Social Research Glossary A B C D E F G H I J K L M N O P Q R S T U V W X Y Z Home
Citation reference: Harvey, L., 2012-24, Social Research Glossary, Quality Research International, http://www.qualityresearchinternational.com/socialresearch/
This is a dynamic glossary and the author would welcome any e-mail suggestions for additions or amendments.
|
|
_________________________________________________________________
Falsificationism
Falsificationism is an inductivist approach to knowledge production that basically asserts that theories cannot be proved but that theories or hypotheses can be disproved, or falsified
Introduction
Falsificationism asserts that theories cannot be proved but that theories or hypotheses can be shown to be false.
There are three main variants: dogmatic falsificationism; naive methodological falsificationism; and sophisticated methodological falsificationism.
Methodological falsificationist approaches to the production of scientific knowledge develop revolutionary conventionalism beyond Duhem's 'simplistic' model, combining elements of justificationist accounts with the basic conventionalist principle.
The work of Popper is central to methodological falsificationism. He 'progressed' from the naïve to the sophisticated approach although he tended to confound the two approaches in his work. Thus, this account follows the extrication of the central tenets of both as reconstructed by Lakatos, and utilises the categories that Lakotos has defined.
Introduction
Dogmatic falsification takes as central the falsifying element of inductivism. The one element that 'classical' empiricists are agreed on is that while inductivism is unable to prove a proposition it is able to falsify a theoretical statement on the basis of observation.
The hallmark of dogmatic falsification is then the recognition that all theories are equally conjectural. ‘Science cannot prove any theory ... but it can disprove (it can perform with complete logical certainty [the act of] repudiation of what is false’ (Medawar (1967), p. 144); that is, there is an absolutely firm empirical basis of facts that can be used to disprove theories.
Thus, inductivism demands that scientific honesty requires specifying, in advance, an experiment such that if the result contradicts the theory the theory has to be given up. 'Criteria of refutation have to be laid down beforehand: it must be agreed which observable situations, if actually observed, mean that the theory is refuted' (Popper (1963) p. 38, footnote 3) ‘... according to the logic of dogmatic falsificationism, science grows by repeated overthrow of theories with the help of hard facts.' (Lakatos, 1970, pp 96-97, emphasis in the original)
Yet despite this bold formulation, and re-emphasis, ‘a kind of 'turning inductivism on its head', the dogmatic falsificationist position provides no advance at all on the problems facing the inductivist. Although, dogmatic falsificationism makes no attempt to prove theories, accepting, as it does the fallibility of all theories, it still assumes the infallibility of the empirical base, to which it addresses all scientific knowledge.
Problems with dogmatic falsificationism
Lakatos, therefore, correctly indictes dogmatic falsificationism for failing to take into account the theory-laden nature of observation and for assuming the validity of the doctrine of observational or experimental proof. Even given that there may be a borderline between theory and hard, observational fact, to assume that observation statements are basic and 'proved' is to commit a fundamental error of logic, it is to confuse singular, particular statements with generalisable propositions. No factual proposition is derivable from an experiment, only propositions can be derived from other propositions. There is no way observation can indubitably ascertain the truth value of such propositions.
Lakatos goes on:
If factual propositions are unprovable then they are fallible. If they are fallible then clashes between theories and factual propositions are not 'falsifications' but merely inconsistencies. Our imagination may play a greater role in the formulation of 'theories' than in the formulation of 'factual propositions', but they are both fallible. Thus we cannot prove theories and we cannot disprove them either. The demarcation between the soft, unproven 'theories' and the hard, proven, 'empirical basis' is non-existent: all propositions of science are theoretical and, incurably, fallible. (Lakatos, 1970, pp 99–100)
Dogmatic falsification has one other profound disadvantage, its inability to account for the development and persistence of the more important scientific theories. For, according to Lakatos, such theories as Newton's gravitational theory, fail, specifically, to forbid any 'observable state of affairs'. In recounting his infamous, hypothetical story of the discovery of a new planet he suggests that such a respected theory will be surrounded by so many loopholes that no single instance will be able to falsify it.
In general, the more significant the theory the less dogmatic falsificationism is able to contend with it, indeed, the less the dogmatic falsificationist will regard it as scientific. The fewer observations needed to refute a theory, then the more scientific it is by dogmatic falsificationist standards.
Thus 'all crows are black', 'all planets move in circles', 'all planets move in ellipses' are propositions that comprise a descending order of scientificity for dogmatic falsificationists. Furthermore, all are infinitely more 'scientific' than Maxwell's or Einstein's or Newton's theories.
Thus Lakatos concludes:
if we accept the demarcation criteria of dogmatic falsificationism, and also the idea that facts can prove 'factual' propositions, we have to declare that the most important, if not all, theories ever proposed in the history of science are metaphysical, that accepted progress is pseudo-progress, that the work done is irrational. If the second assumption is denied then we certainly end up in complete scepticism: then all science is undoubtedly irrational metaphysics and should be rejected. scientific theories are not only equally unprovable, and equally improbable, but they are also equally undisprovable. But the recognition that not only the theoretical but all the propositions in science are fallible, means the total collapse of all forms of dogmatic justificationism as theories of scientific rationality. (Lakatos, 1970, p. 103).
Naive methodological falsificationism
Introduction
Naive methodological falsificationism arose as a criticism of inductivism and of Duhemian conventionalism. As Lakatos put it:
Contemporary [naive methodological] falsificationism arose as a logico-epistemological criticism of inductivism and of Duhemian conventionalism. Inductivism was criticised on the grounds that its two basic assumptions, namely that factual propositions can be 'derived' from facts and that there can be valid inductive (content-increasing) inferences, are themselves unproven and even demonstrably false. Duhem was criticised on the grounds that comparison of intuitive simplicity can only be a matter for subjective taste and that it is so ambiguous that no hard-hitting criticism can be based on it. (Lakatos, 1975, p.7)
What naive methodological falsificationism did was to shift the emphasis from acceptance of theories by convention to allowing singular ('basic') statements to be accepted by convention. Coupled with that was the demand that the procedure for adopting a 'basic' statement be made explicit. Naive methodological falsificationism like all conventionalist approaches denied the infallibility of the empirical base and perceived science as starting from problems rather than observation. While problems stem from observation, they are only problems in as much as they are observed disconfirmations of theory.
Thus naive methodological falsificationism posits progress in science as from problem, to speculative hypothesis, to criticism of the hypothesis, to falsification of the hypothesis and thence to a new problem.
Naive methodological versus dogmatic falsificationism
Falsification for the naive methodological falsificationists is, however, epistemologically quite different from that of the dogmatic falsificationist.
The naive methodological falsificationist makes some singular statements unfalsifiable by fiat. The acceptibility of the taken-for-granted nature of such basic statements is that, according to Popper, relevant techniques exist that enable those who have learned them to decide that the statement is 'acceptable'. Once 'accepted' such a statement is a 'basic' or 'observational' statement. The inverted commas imply that there is nothing immutable or basic about 'basic' statements. 'Basic' statements do not carry the same weight as the basic premises of the dogmatic falsificationist, to whom such ‘basicness’ is the result of an unfalsifiable empirical grounding of observation. In short dogmatic falsificationism only falsifies by empirical observation and does not accept any statements (by fiat or empirically).
Thus, the 'basic' statements of naive methodological falsificationism are fallible, hence the whole enterprise is on shaky foundations. Naive methodological falsificationism takes well-corroborated theories as unproblematic background knowledge and assesses the 'observational' statements in the light of them. To test a theory in the light of certain observations will necessitate taking-for-granted some aspects of knowledge about the world (which may prove to be incorrect in the long run but which at the time were well corroborated) and testing on the basis of this taken-for-granted unproblematic grounding.
The methodological falsificationist might have been 'wrong' to assume the unproblematic nature of the grounding but is not deluded that the testing is grounded in fact and therefore proven (false).
Further, by taking the corroborated theories for granted, the methodological falsificationist can test a far wider range of 'observational' statements than is possible for the dogmatic falsificationist whose theories need to be strictly observational.
The methodological falsificationist needs to clearly distinguish between which statements are under test and which are taken as unproblematic background knowledge. This clearly involves a considerable conventional element, not least of all in the acceptance of the 'truth-value' of 'basic' statements. Naive methodological falsificationism demands that should a statement be rejected, the premises and basis for testing the statement and alternative formulation should be made clear. Science thus advances not on certainty but on the best available corroborations of the time.
It is not satisfactory for a refutation to be couched in subjective feelings of doubt or in conviction about the 'correctness' of the alternative formulation. Yet, the very incorporation of the conventionalist principle, which involves decisions by the scientist, that includes a subjective element. However, it is one that is 'objectively' based because it derives directly from the state of objective scientific knowledge appertaining at the time. As Popper is aware:
the empirical basis of objective science has thus nothing 'absolute' about it. Science does not rest upon solid bedrock. The bold structure of its theories, as it were above a swamp. It is like a building errected on piles. The piles are driven down from above into the swamp, but not down to any natural or 'given' base. And if we stop driving the piles deeper it is not because we have reached firm ground. We simply stop when we are satisfied that the piles are firm enough to carry the structure, at least for the time being. (Popper, 1959, p. 111)
Where 'empirical basis' clashes with theory, then, dogmatic falsificationism declares the theory disproven. Naive methodological falsificationism sees the theory as refuted but not disproved. Thus methodological falsificationism separates disproof from rejection, which the dogmatic falsificationist necessarily conflates. Unlike dogmatic falsificationism, methodological falsificationism may reject a theory as 'false' even if it is true. It is a risk of the enterprise; the inevitable spin-off of a model that hopes to retain rationality yet recognises the fallibilism of theory and observation. Unless there is a way to eliminate some theories, scientific progress will, according to the falsificationists, be chaotic.]
In summary, the methodological falsificationist calls for the ruthless disposal of rejected theories if science is to progress. In so doing, the methodological falsificationist has separated disproof from rejection which the dogmatic falsificationist necessarily conflates. Thus the methodological falsificationist regards theories as scientific if they have an 'observational' ('empirical') basis and are 'refutable', whereas the dogmatic falsificationist demands that a theory must be grounded in immutable observational base if it is to be regarded as scientific.
The methodological falsificationist position permits probabalistic theories to qualify as scientific if decision criteria about their 'falsifiability' can be drawn up, i.e. if levels of improbability may be specified in advance. This is crucial to much quantitative thinking in the social sciences. A disadvantage relates to 'ceteribus paribus' [other things being equal] clauses in a theory. Such clauses, if they are to be 'relegated' to the unproblematic background knowledge, lead to a refutation of a theory should 'empirical' evidence be at variance with the theory. However, this does allow complex theories, such as Newton's theory of gravitation, to be included as scientific for the methodological falsificationist, unlike his dogmatic counterpart who is unable to admit such theories to science.
See also justificationism
Inadequacies of naive methodological falsificationism
The naive methodological falsificationist contrives a way to save falsificationism from the fallibilism that destroyed dogmatic falsificationism, yet retaining falsificationist's creed of scientific honesty, namely that should the result of an experiment specified in advance be such that it falsifies the theory, the theory should be given up. This process leads to risky decisions, risks that, on the face of it, seem to be somewhat reckless. The history of science does not seem to bear out this potential high risk decision making. In fact, the 'principles of progress' propounded by the naive methodological falsificationist are at variance with the 'actualities' of the history of science.
The history of science is full of examples of notoriously slow change. There was nothing daring about the one hundred and fifty years it took for the Copernican system to become established, and Mercury's perihilion was an 'observed fact' for eighty five years before Newton's theories were to come under serious attack. The naive falsificationist position implies a confrontation between theory and experiment and assumes that falsification is the important arbiter of this confrontation. The history of science is portrayed by a succession of refuting instances and new hypotheses, that the turning points in the progress of science are 'crucial' experiments. The history of science, however, shows that most historically reconstituted confrontations are between at least two theories in the light of some experimental information and that some of the more interesting experiemental results have acted to rather than falsify results.
Furthermore, the naive methodological falsificationist position, by assuming that confrontation is between theory and observation, makes no progress towards confronting the problem of the theory-laden nature of observation. On the face of it, naive methodological falsificationism seems to do nothing more than admit the problem of fallibilism without going any further than dogmatic falsificationism in solving the problem.
What has happened is that the naive methodological falsificationist, in overcoming the conventionalist premise that no experiemental result can ever kill a theory, removed the auxilliary (supportive) hypotheses to 'the realm of unproblematic background knowledge, eliminating them from the deductive model of the test-situation and thereby forcing the chosen theory into logical isolation, in which it becomes a sitting target for the attack of test-experimenters. But since this procedure did not offer a suitable guide for a rational reconstruction of the history of science, we may just as well completely rethink our approach.' (Lakatos, 1970, p. 117)
Sophisticated methodological falsificationism
Introduction
Re-thinking the problems raised by naive methodological falsificationism led Popper towards a more sophisticated version. The sophistication of the naive model is primarily Popper's but the articulation of the two levels of methodological falsificationism appears most fully developed in Lakatos' work.
The sophisticated falsificationist is concerned that the growth of science embodies more than the elementary process posited by the naive methodologist. Sophisticated falsificationism takes account of confirmation of theories and the interconnectedness of theories.
The SFs account of science, with its emphasis on the growth of science, switches the focus of attention from the merits of a single theory to the relative merits of competing theories. It gives a dynamic picture of science rather than the static account of the most naive falsificationists. Instead of asking of a theory, 'Is it falsifiable ?', 'How falsifiable is it ?' and 'Has it been falsified ?', it becomes more appropriate to ask, 'Is this newly proposed theory a viable replacement for the one it challenges ?'. In general, a newly proposed theory will be acceptable as worthy of the consideration of scientists if it is more falsifiable than its rivals, and especially if it predicts a new kind of phenomena not touched on by its rivals. (Chalmers, 1978, p. 47)
The sophisticated falsificationist requires then, a theory to be more than just falsifiable if it is to be accorded scientific status. Specifically it needs to lead to novel phenomena if it is to replace an established theory. The theory therefore needs to exhibit excess empirical content, and to have some of the excess content verified. This latter, and possibly indefinite process, embodies corroboration, and as such it is not illogical to 'save' a theory through corroborated adjustments rather than to persist in the pursuit of 'falsification at any cost'. The problem, then, is to lay down demarcation criteria which distinguish between 'scientific' and 'pseudo scientific' D adjustments. The former representing 'progressive' adjustments and the latter 'degeneration'. Non-permissible auxilliary hypotheses are ad hoc hypotheses.
Whatever the change in science, either through the sophisicated 'falsification' of theory or through the 'scientific' adjustment of an established theory in the light of anomalies, what the sophiticated methodological falsificationist is doing is appraising, not a single theory but a series of theories.
For a series of theories to be progressing it is necessary that they need to constitute both an empirical and theoretical problemshift. Growth then becomes a multiple relation between competing theories, the empirical 'base' and the emergence of empirical information as a result of the competition. Sophisticated falsificationism sees science as growing through a proliferation of theories, and is therefore a dynamic model. Unlike the naive falsificationist, for whom science grows through repeated experimental overthrow of theories, the sophisticated falsificationist permits no 'falsification' of theory until a better one is produced, and does not wait until accepted theories are refuted. So long as succeeding theories fulfil the requirements of the sophisticated falsificationist science can progress independent of refuting instances.
'While naive falsificationism stresses 'the urgency of replacing a falsified hypothesis by a better one' (Popper, [1934] 1959, p. 87, footnote 1.), sophisticated falsificationism stresses the urgency of replacing any hypothesis by a better one. Falsification cannot 'compel the theorist to search for a better theory' (Popper, [1934] 1959, section 4), simply because falsification cannot precede the better theory'. (Lakatos, 1970, p. 122).
Consequently as there is no falsification of a theory in the sophiticated version (as there is in the naive version) there can be no crucial or decisive experiments for the sophisticated falsificationist, except in retrospect. Indeed, some of the theories that bring about falsification (in the sophisticated sense) are proposed after the 'counter evidence' has appeared (which is quite paradoxical to the naive falsificationist). The sophisticated version is less constrained to subvert history to its normative prescriptions, there is no requirement that it seek crucial falsifying experiments.
The sophiticated approach to methodological falsificationism is, in some respects, simpler than the naive approach. There is no need to concern oneself, as a scientist, with decisions as to whether a 'falsifying' experiment undermines a theory, its auxiliary hypotheses or the ceteribus paribus clause or some other aspect of the 'unproblematic' background knowledge. Falsification becomes the replacement of one theory by a 'better' one. This bypasses the naive falsificationists preclusion of well corroborated theories from background knowledge because of the naive concern with singular statements.
Provided theories (as background knowledge) either directly or through auxilliary hypothese encompass empirical content progressively, then the sophisticated methodological falsificationist can admit such theories to science.
Thus we do not eliminate a (syntactically) metaphysical theory, if it clashes with a well-corroborated scientific theory, as naive falsification suggests. We eliminate it if it produces a degenerating shift in the long run and there is a better, rival, metaphysic to replace it. (Lakatos, 1970, p. 127)
Simplistic conventionalism and sophisticated methodological falsificationism
Instead of 'falsificationism at any price', the sophisticated methodological falsificationist attempted to provide standards by which a theory may be saved. Such standards pre-dated sophiticated falsificationism.
Duhem's conventionalism posited 'simplicity' as the criterion of change and thus permitted a surrounding set of auxilliary propositions to keep a 'refuted' theory in tact (up to a point). 'Good sense' ultimately prevailing when the theory became 'cumbersome' and a 'simpler' version of reality 'discovered'.
Sophisticated methodological falsificationism demands to know what the criteria of simplicity are; what the nebulous concept of 'good sense' amounts to, and thus, in what sense, for example, is Copernican astronomy simpler than Ptolemaic astronomy ? The sophiticated falsificationist removes these 'subjective' elements and replaces them with a more rigorously objective account.
However, the 'objectivity' of this account depends upon the assessment of the empirical content of succeeding theories. The progressive nature of the theoretical shift is problematic in the light of the theory-laden nature of observation, which falsificationism fails to confront, and, indeed, the immediate intuition of the excess empirical content that the sophisticated falsificationist takes for granted as self-evident, is by no means obvious. The same objection, namely that of the theory base of observation, can be levelled at the intuitive appraisal of the empirical content of theories. In the light of given theoretical knowledge, the empirical content of a subsequent theory may not be at all apparent.
Inadequacies of sophisticated methodological falsificationism
Theory context of observation
Despite the sophisticated falsificationists attempts to account for confirmation of theories, the falsificationist model is still dominated by 'refutation'. The sophisticated version differs fundamentally from the naive version in its demand that refutation is only acceptable when a new theory is empirically better than an old one. The naive falsificationist has no such limitation to deflect attacks on theory, all that is needed is bold conjectures cast in falsifiable terms. The assumption of naive methodological falsificationism is that science progresses from an exposé of what is false.
Nevertheless, in both approaches, theory acceptance is only tentative, it is always of the 'best currently available type'. However, this 'best available theory' principle still relies on the primacy of 'observation' and fails to take account of theory-laden nature of 'observation'.
Popper distinguishes between private and public observation statements thus logically a theory is refuted privately or publicly (only the latter count for science). For the sophisticated falsificationist the public refuting observations need to be incorporated into a 'better' theory. For a public refutation the observation needs to be acceptable to the scientific community, i.e. has become accepted by convention.
But why? What determines such acceptance? Science needs to be convinced that what is 'seen' is real, which requires a theoretical base by which perception is constructed. Such a theory is inextricably related to the theory being posited which the observation is designed to confirm (in the sense of rejecting a previous theory) for the sophisticated falsificationist. Thus, the theoretical base of the observation is self legitimating.
By accepting the 'observation' by convention it is not the observation but the theory that is being accepted. That, subsequently, the same observation might fit into a more advanced alternative theory in no way undermines the fact that the original observation was 'seen', in its theoretical context, at the time and adapted by convention as a 'confirmation' of a theory that was only later to be 'refuted'. The observation has no autonomy; all observation is theory based. Consequently to refer to 'observation' statements as 'basic' and therefore as criteria for scientific advance, is to completely misrepresent the relationship between theory and observation.
Knowledge at the time of Copernicus did not permit a legitimate criticism of the observation that the apparent sizes of Mars and Venus remain roughly constant, so that Copernicus's theory, taken literally, could be deemed falsified by that observation. One hundred years later, the falsification could be revoked because of new developments in optics. Conclusive falsifications are ruled out by the lack of a perfectly secure observational base on which they depend. (Chalmers, 1978, p. 60)
In any test situation the observation or other pertinent conditions may be 'incorrect' as well as, or instead of, the theory. There is no way the falsificationist can decide which in any absolute way.
Historical evidence
There are innumerable instances in the history of science of progress ignoring the ground rules of methodological falsificationism. 'Observational evidence seems to have had little direct impact in causing the overthrow or development of many major theories in science.
Prout's theory of atomic weight, which lies at the heart of modern theories of the structure of atoms, persisted despite numerous experimental results that showed that not all elements had atomic weights that were simple number multiples of the atomic weight of hydrogen.
Newton's theories of gravitation were similarly 'refuted' by observation. The orbit of the moon and, later, the orbit of Mercury were clear 'refuting' instances yet the theory persisted and was one of the most important theories in the history of science. It persisted in spite of these refuting instances, the orbit of the moon was only explained by external factors some fifty years after the 'refutation' was discovered. The perihilion of Mercury was never accounted for by Newton's theories.
Maxwell's kinetic theory of gases was 'falsified' by measurements of specific heats of gases, as was Bohr's theory of the atom by observation that some matter remained stable longer than the theory anticipated. However, like Copernicus, whose system of astronomy faced a barrage of refutations, they persevered, shrugging off the refuting instances.
These examples show that the falsificationist interpretation is, within its own terms of reference, facile. For, while the evidence seems clear, and indeed, on the basis of inductivism and an 'observational base' is uncontestable it is merely counter-evidence offered within the methodological falsificationist model. Methodological falsificationism can thus be shown to be falsified by its own criteria, however, this in no way broaches problems of historicality.
Furthermore, to defend the sophisticated model on the grounds that to suggest, as it logically must, that in the case of these major theories no content-increasing successive theory was proposed (hence the anomalies could not be accounted for by auxilliary hypotheses, or shelved) is to expect the historian of science to tolerate too gross a distortion to accommodate the model. In each case there were documented alternative theories which only in hindsight, in the perspective of the successful theory, might be regarded as non-content increasing.
Thus while methodological falsificationism is 'falsified' by its own accepted standards of historical evidence one can make no historist statements about the nature of this evidence in relation to other theoretical models.
Philosophy of Science.Info
Falsificationism is a rival account of the processes involved in scientific research to inductivism. Inductivism holds that science proceeds from observation to theory, beginning with observations derived from experiments, and extrapolating from these to general laws. Falsificationism suggests that science proceeds in the opposite direction, beginning with scientific theories or “conjectures”, and then conducting experiments and eliminating those theories that are falsified by results.
Falsificationism exploits an important logical point: falsifying instances are more significant than confirming instances. If we have a general law, and conduct an experiment that confirms it, then we still do not know whether the law is true. It remains a live option, but nothing more. If, on the other hand, our experiment contradicts the theory, then we have discovered that the theory is false. Unexpected experimental results are far more significant than expected results.
See also
methodology of scientific research programmes
Researching the Real World Section 2.2.1.6
Chalmers, A.F., 1994, What is This Thing Called Science? Milton Keynes, Open University Press.
Lakatos, I., 1970, 'Falsification and the Methodology of Scientific Research Programmes', in Criticism and the Growth of Knowledge, Lakatos, I. and Musgrave, A. (Eds.), pp. 91–195. Cambridge: Cambridge University Press.
Medawar, P. B., 1967, The Art of the Soluble. London: Methuen.
Philosophy of Science.Info
Popper, K.R.,
Popper, K.R., 1963, Conjectures and Refutations. Routledge, London.
Top