It is reasonable to adopt as true empirically
falsifiable universal hypothesis if it has withstood attempts of
falsification")
and
(2) statement about the state of
testing of an empirically falsifiable universal hypothesis ("X has
withstood attempts of falsification")
to the conclusion – statement that
it is reasonable to adopt as true the empirically falsifiable universal
hypothesis (X).
This
argument is perfectly deductively valid, and its conclusion is not truth of X, but reasonableness of adopting X as
true.
The
distinction is useful also for quite a many other "hypothetico-deductive
reconstructions" of important arguments (for example, the argument known
as inference to the best explanation). But, to repeat, whatever are merits of
the distinction, being helpful for solution of the problem of infinite justificatory
regress (the problem of limits of rationality) is not one of them.
As far as
the problem is concerned, it doesn’t matter whether reasons you provide are
reasons for a statement or reasons for believing (adopting) a statement.
Whatever they are, we can always ask for reasons for those reasons (like above
premises (1) and (2)) – or for reasons to believe those reasons. And we can ask
for such reasons ad infinitum.
The
general scheme of infinite regress with reasons for believing is:
S
(R0)
It is reasonable to believe S.
Justificatory demand: What are
reasons to believe S?
Justification of belief-act:
(R1)
It is reasonable to believe a statement whenever it fits condition C1
(J1)
S fits condition C1
Hence, it
is reasonable to believe S
Justificatory demand: What are
reasons to believe R1 and J1?
Justification of belief-acts:
(R21)
It is reasonable to believe a statement whenever it fits condition C21
(J21)
R1 fits condition C21
Hence, it
is reasonable to believe R1
(R22)
It is reasonable to believe a statement whenever it fits condition C22
(J22)
J1 fits condition C22
Hence, it
is reasonable to believe J1
Justificatory demand: What are
reasons to believe R21, J21, R22, J22?
…………………………………………….
ad infinitum
So, we are
involved into infinite justificatory regress which ramifies on each
justificatory step. Some branches of this regress may be cut off with the help
of some subsidiary principles (and Musgrave proposes some such principles), but
then those principles themselves must be reasonable beliefs, and if we try to
provide reasons for believing them we are inevitably caught in infinite
justificatory regress.
To break
these infinite regresses we must stop somewhere without providing further
reasons. But then, believing those stopping points is unjustified, and all the
chain of justifications collapses. No justification can be achieved this way.
7.2. Problematic Questions
Now we
will be interested in the first part (before italicized text) of what Musgrave
calls "critical rationalism’s central contention"
(CRM1'') "…if a hypothesis has withstood our best
efforts to show that it is false, then this is a good reason to believe
it..." (Musgrave 2007, 181)
Musgrave
gives also more precise formulation of what he considers to be "the heart
of critical rationalism" (Musgrave 2007, 182):
(CRM2) "…it is reasonable to believe
that P (at time t) if and only if P is
that hypothesis which has (at time t)
best withstood serious criticism." (Musgrave 2007, 183)
When discussing CCR,
Musgrave quotes Bartley’s statement which I have criticized above as "Troyan
horse" of justificationism – that the position is held rationally if it is
held open to criticism and survives
severe examination – as identical with his interpretation of CR. So, my
criticism of this "Troyan horse" in the section 4 holds for
Musgrave’s interpretation of CR.
Besides, Musgrave adds:
"I assume that to hold a position rationally is to have a reason for
adopting it as true or believing it". But what is a reason for this
assumption? If there is any, what is a reason for that reason etc. ad infinitum? Isn’t this assumption
irrational on its own criterion?
Musgrave approves
Bartley’s argument for "comprehensively critical rationalism, a
rationalism that could subsume itself and be rational by its own lights"
(Musgrave 2007, 189). And he thinks his interpretation of CR satisfies this
demand. But it doesn’t: the principle "to hold a position rationally is to
have a reason for adopting it as true or believing it" can’t be comprehensive and self-subsumptive, for it generates infinite justificatory regress.
In part,
Musgrave seems to admit it when he gets down to observational statements
(which, from CR viewpoint, also are to be qualified as hypotheses). He concedes
that it would hardly be reasonable to demand that all such statements "has
withstood our best efforts to show that it is false", "has (at time t) best withstood serious
criticism".
So, he modifies
"critical rationalism’s central contention" (Musgrave designates it
as CR, I will designate it as CRM) and supplements it with
subsidiary principles E and T. So we have:
CRM
It is reasonable to non-perceptually believe that P (at time t) if and only if P is
that hypothesis which has (at time t)
best withstood serious criticism.
E It is reasonable to perceptually believe that P
(at time t) if and only if P has not failed to withstand criticism
(at time t).
T A’s
testimentory belief that P is reasonable
(at time t) if and only if P has not failed to withstand criticism
from A (at time t)
But is this enough? Let
us ask further questions:
1) If the belief
supported by E conflicts with the belief supported by T, how are we to decide
which one is reasonable? To answer the question we need yet another subsidiary
principle – let us call it ET.
2) About each of
statements CRM, E, T, ET we can ask further questions: is it
reasonable to believe (adopt) this statement? What are good reasons to believe (adopt)
statements of this kind? With respect to this kind of statements, how are we to
distinguish reasonable beliefs from unreasonable? Do we need some principle of reasonableness
of second level to judge reasonableness of proposed principles of reasonableness
of first level? But how do we know if that second-level principle of reasonableness
is reasonable? Do we need third-level principle of reasonableness, forth-level
principle of reasonableness etc. ad
infinitum? As consequence: we never know which of all possible propositions
for principles of reasonableness are reasonable and which aren’t.
3) What about our
judgements that a belief fits – or doesn’t fit – certain principle of reasonableness?
(Let us call such judgements 'fitness-judgements') When it is reasonable/unreasonable
to believe that a fitness-judgement is true? Do we need principle of reasonableness
of fitness-judgements? Suppose some multitude of such principles are proposed?
How are we to distinguish which of these propositions are reasonable? Do we
need second-level principles, third-level principles, forth-level principles ad infinitum?
4) How about other kinds
of beliefs: such as moral beliefs, metaphysically-psychological and metaphysically-ontological
beliefs? (Is killing innocent people good or bad or indifferent? Is our mental
self the same with our brain or it is a different entity? was our world created
by omnipotent or superipotent Creator?)
Musgrave’s attempted
answer to these questions seem to be: the principle to decide on reasonableness
of statement about reasonableness is CRM itself; so we need no
further principles of reasonableness.
Now there are three further
questions to discuss:
(1) can we really apply CRM
to all non-perceptual beliefs,
especially those which are empirically unfalsifiable?
(2) can we effectively apply CRM to non-perceptual empirically
unfalsifiable beliefs? Is CRM really able to serve as a criterion of reasonableness of beliefs?
(3) can CRM be reasonably
accepted as a good reason for believing CRM itself?
I argue that answers to
all these questions are negative.
7.3. Why the Demand to Withstand Serious Criticism Can’t Be Applied to
All Nonperceptual Beliefs
To see that the answer
for questions (1) and (2) is negative, we are to note that CRM
contains nothing substantial about criticisms (examinations). CRM
itself does no critical job, it just demands that the job was done. But this
brings us back to the question of what count as severe criticism and to
infinite regress of criticizers of criticizers of criticizers … For any
critical discussion to be possible we must accept some empirically unfalsifiable
positions (such as beliefs about conditions of reasonableness,
fitness-judgements etc.) without anything which deserves to be qualified as
severe examinations (criticisms).
Let us compare this with
the case of universal empirical (empirically falsifiable) hypotheses, how it is
discussed by Musgrave in the article "How Popper [Might Have] Solved the
Problem of Induction". In section 2 Musgrave criticizes – on my view,
quite correctly – logomania ("the view that only reason or reasoning
provides a good reason for believing anything") as leading to infinite
regress and proposes, to stop this regress, to admit that some non-inferential
beliefs are reasonable and that perceptual beliefs (expressed in the form of
observational statements) are of this kind. (Musgrave 2004, 22) So far so good.
To be more precise, in the
case of universal empirical (empirically falsifiable) hypotheses critical job
is done by
– singular observational statements
which, as Musgrave admits in principles E and T, needn’t have "withstood
our best efforts to show that it is false" (have "best withstood
serious criticism")
and
– epistemological principles (which
may be formulated as statements about conditions of reasonableness) which are
"taken for granted".
But what is to do
critical job with respect to those epistemological principles (or hypotheses
about conditions of reasonableness) themselves? Observational statements can’t
do the job. The fact that event e was
observed at time t at place p (whatever e, t and p are) is of no
argumentative relevance to whatever epistemological hypothesis or principle. It
can’t serve as "a reason for adopting it as true or believing it", –
neither 'positively' (as verification) nor 'negatively' (as serious attempt of
falsification which has failed).
A hypothesis about
reasonableness can be criticized by some other hypotheses about reasonableness.
But then these other hypotheses must be accepted without complying with demand
that they "withstood our best efforts to show that it is false". If all
hypotheses about reasonableness were demanded to "withstood our best
efforts to show that it is false", this would entail infinite regress!
So, to avoid infinite
regress in discussing such theories and principles, we need some other
reasonable non-inferential beliefs besides observational statements.[3]
The same goes for moral
beliefs and also some other kinds of "metaphysical" beliefs. I think
one of the most important points of Popper’s philosophy is that we can’t do
away, as positivists hoped, with "metaphysics", can’t avoid holding
some metaphysical beliefs; that we can hold them rationally; that these
beliefs, as well as empirical ones can be subjected to critical discussion;
they are "rationally arguable".
So, if we mean by CR the
general conception of rationality (as Popper did use the title) rather than its
special application for empirical sciences, then E, T and ET are not exceptions
for CR, but particular cases of general rule.
My view is that initial
non-inferential beliefs (which serve as tentative reasons for other beliefs) can
be anything we are psychologically disposed to believe. They can be suggested
by perceptions, other peoples, common sense etc. Popper especially emphasized
that we are already born with sort of unconscious beliefs-expectations which
are later "criticized" by our experiences and modified. After we have
mastered language, this process continues on higher, conscious level.
Believing many such
unexamined positions by itself is neither rational nor irrational; it is just
necessary starting platform. Rationality and irrationality depends mainly on
how we treat those positions we already believe or tend to believe – as
unassailable dogmas, arbitrary commitments or as tentative, fallible, open for
critical discussion and revisable hypotheses.
Really, critical process
departs from unexamined beliefs, and we examine these beliefs by confronting
them with each other, and finding out contradictions, and trying to eliminate
them in such a way that would save all the most plausible and most important. (To
anticipate the charge of "psychologism". Estimation of plausibility
and importance is a matter of "psychology" – informal understanding,
judgement, intuition, "gut-feeling". Beliefs never are – and can’t be
– a matter of logics alone. They are always a matter of a psychology, though
psychology can check and correct itself by means of logics.
To illustrate with a
simple example.
I see a cat and this
makes me believe there is a cat. To be rational, I am not bound to give reasons
why I believe there is a cat. If someone denies it, then burden of argument is
his. If he will succeed to pesuade me, then I stop believing there is a cat. If
not, then not. My rationality is my willingness to listen to arguments of my
opponent.
Suppose, you say that my
reason to believe that there is a cat is my seeing it. Suppose, I agree. This
just shifts the problem of providing a reason to a meta-level. To be rational I
am not bound to give reasons why I think that seing X is a good enough reason
to believe there is X. If someone denies it, then burden of argument is his. If
he will succeed to pesuade me, then I stop believing that seeing a cat is a
good enough reason to believe there is a cat. If not, then not. My rationality
is my willingness to listen to arguments of my opponent.
Besides,
"justification" of my belief that there is a cat by reference to the
principle of reasonableness of perceptual beliefs presupposes yet another
belief which goes without providing reasons – belief about perception. Really,
this "justification" goes as follows:
(1) it is reasonable to believe that
seing X is a good enough reason to believe there is X;
(2) I see a cat over there.
Hence, it is reasonable to believe that there
is a cat.
But there are two
premises in this arguments which go without reasoning. Besides the principle
(1), there is a statement about perception (2). Am I to provide reasons why I
believe that I see a cat, and then reason for believing that reason etc. ad infinitum?
The fact is, that in
such cases our beliefs aren’t based on reasons; perceptions don’t serve as a reason for acceptance (believing); they
just cause – or suggest – beliefs. For some
of such beliefs reasons can be provided, if asked for, by a hindsight. But it
is not the case that those reasons, if provided, are themselves based on
further reasons which are based on yet further reasons ad infinitum. Something must be accepted (believed) without
providing reasons. But if some belief is accepted now (whether with or without
providing reasons), it can possibly be criticized in future; and these
criticisms can possibly persuade us to renounce it.
To
recapitulate. We can escape infinite regress and vicious circularity only if we
admit that position can be believed
reasonably (rationally) even if no good reason for believing it can be provided
and even if we didn’t yet subjected it to critical examination. But in this
case we need some other way to demarcate rational and irrational. As far as I
see it, openness to criticism is the only feasible candidate.
As far as I understand Popper
and Bartley, their view is not that
for any hypothesis to be rationally believed it is necessary that it "has
withstood our best efforts to show that it is false". True, it is Popper’s
demand for acceptance of scientific
theories (i.e. empirically falsifiable universal statements), but it
belongs to Popper’s more specific theory, falsificationism, not to CR as
general theory of rationality. CR (≡rejection of justificationism) as a general
conception of rationality which applies
to all beliefs – whether scientific or not, whether universal or singular –
is not so exacting. It demands only that we admit our fallibility and were
willing to consider criticisms whenever we meet them, and to revise our beliefs
in their light. So, on my view, "critical rationalism’s central
contention" is that rationality is
to be understood not as justification but as openness to critical discussion
aimed at truth.
7.4. Is Belief in CR as Reasonable
as Belief in All Granny Told Me?
On my view, CRM makes
it reasonable to believe that CRM (or to believe that CRM is
reasonable) no more than my statement that all I have ever said is true makes
it reasonable to believe that all I have ever said is true.
Musgrave admits that on
his account justification of believing CRM is circular; believing CRM
is justified only on CRM own terms: CRM states that to be
reasonable non-perceptual belief has to "best withstand serious
criticism", and CRM itself "best withstands serious
criticism" (comparing with its alternatives – different versions of
justificationism, dogmatism and irrationalism). The reason for accepting CRM
is CRM itself!
Musgrave argues that
this is "not devastating" for CRM, for there can be no
better alternative for any general
theory of reasonable beliefs. Any such a theory TRB, when asked why it is
reasonable to believe it, can
either (1) give no
reason and, so, admit that belief in TRB is unreasonable;[4]
or (2) give some reason extraneous to TRB, which would mean that
TRB is not an adequate general theory of reasonable beliefs;
or (3) to say that a
reason for believing TRB is that it complies with TRB’s criteria of
reasonableness.
Musgrave thinks that 3)
is the best of three, for 1) and 2) are self-defeating.
But now the question is:
if CRM is the only general theory of reasonable beliefs which can provide
(3) as a reason for believing itself? Musgrave admits that it isn’t; such
self-justifucation is easily achievable:
“Self-subsumption is too
easy to obtain. "It is reasonable to believe anything said in a paper by
Alan Musgrave" subsumes itself, since it occurs in this paper, but it is
crazy epistemic principle. So is "Granny told me I ought to believe
everything she tells me." And "The Pope declared ex cathedra that everything declared ex cathedra by the Pope is a matter of faith" is no triumph
either.” (Musgrave 2007, 190)
But, if so, why those
"epistemic principles" are "crazy", while CRM (presumably)
is not? Is it more reasonable to believe CRM rather than any other logically
possible self-subsumptive epistemic principle?[5]
If self-subsumption is
acknowledged as permissible reason for believing a general theory of reasonable
belief, then all self-subsuming general theories of reasonable beliefs are
equally reasonable (or equally unreasonable). If it is not, then to demonstrate
a reasonableness of CRM we need some other reason, extraneous to CRM.
But then (2) is the case, and CRM is not an adequate general theory
of reasonable beliefs! And – here Musgrave is right – this holds for any
general theory of reasonable beliefs!
Musgrave proposes critical
rationalists to "take comfort from this – any general theory of reasonable belief will be subject to the same
objection", so "this cannot be used to discriminate against it {CR}
in favour of some rival theory of rationality". "An objection that
hits all competitors with equal force is no objection to any particular
one." (Musgrave 2007, 190)
I see nothing to take
comfort of. It is the same as to say: "If to punch everybody in the face
with equal force it will hurt nobody". If it is shown that all general
theories of reasonable beliefs are unsatisfactory for the same reason, they all
are unsatisfactory. Full stop.
Does objection of
circular justification hits (C)CR too? I think no, for (C)CR is not a general
theory of reasonable beliefs.
(C)CR doesn’t make
belief in (C)CR reasonable. (C)CR doesn’t claim for such a privilege of
self-justification. It admits that it itself is not justified – whether is the
sense of demonstration of truth, or in the sense of demonstration of high probability,
or the sense of demonstration of reasonableness. And it denies that such
justification is necessary for rationality. (C)CR may well be true without justification.
Possibly, it may also well be reasonable without justification?[6]
8. How (C)CR Manages to Be a General Account of Rationality
without Being a
General Theory of Reasonable Beliefs
As far as empirical matters are concerned, Musgrave’s theory of reasonable beliefs is
congenial with Popper’s solution of the problem of induction – falsificationism.
True, Popper’s own explanations about what he thinks reasonable to believe
about scientific theories not quite fit with Musgrave’s formulation that
"…if a hypothesis has withstood our best efforts to show that it is false,
then this is a good reason to believe it..." (Musgrave 2007, 181) But, I
think Popper would agree with it, if we interpret "to believe it" not
as "to believe that it is true sensu
stricto", but as "to believe that it is either true or the best
approximation to truth we have".
But Popper’s philosophy
contains no general theory of
reasonable beliefs (especially, about non-empirical matters), no general criterion of reasonableness for beliefs,
just as there is no general criterion of truth (Popper argued that no general
criterion of truth is possible). CR and CCR are not such theories. They give a
general account of rationality not as
reasonable beliefs, but as critical attitude, attitude of openness to critical
discussion.
CR and CCR use the word 'rational'
in the meaning very different from that of 'reasonable beliefs':
(1) 'rationality' as
property of attitude toward our beliefs and as property of scientific method,
on CR and CCR accounts, means willingness to subject our beliefs and scientific
theories to critical discussion, examinations, testings.
(2) 'reasonableness' of
belief means quite another thing; I think that when discussing what is
reasonable to believe about scientific theories Popper did hardly mean by
'reasonable' anything more than plausible
in the light of present state of critical discussion.
Generally, I think that
estimation of beliefs as reasonable/unreasonable is no more than personal
judgement of what is plausible/implausible in the light of common sense and arguments and evidences which a person knows about.
And it is impossible to make an exhaustive codification of common sense. If it
is so, no satisfactory general theory of reasonable beliefs is possible;
neither it is needed.
There is no all-purpose
system of principles to account for conditions making some beliefs reasonable
while others unreasonable (though, surely, we may formulate some particular
common sense principles for what we think reasonable to believe).
May be, it would be useful
to distinguish personal and public levels of the question of reasonableness of
beliefs.
On the personal level, I
doubt that distinction of reasonable and unreasonable beliefs does matter: for a
person always believes in what he/she
is made to believe by all evidences and arguments he/she did consider up to the
present moment (which is for many beliefs, inevitably, no arguments and
evidences). What does matter for rationality is his/her willingness to pay
attention to further evidences and arguments, his/her openness to critical
discussion. There is only one general "principle": do deploy
alternative positions and arguments pro
and contra, think them through,
propose your own arguments for discussion, and see where it gets you, what
beliefs you have acquired and lost.
On the public level,
question about reasonableness of beliefs may be a question of appeal to common
sense and reasons (arguments). Really, we usually talk about reasonableness or
unreasonableness of beliefs when we want to persuade others that belief X is
preferable over its alternatives. In this case "reasonableness" of
beliefs is their final appeal to common sense. As Popper formulates it,
"we should not depart from common sense … without some fairly good
reason". (Popper 2005, 47) Let us designate this principle as CS.
It is to be remarked
that in accordance with CS "fairly good reason" is made "fairly
good" either by appeal to another "fairly good reason" or by
appeal to common sense. And, as there can’t be infinite regress of "fairly
good reasons", all "fairly good reasons" are finally grounded in
common sense.
It is also to be
remarked that though CS is public, its
application is personal: there is no one but a person himself/herself to
judge what is (or is not) common sense and what is (or is not) a fairly good
reason.
9. The Myth of Belief-act
I hope that the thesis
that we needn’t a general theory of reasonable beliefs will become more
plausible if we take into account one point which Musgrave explains in the
article Can I Decide What to Believe?
(Musgrave 2009b). Musgrave’s answer to the titling question is: we never do –
and can’t – decide what to believe;
our beliefs aren’t matter of our decisions. If I don’t believe that X is the
case, I can’t decide that from now (or from tomorrow) on I will believe that X
is the case, and fulfil this decision.[7] Our beliefs are a matter of
evidences and arguments which persuade us into believing some things and disbelieving
some other things.
Though Musgrave makes
this point in one of his articles, he seems not to think through its
consequences for his theory of reasonable beliefs. So, Musgrave emphasizes on the
distinction between belief-content (what is believed) and belief-acts (acts of
believing). Now, the usual meaning of the word "act" is something we
do on our decision, something we are free to do or not to do. In this sense,
there are no belief-acts. What Musgrave describes as "belief-acts"
are really belief-states (psychological states of believing this or that) and
belief-events (changes of belief-states, getting into and out of different
beliefs).
If so, to be rational
about belief X we needn’t a theory showing that X (or our believing X) is
reasonable; all we need is to look for arguments pro and contra X.
We don’t – and can’t –
decide what to believe. But we can decide to look for arguments and evidences
here or there, to search for mistakes in our beliefs, to enter critical
discussion etc. Or we can decide not to listen to arguments and solve all disagreements
by resort to physical force. Popper wrote that the seed of his Open Society (the book where Popper first
describes the position he calls 'critical rationalism') was planted by young
Nazi in an encounter which took place in 1933 (the year that Hitler came to
power): Popper’s attempt to start discussion was met by a statement: "What,
you want to argue? I don’t argue: I shoot!" (Popper 1994, xiii). For
Popper, this was the most vivid representation of the alternative of
rationalism and irrationalism: rationalists are those who are willing to argue;
irrationalism despises arguing and prefers violence. This preference is where
we decide. And this is what CR, as general account of rationality, is about. To
recapitulate with Mark Notturno’s splendid explanation:
"Rationality, according to Popper, is
not so much a property of knowledge as a task for humans. What is rational is
not so much the content of a theory or a belief as the way in which we hold it.
We are rational to the extent to which we are open to criticism, including
self-criticism; and to the extent to which we are willing to change our beliefs
when confronted with what we judge to be good criticism. We are, in short,
rational to the extent to which we are willing to appeal to reasons and
arguments, as opposed to violence and force, to resolve our disputes." (Notturno 2002, 36)
10. (C)CR and Immediate Knowledge
In Common
Sense, Science and Scepticism, chapter 1, Musgrave describes an answer to
the problem of infinite regress of philosophers he calls
"dogmatists". The answer is a distinction between
immediate knowledge which does not require further justification
and
mediated knowledge which requires justification by immediate knowledge.
"Dogmatists"
differ as to what is a source of immediate
knowledge. Some (called 'empiricists') believe that it is experience
(perceptions), others (called 'rationalists' or 'intellectualists') – that it
is reason (its inherent undoubtable ideas).
I think that this
distinction is useful to understand (C)CR better. What is (C)CR’s view about immediate knowledge? Does it admit that
there is such a knowledge? If yes, what kind of knowledge it is – whether
perceptual or inherent to reason?
I think that it would be
correct to say that (C)CR agrees with "dogmatism" that there is
"immediate knowledge" which does not require further justification.
And it admits that there is "immediate knowledge" of both –
"empiricist" and "rationalist" – kinds. On
"empiricist" side, it is observational beliefs, on
"rationalist" side – logics, common sense epistemological
assumptions, inborn or learned expectations, basic moral ideas etc. Immediate
knowledge has as its sources experience, inborn expectations, intuition etc.
But, unlike "dogmatism", (C)CR does not dogmatize this
"immediate knowledge". (C)CR denies that it is certain or highly
probable (in the sense of probability calculus). (C)CR also denies that it is
unrevisable, that there is no way to check and correct it. Instead, (C)CR
contends that "immediate knowledge", as well as mediated, is fallible
and open to examinations, critical discussion and revision.
Besides, (C)CR
recognizes that, except some special cases (which may be of great importance –
as in the case of observational statements and scientific theories), there is
no neat demarcation between immediate and mediate knowledge. "Immediate
knowledge" can be criticized by other – immediate or mediate – knowledge,
and can be supported by yet other – immediate or mediate – knowledge. When some
points of "immediate knowledge" are challenged, we try to find
reasons pro and contra and weigh them and make our reasoned judgement. Those reasons
are just appellations to other – immediate or mediate – knowledge.
Hence, there can be no
overall inventory of "immediate knowledge" and relationship of its
components. And "immediate knowledge" has no absolutely "hard
core" – ideas and principles which are absolutely basic and, so, beyond
possibility of criticism and revision.
And, (C)CR contends that
all that "knowledge" is not knowledge in the sense of justified true beliefs, – it is conjectural knowledge without any
certificate of truth (certainty) or high probability.
Addendums
1. Definitional Character of CCR’s
Solution of the Problem of Limits of Rationality
On my view,
"circular reason" is not be permitted as a reason for general
theories of reasonable beliefs, as it is not permitted for all other theories
(beliefs). As all self-subsuming general theories of reasonable beliefs are
equally "reasonable in their own lights" and unreasonable in any
other lights, we have no reason for
preference of one self-subsuming general theory of reasonable beliefs over
any other.
Musgrave submits that
objection about circularity is not devastating for CR in his interpretation (CRM).
It wouldn’t be devastating if CRM would admit that for beliefs to be reasonable it is not
necessary that there were reasons to believe them. Or it wouldn’t be
devastating if CRM would admit that "unreasonable" beliefs can serve as proper reasons for
reasonable beliefs. (In this case, "unreasonable" means not contrary to reason but preceding reason.) But Musgrave never
admits this. Besides, he identifies "reasonableness" as having
reasons with "rationality":
“I assume that to hold a position rationally is
to have a reason for adopting it as true or believing it” (Musgrave 2007, 195)
So, on Musgrave’s definitional assumptions, believing CRM
is unreasonable, and critical rationalists hold CRM irrationally.
Believing CRM is nothing but irrational commitment which is not
principally different from commitment to the belief that everything granny told
me is true.
Bartley proposed CCR
exactly as the way to break this impasse.
And his solution of the problem is as well definitional. He proposes to use a
word 'rationality' in such a sense which preserves all valuable in the idea of
rationality and makes holding (C)CR rational. This sense is: for a person to
hold a position rationally it is necessary that he/she held it open to critical
discussion[8] (and it is not
necessary that he/she had a reason to believe it).
In this sense, the
preference for (C)CR needs not be irrational. It is rational as far as (comprehensively) critical
rationalist is open-minded about (C)CR, doesn’t avoid its critical discussion,
admits that it can possibly be mistaken and critical discussion can possibly
reveal its mistake. In this sense, rationality of my beliefs and preferences
is just the fact that I don’t aviod arguments which can possibly succeed to
persuade me out of these beliefs and preferences.
It is simple honesty to
admit that we believe some positions just because they seem true or plausible
to us (or because we have learned them from our social environment); and there
is nothing irrational about it, for it is unavoidable, on simple logical
reasons (those of justificatory infinite regress).
Popper had introduced
the term 'critical rationalism' (in "Open Society and Its Enemies")
explicitly as recognition of this fact. He went so far that declared CR as
variety of rationalism which frankly admit its own origin in irrational
decision, irrational faith in reason.
Bartley, when he
formulated CCR and argued that rationality can be comprehensive and
self-subsumptive did not contest
Popper’s admission that we can’t avoid accepting some positions (CR included)
without providing good reasons (justification) for them (or for believing
them). He admitted this as well as Popper. What Bartley did contest is that
holding some positions without providing 'justification' for them (or for
believing them) is irrational in the
meaning inherent to CR (and CCR). Bartley argued that as critical rationalism identifies rationality
not with justification but with critical discussion, then holding CR and
other unjustified positions is not irrational, provided "willingness to
submit these to critical considerations when we discover them or when they are
pointed out to us". Rationality as openness
to critical discussion is comprehensive and self-subsuming; rationality as
providing reasons-justifications (whether for statements or for believeing
statements) can’t be comprehensive and self-subsuming, – moreover, nothing
(exept tautologies) can be justified as demanded.
2. Cheese-Made Moon, Copernicus and Rationality
The interesting point to
discuss here is Musgrave’s example with the belief that Moon is made of green
cheese. Musgrave proposes this example to demonstrate craziness of a thesis that
a belief is reasonable if what is believed is criticizabele, which he
attributes to Bartley. In section 4 I have explained that this is a misattribution:
it is not Bartley’s thesis; CCR means nothing of the kind. But it is
interesting to discuss on this example what CCR really means.
Musgrave writes that the
belief that Moon is made of green cheese was thoroughly criticized and
"according to CR it is thoroughly unreasonable
for us to believe that the moon is made of green cheese, because that view has been thoroughly criticized and refuted."
I am not sure what Musgrave means. If he means that it would be unreasonable
for us to believe that the moon is made of green cheese if we admit that this view has been refuted, then, surely,
according to CCR it is unreasonable too. It would be irrational from CCR
viewpoint, for CCR takes principle of non-contradiction as inalienable from
rationality. For Bartley, as well as for Popper, logics, its central
core – principle of non-contradiction, is "the organon of criticisms".
(See "Retreat to Commitment", appendix "The Criticizability of
Logic") But may be Musgrave means something else? Let us abstract from
concrete example and consider the question in general terms. Musgrave’s
contention then can be reformulated as follows:
For any position X it is unreasonable to believe X if it was thoroughly
criticized and refuted.
I don’t know how to
understand it and how to relate it with CCR.
First, I don’t know what
does it mean for position to be "thoroughly criticized". The relevant
question is not "if X was thoroughly criticized?", but "if X was
soundly criticized?". But
(supposing criticisms were valid)
answering this question demands judgement
on truth of premises of criticisms (which, as we know, can never be
demonstrated). So, we never know if X was soundly criticized. We can – and,
rationally, must – believe that X was soundly criticized, if we see that
criticism of X is valid and if we believe
that its premises are true. But person who believes that X is true and that
premises of valid criticism of X are false is just as rational. At least, this
is CCR view, how I understand it. Otherwise, who is to adjudicate?
Principally, the same
goes with X being "refuted". Really, if a person admits that belief that Moon is made of green cheese
was refuted (as we all surely would admit), then
it would be irrational for that person to believe that Moon is made of green
cheese. But, if we imagine a person who believes that Moon is made of green
cheese, and have considered all proposed "refutations", and have
judged their premises false, then he rationally (though, I am sure, mistakenly)
believes that Moon is made of green cheese. At least, this is CCR view, how I
understand it.
This can seem rather odd
prima facia; but I think it is right,
for there is no general rule on how we are to decide between conflicting
beliefs. Surely, we can propose some specific rules for some important typical
cases (as those proposed by Popper for empirical sciences) and argue in favour
of those rules. But now, if some person is consistently
unpersuaded by our arguments, this does not entitle us to accuse him of being
irrational. We better try to explain our position better or find other
arguments which may succeed better. At least, this is CCR view, how I understand
it.
I think this position
can be made better undertood if we apply it to real situations of progress of
scientific knowledge. Suppose, there is a revolutionary scientific theory H
which have great appeal for elegance, simplicity etc., but which prima facia badly fits with known
empirical data, as far as it is understood at present moment. And there is old
theory G – very cumbersome, but much better fitting with known empirical data, as far as it is
understood at present moment. There are few scientists-revolutionaries who,
being well aware of empirical problems with H, nevertheless, believe that H is
true. And there are majority of scientists-conservatists who, being well aware
of appeals of H, believe that H is false, because it doesn’t fit with known
empirical data, as far as it is understood at present moment. Both sides act on
the same evidences and arguments. I think such situations were rather usual in
scientific revolutions; as example, we can take Musgrave’s description of
situation with heliocentric (H) and geocentric (G) cosmological theories before
Kepler in the article "Wandering Stars and Falling Stones" (Musgrave 2009b).
Now who – revolutionaries or conservatives – believe rationally? To say that
conservatives are irrational in their beliefs would directly contradict with
falsificationism. To say that revolutionaries are irrational in their beliefs
would mean that irrational beliefs play at least as important role in
development of scientific knowledge as rational ones, which is hardly acceptable
for rationalists. CCR’s answer, as far as I do understand CCR, is that both are
rational. With respect to 'revolutionaries': if you have a gut-feeling that
theory X is true and important, you may be quite rational to take a personal
risk to side with it against all known arguments and evidences, as far as you
don’t ignore them but admit thet they constitute a problem for the theory to be
solved in the future.
3. On Skeptical “Suspension of
Judgement”
In
Critical Rationalism (2007), section 6, third paragraph, Musgrave writes:
"If two competing hypotheses equally well
withstand criticism, the critical rationalist will accept neither of them. Here
the proper attitude is skeptical “suspension of judgement”, supplemented by
hope that future critical activity will enable us rationally to choose."
I think that it may be
the case if we have to do with two competing scientific theories which have
some difference in empirical content, but at present moment we have no means to
check areas of their divergences.
But generally, with
respect to other kinds of hypotheses, it is not the case, or not good enough
description. And it does not suit with what Musgrave writes in Can I Decide What to Believe?: we don’t
– and can’t – decide what to believe; rather we get into or out of believing in
truth of some positions under influence of evidences and arguments. Generally,
there is no objective way to establish: how well competing theories withstand
criticism? if they withstand it equally well? etc. It is always easy to propose
valid argument against any
hypothesis; the question is: if it is sound?
This depends on premises, but we can never establish their truth or high
probability.
So, what we really do
is: rely on our informal understanding, judgement, intuitive evaluation of
plausibility and importance. We investigate whether hypotheses at consideration
contradict some plausible statements (those we believe, or tend to believe, to
be true); make estimation of their plausibility and importance; and this
"cost" of acceptance (beliefs we must give up if we accept hypothesis
under consideration) weighs against respective hypothesis. We choose a hypothesis
trying to achieve two aims: 1) to make our system of beliefs internally
consistent and 2) to maximize "gains"/"costs" balance.
It is all for a person
to judge, what – in the light of arguments available – seems most plausible. We
consider arguments ‘pro’ and ‘contra’ each of competing hypotheses, weight them
on our “mental balances” against one another. And this process of considering
and weighing arguments either makes us to believe (or tend to believe – with a
deal of wavering) that one of competing hypotheses is true or the best
available, or leaves us in the state of suspension. This is the case with
everyone – whether critical rationalist, or justificationist, or irrationalist.
But, I think there is a difference: unlike justificationists and
irrationalists, critical rationalist believes that this process is the only way
to improve our knowledge and beliefs and social institutions, and wittingly
aspires to make this critical process as efficient as possible.
This situation happens
even in science, as in the case of quantum mechanics. There is the system of
equations which work very well for predictions, and several competing
interpretations of how it is to be understood. The grandest physicists, far
from taking attitude of skeptical “suspension of judgement”, defend different interpretations.
4. On Belief, Preference and
Classifying as True
Musgrave’s formulation
of CR says that "…if a hypothesis has withstood our best efforts to show
that it is false, then this is a good reason to believe it..." (Musgrave
2007, 181) It seems not quite agree with Popper’s own explanations about what
he thinks reasonable to believe about scientific theories. As Musgrave admits:
"Popper preferred to talk, not of
believing a theory or thinking it true, but of preferring one theory to another, or choosing one theory over another, or adopting one theory rather than another." (Musgrave 2004, 25)
Musgrave seem to explain
it as a purely verbal preference:
"Popper also said that the aim of
scientific theorizing is truth. But to prefer or to choose a theory, when your
aim is truth, is to believe it. …to classify something as true is also to
believe it." (Musgrave 2004, 25)
So, Musgrave qualifies
Popper’s replacement of "believing" with "preferring",
"choosing", "adopting" (and D.Miller’s "classifying as
true") as "terminological fads".
I think it is mistaken.
Popper’s "talk" about preferring
one theory to another, rather than believing
it to be true, is not merely verbal preference and not "the tendency
to squabble about words, in particular, about the word 'belief'" (Musgrave
2004, 24). Popper’s further explaination clearly shows that the difference is
not about the word 'belief', but about content
of belief: what it is that is
reasonable to believe about scientific theories?
“A justificationist may, however, object that I
have not really replaced one problem by another. He may argue, first, that instead
of 'reasons why we prefer one theory to another' I could have said 'reasons why
we believe that one theory is better
than another'. To the extent that this point is verbal, I readily grant it…
…there is a world of difference between a meta-theory that asserts that theory
A is better than a theory B, and another meta-theory that asserts that theory A
is, in fact, true (or 'probable')”. And there is a world of difference between
arguments that might be considered as valid or weighty reasons in support of
the one or the other of these meta-theories.” (Popper 2005, 23)
Now how about Musgrave’s
argument that "to prefer or to choose a theory, when your aim is truth, is
to believe it"? It depends on what "to believe a theory" means.
If it means "to believe that a theory is true sensu stricto", than Popper would disagree. To prefer or to
choose a theory, when your aim is truth, you needn’t necesssarily believe that
it is true sensu stricto; you may
believe that (as far as you can judge in the light of present state of critical
discussion) a theory has more to do with truth, is nearer to truth than its
known alternatives.
The difference is best
illustrated with an example of
There was never in the
history of science a theory more (or as) successful as
I think that this is
what Popper did mean when writing that "there is a world of difference
between a meta-theory that asserts that theory A is better than a theory B, and
another meta-theory that asserts that theory A is, in fact, true", and
that as far as scientific theories are concerned it may be reasonable to
believe the first, but not the second.
Up to the moment I did
eschew to use Popper’s own favourite term to designate a truth-approximation –
'verisimilitude' ('objective truthlikeness'). Notoriously, Popper’s attempt to
give this concept a strict logical definition has failed, and Popper have
readily recognized it. But he retained the concept of verisimilitude in a
weaker intuitive sense, in which, it seems, we can say that (as far as we can
judge in the light of present state of critical discussion) Newton’s theory
(though false sensu stricto) is very
near to truth in a very wide domain and Einstein’s theory (though we don’t know
if it is true, and have no good reason to believe that it won’t follow Newton’s
suit) is a further progress toward truth.
Musgrave notes the
connection between Popper’s "talk" about a preference for a theory
and the concept of verisimilitude, but is too ready to dismiss it as
"obfuscating smokescreens" which "cut no epistemic ice" (Musgrave 2007, 178). Also, Musgrave’s charging
Popper with "dropping truth in favour of verisimilitude" (Musgrave
2004, 25), "replacing truth with verisimilitude" (Musgrave 2007, 178)
seems to me not quite just: Popper did not propose believing in verisimilitude
as overall (for whatever statement) replacement for believing in truth. It may
be appropriate replacement with some kinds of statements, but not with others.
And this replacement, whenever it is proper, really means (as Popper readily
admits) sort of a "meta-belief"
– belief in the truth of meta-statement
that a scientific theory under consideration is (as far as we can judge in the
light of present state of critical discussion) the best approximation to truth
we have at our disposal.
I agree with Popper that
in the case of Einstein’s theory and quantum mechanics it is more reasonable to
believe that they are the best approximations to truth we have, than that they
are absolutely true. But is it always the case with empirical scientific
theories? In e-mail correspondence A.Musgrave objected that it is crazy to say
that no belief in the truth of a theory is reasonable, and proposed examples:
beliefs the earth is round not flat, that mammalian blood circulates, that
water contains hydrogen and oxygen. I agree, but must point out that Popper
surely din’t mean theories like that the earth is round not flat. (Even so,
this tstatement is true only if the word 'round' is used in a very fuzzy
meaning.) Popper’s explanation means only that kind of theories which are
unrestricted universal statements (like laws of physics). The statement that
the earth is round not flat is a singular statement. I am not sure about the
status of other two examples. Though a statement that mammalian blood
circulates can be given a form of
unrestricted universal statement, it seems, nevertheless, to be of some other
class, for it is derivative of unique historical process – evolution of life on
Earth. Also about water, if it is to be reformulated into more precize statement
that all water consists of H20 molecules; then this statement may be
definitional or, if water is defined by reference to its function for life, is
of the same kind as the statement about mammalian blood circulation.
Nevertheless, I am going to admit that probably there are some genuine
unrestricted universal empirical statements such that it is reasonable to
believe that they are true, – not merely well approximating truth. (Possibly, a
statement that H20-substance under normal pressure and temperature
between 273 K and 373 K is liquid will do.)
So, I tend to admit that
the answer to the question "What is more reasonable to believe about
accepted scientific theory: whether that it is true or that it is the best
approximation to truth we have?" is case-dependent. And, I think there is
no neat delimitation, though generally it seems that the more complex and
mathematized the theory is, the more reasonable is option for verisimilitude
rather than truth sensu stricto.
******
Yet another
"talk" is David Miller’s "classifying a theory as true".
Its meaning may belong to a different plane than that of beliefs. It may have
to do with the emphasis on rationality if
scientific method rather than of beliefs. Suppose, some general
methodological rules about tha way to test and accept scientific theories are
accepted (which, it may be said, presupposes the belief that they are
reasonable) by scientific community. Now scientists consider a theory A and
judge that according with accepted methodological rules it is to be
(tentatively) accepted as true (or classified as true). This judgement does not necessarily means that all
scientists who support it believe that A is true. Some of them may believe that
A is true; some others may believe that in the light of present state of
critical discussion A is the best approximation to the truth we have at our
disposal. Some may not believe even this. So, in science, tentative acceptance
or classification of a theory as true may be not quite the same thing as
believing that it is true. Surely, it is dependent on personal beliefs of
scientists, but may be not directly and unambiguously translatable into terms
of personal beliefs.
It well accords with
Popper’s statement that “The method of science is rational. It is the best we
have. It is therefore rational to accept its results…” (Popper 2005, 61)
Surely, it presupposes some beliefs – a meta-belief that the method of science
is rational, is the best we have, and that rationality is transferable from a
method to acceptance of its results.
5. Can We Replace False Theories of High Verisimilitude
with Partial Truthes?
In the article The 'Miracle Argument' for Scientific
Realism Musgrave proposes to replace “the verisimilitude project” with explanations
in terms of partial truth and approximate truth:
"Partial truth is not the same as verisimilitude. Verisimilitude is
closeness to the truth − the 'whole truth' − of a false theory taken as a whole.
Partial truth is just truth of parts... A false theory T might be successful
(issue nothing but true predictions) in a
certain domain D. Explain this not by saying that T is close to the truth,
but by saying that 'In Domain D, T' is true... It is the same with approximate truth... Approximate truth
is a species of partial truth. ... I have come to believe that the entire
verisimilitude project was a bad and unnecessary idea. ...what was the point of
the verisimilitude project? Precisely to explain how a false theory can have
partial success... We should eschew the near-truth of false wholes in favour of
the simple truth of their parts. We should explain partial success in terms of
truth of parts. ...it is simple truth, not verisimilitude, that is doing the
explaining." (Musgrave 2007b, 12-13)
I agree that what
remained of Popper’s conception of verisimilitude, after it was shown that his
attempt to make it into logically strict definition is failure, is just what Musgrave
describes as 'partial truth' and 'approximate truth'. But I don’t see why
combination of these two aren’t to be called 'verisimilitude' or 'closeness to
the truth' or 'nearing the truth'. But this becomes dispute over words which is
not worth taking. For me, to say that theory X is nearer to the truth than
theory Y is the same as to say that X has in it more 'partial truth' and
'approximate truth'; and I did always understand Popper’s conception of
verisimilitude in this sense.
Surely, I agree with
Musgrave that we needn’t a measure for false theorie’s approximation to the
'whole truth'; for all real needs I can imagine all we need is domain-related
estimation, estimation relative to the domain where theory is true or well
approximates truth. I think that Popper’s own final view of verisimilitude
accords with this, − see his remark in Appendix 2 "Additional
Remark" of "Objective Knowledge" about the problem of definition
of verisimilitude, that "perhaps it cannot be solved by purely logical
means but only by a relativization to relevant problems…"
But the problem is that we never know what is the domain of truth
(or good enough truth-approximation) of scientific theory.
If theory X has passed
all examinations up to present moment, it can be still reasonable not to
believe that X is absolutely true (i.e. true sensu stricto); to believe that it is more likely that X is only
partially or approximately true (i.e. false sensu
stricto); that X is false relative to
some unknown domains.
If theory X is refuted
and superceded by theory Y, which is accepted as true at present moment, then
we can determine domain where X is to be
accepted as true or approximately true only by comparision with Y, whose
truth-domain is unknown. But we have no good reason to believe that X is true (or approximately true) in all that
domain where it coincides with (or approximates) Y, for we have no good reason
to believe that Y is true in all that domain.
Literature
Artigas, M. 1999. The Ethical Nature of Karl Popper’s Theory
of Knowledge.
Bartley, W.W. 1984. The Retreat to Commitment.
Musgrave A., 2000. Common Sense, Science and Scepticism.
Musgrave, A. 2004. How
Popper [Might Have] Solved the Problem of Induction. Philosophy, 79(1): 19-31.
Musgrave, A. 2007.
Critical Rationalism. In E.Suárez-Iñiguez, ed., The
Power of Argumentation (Poznań Studies in the Philosophy of the
Sciences and the Humanities, vol.93).
Musgrave, A. 2007b. The
'Miracle Argument' for Scientific Realism. The
Rutherford Journal. http://www.rutherfordjournal.org/article020108.html
Musgrave, A. 2009a.
Experience and Perceptual Belief. In Rethinking
Popper (
Musgrave, A. 2009b. Secular Sermons.
Musgrave A., 2012.
Deductivism Surpassed: Or, Foxing in its Margins? Journal for General Philosophy of Science 43 (1):125-132.
Notturno, M. 2002. On Popper. Wadsworth Publishing Co Inc.
Percival, R. S. 2012. The Myth of Closed Mind.
Popper, K. 1974. Intellectual Autobiography. In P.A.Schilpp, ed., The Philosophy of Karl Popper.
Popper, K. 1977. The Open Society and Its Enemies. Vol.2.
Popper, K. 1979. Objective Knowledge: An Evolutionary
Approach.
Popper K., 1994. The
Myth of the Framework. ed. by M.A. Notturno. Routledge,
Popper, K. 2008. Conjectures and Refutations.
Popper, K. 2005. Realism and the Aim of Science.
Routledge,
[1] I think it is obvious that when
saying that "we stop criticizing … when we reach positions against which
we can find no criticisms", Bartley didn’t mean by 'criticism' of X any
statement which contradicts X (such criticisms are always easily found for any
X); surely, he did mean only plausible
criticisms – those Ys which we (tend to) believe to be true.
[2] Let us consider yet one Popper’s
statement:
“…since
we have to choose, it will be 'rational' to choose the best-tested theory. This
will be 'rational' in the most obvious sense of the word known to me: the
best-tested theory is the one which, in the light of our critical discussion,
appears to be the best so far, and I do not know of anything more 'rational'
than a well-conducted critical discussion.” (Popper 1979, 22)
It
is remarkable, that even this statement (though it is about scientific
theories) is equivalent with a very weak "theory" of reasonable
beliefs: it will be 'rational' to choose the theory "which, in the light
of our critical discussion, appears to be the best so far". There is no
criteria of reasonableness of belief at all. It says just: do enter into
critical discussion, consider arguments, and see where it gets you, what will
"appear to be the best so far".
[3] In the article "Deductivism
Surpassed: Or, Foxing in its Margins?" Musgrave writes about deductivist
reconstructions of ampliative reasoning: "The suppressed epistemic
premises are true and known to be so." I disagree: they often are true;
but how they are "known to be so"? (Musgrave 2012, …) They are
"known" no more than anything else, i.e, not "known" in
traditional sense (that of demonstrable truth or high probability), but
guessed, believed, judged to be true.
[4]
Musgrave råmarks that "Bartley accused
Popper of admitting this, and Popper accepted the accusation" (Musgrave
2007, 189). It is true, but it is to be remarked, that Bartley didn’t accuse
Popper for giving no reason for accepting CR. As far as I understand Bartley, his
view is that giving no reason (justification) for accepting X does not (or, at least, not always) mean that believing X is irrational
(unreasonable).
[5] Also, the fact that there are
(infinitely many) such (logically possible) epistemic principles relativizes CRM.
To "withstand severe criticism" means a very different things if we
take as dogmas beyond criticism everything said in a paper by Alan Musgrave or
everything granny told me or everything the Pope declared ex cathedra.
[6] Though not demonstrable (by
logically valid argument whose premises are known-demonstrable to be true),
truth and reasonableness of (C)CR may be well arguable – by arguments whose
premises are plausible or even, as Bartley would say, unproblematical at present. Arguments can succeed to persuade some
(though probably they won’t succeed to persuade some others) that (C)CR is true
and reasonable. And this is all arguments can do, generally.
[7] The
same point is emphasized by R.S.Percival in his book "The Myth of Closed
Mind"
[8]
Compare with Popper’s statements (accentuation mine):
“This
will be 'rational' in the most obvious sense of the word known to me: the
best-tested theory is the one which, in the light of our critical discussion,
appears to be the best so far…” (Popper 1972, 22)
“In The Open Society … I argued that one of the best senses of
"reason" and "reasonableness" was openness to criticism…”
(Popper 1974, 92)