-
APPLYING POPPER TOSOCIAL REALITIES:
Practical Solutions toTYRRELL BURGESS* Practical Problems
AMONG THE MOST OBVIOUS of intellectual developments in the last
fewdecades has been the expansion of the social sciences. The
numberof departments of economics, sociology, social and public
administrationhave multipled in institutions of higher education.
The number of stu-dents in degree courses in these fields has grown
astonishingly. At the sametime governments and other public
authorities, international agencies, andeven commercial
enterprises, have taken to commissioning quantities ofsocial
science research. The results of all this activity have been
disappoint-ing. Students have been disillusioned by the courses,
governments and otheragencies by the return on expenditure. The
social sciences, which seemto ofTer the possibility of solutions to
intractable social problems, appearinstead to be remote, academic,
and inconsequential.
There are innumerable examples of the social sciences failure.
Let us beginwith the desire to improve the outcomes of education.
Many children donot learn to read at school and it is widely
recognized that individual happi-ness and social peace may require
a remedy for this deficiency. Unhappilythe instinct of most social
scientists is to fmd out yet more about the chil-dren, their social
circumstances, family background, standard of accom-modation, and
so on. We thus know innumerable facts with which theinability to
read can be said to be associated. There is an industry devotedto
this sort of investigation: social class, family size, home
circumstances,race, and much else are included in what I may be
forgiven for callingmultiple digression analyses-all in order to
extend what we know about
Tyrrell Burgess is Reader in the Philosophy of Social
Institutions at Northeast LondonPolytechnic (London, England).
299
-
300 Et cetera FALL 1985
children who cannot read. Very little of this is concerned with
the questionof how children are to learn to read. The knowledge
from these investi-gations may or may not help: if it does, it will
be a lucky chance. Theseinvestigations are not specially designed
to ofFer practical help.
I would suggest that the way to solve the practical problem of
teachingchildren to read, whatever their circumstances, is to try a
number of readingmethods until we fmd one that works. The kind of
knowledge that is re-quired for this is quite different from the
kind of knowledge which is soughtby those who use the problem as an
occasion for research. The usual out-come ofthe kind of
investigation that I have just described is the assertionby
sociologists that improvement in reading is impossible until all
the otherassociated factors are mitigated-that we cannot change
anything unlesswe change everything.
Here is an example taken from David Donnison, then Director of
theCentre for Environmental Studies, introducing the longitudinal
study ofchild development undertaken by the National Children's
Bureau. (1) Pro-fessor Donnison writes:
Living conditions for families with young children probably vary
moregreatly-inequalities are sharper-than for any other type of
household. Manychildren live in the newest and leafiest suburbs
within easy reach of well paidjobs in expanding industries, new
schools and shops, extensive parks, andall the advantages of urban
and rural life. But many others live in overcrowdedquarters where
people are constantly on the move, social organisation is
weak,unemployment is rife, schools are old and under-staffed, and
there is no openspace or legitimate playground.
Such patterns are the outcome of a long history of economic and
socialdevelopment, reinforced or modified by the policies followed
by central andlocal authorities for family allowances, employment,
housing, transport, andland uses. Too often they are re-emphasised
rather than corrected by thedeployment of educational resources.
There is no time to be lost in settingabout the task of changing
them.
Professor Donnison concludes:How much do children learn? How far
behind the others do the weaker
performers fall? . . . What can we do to improve the
situation?The patterns glimpsed in the National Child Development
Study are so
deeply embedded in this country's economic and social structure
that theycannot be greatly changed by anything short of equally
far-reaching changesin that structure.
Professor Donnison is explicit that the performance ofthe weaker
learnerscannot be greatly changed unless we change the country's
economic andsocial structure. Professor Donnison is a well-meaning
man, but his mistakenview of social science leads him to write
absurdities and greatly demoralizesthose who might help children to
learn.
-
APPLYING POPPER TO SOCIAL REALITIES 30t
To take a further example, I have been present at a meeting to
discussthe improvement in the attractiveness of a school to local
parents, to whichthe education officer's contribution was, "The
school is unattractive becausethe pupils are mostly black and the
buildings are old, and we cannot doanything about either of these."
The knowledge here had all the fmalityof a solution when it should
have been part of the formulation of a realproblem for which a
solution was to be sought. Explanation has becomea substitute for
action.
Another set of misunderstandings which bedevil the contribution
of thesocial sciences is that which surroimds quantitative work. A
common mistakeis to confuse statistical significance with
significance. A British exampleof this which has had far-reaching
consequences is the formula for dis-tributing central government
support to local authorities. Part of this hasbeen called a "needs"
element, arrived at by running a multiple regressionanalysis on a
number of factors held to represent need and weighting
thedistribution according to the relative statistical significance
of the factors.The object of this is to produce a satisfactory
distribution based upon "ob-jective" criteria. Yet the consequence
is to produce a distribution whichvaries unacceptably from year to
year (calling for additional modifying for-mulae) and which bears
little relation to the actual needs of any individuallocal
authority. (2)
Finally, yet another ground for confusion derives from the
uneasinesswith which the social sciences accommodate philosophical
questions, inparticular questions of value. The place of value
judgments is a matterfor debate in both sociology and economics. It
takes the form, in economics,of a discussion about positive and
normative statements. As Harry Johnsonput it, positive economics is
concerned with how the economy works andnonnative with how it
should be made to work to maximize social welfare. (3)In his widely
used undergraduate textbook, Lipsey (4) argues that argumentsabout
positive statements can be settled by an appeal to the facts
whereasnormative statements depend upon value judgments and
disagreementsabout them cannot be settled in this way. By way of
example he suggeststhat the questions "What government policies
will reduce unemployment?"and "What policies will prevent
inflation?" are positive ones; whereas thequestion "Ought we to be
more concerned about employment than aboutinflation?" is a
normative one. Lipsey immediately recognizes that thedistinction
may break down, and in a half-page footnote he suggests
hownormative statements may depend upon or lead to positive
ones-for exam-ple, "Unemployment is worse than inflation because
the (measurable) ef-fects of unemployment on human beings are
judged by the majority ofadult citizens to be . . . more serious
than the (measurable) effects of infla-tion." This has now become
such a mixture of positive and normativestatements as to be merely
confusing.
There can be many other examples of ways in which the developing
social
-
302 Et cetera FALL 1985
sciences bewilder and mislead professionals and public alike. I
do not believe,however, that this is inevitable. The bewilderment
stems from mistakentheory and practice, and both can be
corrected.
IIThe best hope of corrertion lies through developing the work
of Karl Popper.His epistemology, developed initially in relation to
the physical sciences, (5)is equally promising in the social
sciences. (6) In particular his advocacyof "piecemeal social
engineering," his plea to "minimize misery," and hisunderstanding
of the importance of institutions together offer a basis forthe
harnessing of social science for social improvement. Unfortunately
theseinsights have been largely neglected. This is partly because
his ideas arescattered throughout what is now a very large volume
of work.
Popper's most convenient short account of his view of the social
sciencesis set out in a relatively recent symposium, (7) summarized
as an appendixto the present paper (Popper's "twenty-seven theses"
on the logic of thesocial sciences).
Popper's main thesis is one which he summarizes elsewhere in the
schemaPi^TT->EE->P2: scientific discussions start with a
problem (Pi) towhich we offer a tentative solution, or theory (TT);
this theory is thencriticized to eliminate error (EE); the theory
and its critical revision leadto new problems (P2). It seems to me
a more fruitful explanation thanothers of how knowledge advances,
and in particular it avoids the problemof the logical impossibility
of induction. It is important to realize, however,that it is a
logical explanation, not a psychological one. It does not implya
belief that that is what all individual scientists consciously do.
Indeedit accommodates the immense variety of practices of
individual scientists,including the random, accidental, and
creative insights which are indispens-able to human progress. Such
strokes of genius can be readily dealt withand made more fruitful
if they are regarded as solutions to problems (TT).
There is indeed an implied lesson to be learned from Popper's
schema,though so far as I know he himself has not made this
explicit. It is thatknowledge will be more quickly advanced if
scientists are aware of the logicof the process and if their
practice is in tune with it. In particular it suggeststhat as much
care must be taken in formulating problems as in searchingfor
solutions, and that solutions must be capable of being tested. All
threesteps require creative insight.
I believe that it is in the first step-the initial formulation
of problems-that the social sciences are at their weakest, and that
this weakness is respon-sible for the social sciences' continuing
ineffectiveness. Although Popper'stwenty-seven theses represent the
strongest statement about what the socialsciences could be, the
theses themselves contain echoes of this weaknesswhich diminish the
force of Popper's concluding suggestion for the theses'
-
APPLYING POPPER TO SOCIAL REALITIES 303
further development (see Popper's "suggestion" at the end of the
appendixto the present paper).
From the fourth thesis it is clear that in discussing the social
sciencesPopper is concerned with knowledge and with the theory of
knowledge.He argues that so far as one can say that knowledge
starts from something,one can say that it starts, not from
observations, perceptions, or the collec-tion of data or facts, but
from problems. It starts from the tension betweenknowledge and
ignorance and thus meets the demand built up in the firstthree
theses: no problems without knowledge; no knowledge without
prob-lems; no problems without ignorance. Each problem arises from
the dis-covery that something is not in order with our supposed
knowledge: thatthere is an apparent contradiciton between our
supposed knowledge andthe supposed facts. The problems from which
knowledge starts are thosewhich arise when what we think we know
does not fit with what we under-ftand to be the facts.
This is made explicit in the fifth thesis, in which Popper
asserts thatin the social sciences, as in all sciences, success,
interest, and fruitfulnessdepend on the significance of the
problems and the honesty and simpli-city with which we tackle them.
In this, he continues, we are not confinedto theoretical problems:
"Serious practical problems, such as the problemsof poverty, of
illiteracy, of political suppression, or of uncertain legal
rightswere important starting points for research in the social
sciences."
This thesis thus allows that the problems of social science may
be practicalones: it cites some practical problems that are indeed
"serious." Yet in thisformulation, these serious practical problems
serve simply as the startingpoint for research and thus presumably
to the extension of knowledge. AsPopper says, they "led to
speculation, to theorizing, and thus to theoreticalproblems."
Practical problems are viewed, in other words, as a spur to
thesearch for truth rather than to practical improvement.
inI find this unsatisfactory. This is partly because serious
practical prob-
lems demand better of us, and partly because the formulation
glosses overa crucial problem for the social sciences, that is the
relationship betweenknowledge and pratice. I do not think that
Popper himself has ever expli-citly worked out quite what this
relationship should be. He is not, of course,uninterested in
practical matters. On the contrary, his plea to minimizemisery is
as passionate as "ecrase rinfame"-and his concept of
piecemealsocial engineering is the world's best hope of achieving
it. But he himselfhas confessed that he has not followed this
through.
In the hands of most social scientists, however, the flaw in
Popper's fifththesis becomes extremely damaging. It leads to the
relegation of seriousproblems to the status of mere occasions for
research. It relieves the researchitself from the discipline of
relevance. It encourages the endless elabora-
-
304 Et cetera FALL 1985
tion of detail and the "replication" of research findings. It
also leaas to"Utopian solutions" of the worst kind, like the one
suggested by DavidDonnison quoted earlier.
I fear that an unintended consequence of the fifth thesis will
be tostrengthen these tendencies in the social sciences. What we
need insteadis a way of seeking social knowledge explicitly as part
of a solution to socialproblems. The knowledge that we seek in
these circumstances will be quitedifferent from that which is
sought in pure research.
This distinction is obscured in the fifth thesis. And it is not
clear whatis the status of the problems which are described as
practical. It is surelynot the same as that of the problems
mentioned in the fourth thesis-which arise from an apparent
contradiction between our supposedknowledge and the supposed facts.
The problem of poverty is not a prob-lem of knowledge or a problem
of fact. Of course, we need to know howmuch poverty there is-and
there may be more or less than we think. Weno doubt need to know
the incidence of poverty and the nature of its con-sequences. But
this is not what we mean by the problem of poverty. Povertyitself
is not a problem. It becomes a problem only if we accept or
decidethat something must be done about it: the problem of how to
get frommore poverty to less.
(This is somewhat clarified in the fourteenth thesis. Here
Popper dis-tinguishes between the question of the truth of an
assertion-of its relevance,interest, and significance to the
problems of knowledge which we investi-gate-and the question of an
assertion's relevance, interest, and significancefor
extra-scientific problems. Examples of these latter problems are
thoseof human welfare or national defense, of an aggressive
nationalist policy,industrial expansion, or the acquisition of
personal wealth. Popper saysthat such extra-scientific interests
cannot be eliminated or prevented fromaffecting the course of
scientific research-in the natural as in the socialsciences. But he
believes that it is possible, important, and peculiar to scienceto
differentiate between these interests "which do not belong to the
searchfor truth" and the "purely scientific interest in the truth."
The problemof poverty is of extra-scientific interest. Popper
himself gives the problemof human welfare as an example of such an
interest, so I think we maybe justified in assuming that Popper
would regard the problem of povertyin the same light-when we want
to do something about it and it becomesan issue of public policy
[just like Popper's other examples of nationaldefense and
industrial expansion].)
IVIn the light of this discussion, I propose an addition to
Popper's twenty-seven theses which will help to clarify the
relationship between questionsof truth and issues of public policy.
My suggestion emphasizes the impor-tance of care in the formulation
of problems and of restraint in proposed
-
APPLYING POPPER TO SOCIAL REALITIES 305
solutions. My addition to Popper's theses might perhaps be
numbered 14aand it would go something like this:14a. There are
different kinds of problems which demand different kinds
of solutions and different kinds of tests. One can distinguish,
for exam-ple, scientific, engineering, formal and philosophical
problems. Thelogic of the process (Popper's Pi->TS->EE-*P2
spelled out in thesixth thesis) remains the same whatever the kind
of problem.
Let me elaborate. In Popper's formulation of the logic of
discovery heuses, interchangeably and often all together, a number
of words for thesecond term of his schema (above): theory,
solution, hypothesis, conjec-ture. I presume that this practice
derives from his impatience with discus-sions of meaning which he
regards as trivial. He does not wish under-standing to be limited
by definitions. What is more, a part of what he isarguing is that
theories are solutions to problems, and solutions, even topractical
problems, are theories. But I find that it is useful to carry
thesedifferent ideas into the first term of the schema and
distinguish differentkinds of problems.
The first consists of problems of what is the case, and why: we
may callthese scientific problems. Their nature is well illustrated
in the twenty-seven theses. Problems arise when facts contradict
our expectations. Tenta-tive theories enlarge our understanding of
what is the case. As Popper says,this applies in the social
sciences as in all sciences.
The problems which I call engineering problems are different.
They arisenot from difficulties about what is the case, but from
the need to get fromone state of affairs to another-to use the
illustrations of one engineer, toget from one side of a river to
another, from bread to toast. Examples from"social engineering"
include the need (desire, policy) to change people'scircumstances
from poverty to a modest competence or to make illiteratechildren
literate. As I said earlier, the "problem of poverty" arises whenwe
propose to do something about it.
Formal problems are different again: they arise in closed
systems, likemathematics or chess, and do not necessarily relate to
reality at all.
The fourth group of problems, which I have called philosophical,
arenot so clearly described as the other three. One is tempted to
assert thatthey include all these problems which cannot be
elsewhere classified, andthese include problems of method, policy,
value, ethics which may ariseout of the other problems which we
tackle or, in their turn, give rise tothem. This present discussion
of different kinds of problems is itself anattempted solution to a
problem in the philosophy of social science.
The examples given at the beginning of this paper, of confusion
andstultification in the social sciences, arise from a failure to
be clear aboutthe kind of problems being tackled, in particular a
confusion of scientificand engineering problems. It is a confusion
of two traditions of learning.
-
306 Et cetera FALL 1985
which I have elsewhere characterized as the "autonomous" and the
"ser-vice" traditions. (8)Those who work in the autonomous
tradition speak ofthe "preservation," "extension," and
"dissemination" of knowledge, often"for its own sake." They speak
of pursuing the truth wherever it may lead,regardless of the
consequences. In its extreme version, as advanced for exam-ple by
George Steiner, (9) it would place the pursuit of truth above
theexistence of the human race. The service tradition starts from a
differentplace. Its activity is not self-justifying but explained
and defended in termsoutside the pursuit of knowledge, usually in
terms of a change that mightbe accomplished in an individual or in
society. Both these traditions areof enormous value, and both ought
to be protected. It is of paramountimportance, too, that they be
distinguished, and that we should always beclear, in any particular
enquiry, which of them forms the basis for our work.Teaching
children to read is not a matter of pursuing the truth
aboutchildren, it is a matter of trying a number of reading
methods. It is anengineering, not a scientific problem.
We can similarly understand the nature of the difficulty about
statisticalsignificance by recognizing that it is a matter of
imagining that the solu-tion to a formal problem is apt for a
practical problem. The Greeks illu-strated the confusion very well
with their famous problem of Achilles andthe tortoise. Achilles ran
twice as fast as the tortoise, so the tortoise hada mile start;
when Achilles had covered the mile, the tortoise had gone halfa
mile; when Achilles had done that half mile, the tortoise had done
a quarterof a mile. . ., and so on, Achilles never passed the
tortoise. There is nothingwrong with the mathematics; it just has
nothing to do with reality. A similarmisunderstanding lies at the
heart of much of the quantitative work in thesocial sciences.
Similarly, in the economic example given earlier, the
distinction betweenpositive and normative statements in the social
sciences is unhelpful. Whatwe have instead are different kinds of
problems. There are scientific prob-lems like "What is the
incidence of unemployment and how does it arise?"and there are
problems of social engineering like "How do we get frommore
imemployment to less?" There is nothing "normative" about the
secondquestion: it is a matter of choosing a solution and testing
whether or notit is effective, in other words by an appeal to the
facts. Similarly the ques-tion of whether unemployment is worse
than inflation is clarified by identi-fying what problem (of social
engineering) one is concerned to solve. Ifour problem is to get
from more human misery to less, we should test (byreference to the
facts) which of a number of policies (including
reducingunemployment or curbing inflation) are most apt. The
determination toreduce human misery itself may rest upon a value
judgment. It may bethat a value judgment determines which of our
infinity of possible prob-lems we determine to tackle. But in this
respect problems of (social) engineer-ing are no different from
problems of (social) science. What is more, our
-
APPLYING POPPER TO SOCIAL REALITIES 307
determination may not depend upon a value judgment at all but
upon thetestable theory that a reduction in human misery will lead
to the Govern-ment's re-election.
It is in this context-in the recognition that we have different
kinds ofproblems-that Popper's statements about values are of such
importance.Whether our problems are those of science or of
engineering we must dis-tinguish between scientific and
extra-scientific values. In this there is nodistinction between
"what is" and "what is to be done." The question ofwhether we ought
to do this or that remains for scientists as well as forengineers.
Both are under an equal obligation to be clear about such
valuequestions and about their answers to them. But in each case it
is possibleto bring one's values into the realm of rational
discussion by recasting themin the form of problems which are
amenable to testable solutions. It is inthis way that social
science and social engineering can be advanced.
This brings us to Popper's "suggestion" that the fundamental
problemsof a purely theoretical social science are the situational
logic of and theoryof institutions and traditions. The task of this
theoretical social science,as Popper puts it elsewhere, (10) is "to
try to anticipate the unintendedconsequences of our actions."
The task of social engineering is different. It is to devise
solutions tosocial problems and to test these solutions. Indeed it
may have to reformu-late social problems so that solutions are
possible. And the solutions willalways be institutions which are
testable for their aptness and success.
The relationship between the two is like that between physical
scienceand engineering. Engineering problems may be solved with or
withoutthe aid of theory. The engineer cannot fly in the face of
the physical facts,though he may propose ways of overcoming them. A
developed social sciencemay not only improve understanding but also
give grounds for a socialengineering that in tackling our urgent
problems may be fruitful, apt, andfree from harm.
NOTES AND REFERENCESL Ronald Davie, Neville Butler, and Harvey
Goldstein, Prom Birth to Seven, a report
of the National Child Development Study (Longman in association
with the Na-tional Children's Bureau, 1972).
2. See Tyrrell Burgess and Tony Travers, Ten Billion Pounds:
Whitehall's Takeover of theTown Halls (Grant Mclntyre, 1980).
3. Harry Johnson, The Economic Approach to Social Questions, an
inaugural lecturedelivered October 12, 1967 at the London School of
Economics (Weiderfeld andNicolson, 1968).
4. Richard G. Lipsey, Positive Economics, founh edition
(Weidenfeld and Nicolson, 1975).5. See especially K.R. Popper, The
Logic of Scientific Discovery (Hutchinson, 1959).6. See especially
K.R. Popper, The Open Society and Its Enemies, fifth edition
(Routledge
and Kegan Paul, 1966).
-
308 Et cetera FALL 1985
7. K.R. Popper, "The Logic of the Social Sciences," in Theodor
W. Adorno et al,The Positivist Dispute in German Sociology, Glyn
Adey and David Frisby, eds.(Heinemann, 1976).
8. Tyrrell Burgess, Education After School (Gollancz and
Penguin, 1977).9. George Steiner, "Has Truth a Future?," The
Listener, January 12, 1978.
10. K.R. Popper, "Reason or Revolution," in Adorno, op.cit.
AppendixPopper's "The Logic of the Social Sciences"
The "twenty seven theses" in which Popper sets out the logic
ofthe socialsciences can be baldly summarized as follows (thesis
numbers in brackets).
We know a great deal, yet our ignorance is boundless: with each
problemwe solve we discover new problems and undermine previous
certainties [1,2].The logic of knowledge must accommodate this
tension between knowledgeand ignorance [3]. Knowledge starts from
problems, and our achievementin advancing knowledge is
proportionate to the significance ofthe problemswe tackle
[4,5].
The main thesis is that the method of social science consists in
trying ten-tative solutions to problems: these solutions are
proposed and criticized, andcriticism consists of attempts at
refutation. If the solution is refuted wetry again; if it survives
we accept it temporarily, as worthy of being criti-cized. This is a
consciously critical development of the process of "trialand
error." The objectivity of science lies in the objectivity of
criticalmethod. [6]
The tension between knowledge and ignorance leads to problems
andtentative solutions, a thesis which contrasts strongly with the
"misguidednaturalism" of induction [7]. The recent preeminence of
anthropology, analleged descriptive, objective science, over
theoretical sociology is a vic-tory for misguided naturalism. Even
though a "subject" is simply an arti-ficially demarcated
conglomerate of problems and solutions, the continuingvictory of
anthropology, of misguided naturalism, woiild be a disaster.
Thereis no such thing as a purely observational science: there are
only sciencesin which we theorize [8-10,21].
The objectivity of a science does not depend on the objectivity
of thescientist, but upon a critical tradition: not on individuals
but on the socialresults of mutual criticism. Objectivity is to be
expressed in terms of socialideas, like competition between
individuals and "schools," tradition (espe-cially the critical
tradition), social institutions like publishers, and the powerofthe
state in tolerating debate [11-13].
In critical discussion we can distinguish the question ofthe
truth of anassertion, its relevance, interest, and significance to
problems of interest;and the question of its relevance, interest,
and significance for various extra-scientific problems [14].
-
APPLYING POPPER TO SOCIAL REALITIES 309
The most important function of purely deductive logic is that of
an in-strument of criticism. Deductive logic is the theory of the
transmissionof truth from the premises to the conclusion, and it is
also the theory ofthe retransmission of falsity from the conclusion
to at least one of thepremises. It is the theory of rational
criticism, and deductive systems-that is, what we work with in the
sciences. The concept of truth is in-dispensable for this approach
[15-20].
Sociology is autonomous in that it can and must make itself
indepen-dent of psychology (which itself is a socal science) and in
that we cannotreduce the sociology of understanding to psychology
[22-24].
There is a purely objective method in the social sciences, the
methodof objective understanding or situational logic. The
explanations of situa-tional logic are rational, theoretical
constructions and, in their over-simplification, are false-but with
considerable truth content. Situationallogic assumed a physical
world in which we act, including a social worldand social
institutions [25-27].
These theses lead to a "suggestion" for the social sciences.
This is:We may, perhaps, accept provisionally, as the fundamental
problems of a
purely theoretical sociology, the general situational logic of
and the theory ofinstitutions and traditions. This would include
such problems as the following:1. Institutions do not act; rather,
only individuals act, in or for or through
institutions. The general situations logic of these actions will
be the theoryof the quasi-actions of institutions.
2. We might construct a theory of intended or unintended
institutional con-sequences of purposive action. This could also
lead to a theory ofthe crea-tion and development of
institutions.