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., 2012-19, Social Research Glossary, Quality Research International,

This is a dynamic glossary and the author would welcome any e-mail suggestions for additions or amendments. Page updated 23 January, 2019 , © Lee Harvey 2012–2019.


A fast-paced novel of conjecture and surprises



core definition

Conventionalism argues that the body of established scientific knowledge is established by convention.

explanatory context


Conventionalism argues that scientific theory is not underpinned by adequate empirical evidence and formal reasoning, i.e., the prevailing accounts are but one of a set of theoretical alternatives. The prevailing account is adopted by convention, not because it is a ‘truer’ account. Objectivity, for conventionalists, derives from general agreement among the scientific community over the conventions adopted.


Conventionalism contrasts with rationalism and empiricism. Rationalism and empiricism attempted a distinction between (justified) scientific belief, based on ‘provable’ propositions (either formally from indubitable premises or via operations on empirical evidence) and the non-scientific (faith, opinion, prejudice). Further, empiricism sees science as a progressively and continuously accumulating body of knowledge. Conventionalism denies the necessity of this view.


Conventionalist accounts of the production of scientific knowledge accept the fallibility of the empirical base and the impossibility of proven knowledge. As observation is not theory-neutral then empirical evidence is not the final arbiter that empiricist approaches would want. Science is constructed, and ethical and/or aesthetic values, instrumental or other ‘extrinsic’ conditions can then enter into scientific theory construction. Science is thus not a simple (internal) logical account.


Conventionalists, thus, allow science to be taken seriously as an historical process subject to transformation and locked into relationships with other social practices. Any account of science from this conventional point of view requires an engagement with concrete episodes.


However, a problem arises for conventionalism if neither rational proof nor empirical evidence can be called in to ‘guarantee’ a correspondence between ‘reality’ and theory. If theory is merely conventional how can cognitive value be ascribed to scientific findings? In short, how can conventionalism be ‘saved’ from radical relativism?


Poincaré was probably the first to propose a conventionalist approach, which was developed by Duhem and Mach in the earlier part of the 20th century. More recently, conventionalism was taken up by Bachelard.



Types of conventionalism


There are two strands to this conventionalist approach, conservative conventionalism, which denies the possibility of transcending the conventional framework and revolutionary conventionalism, which accepted the process of change in conventionalist accounts.


Conservative conventionalism

Conservative conventionalism embodies the central tenet of conventionalist approaches. This view suggests that scientific propositions are useful devices for ordering the world rather than being isomorphic to reality.


Poincaré (like Mach) rejected the idea of a-priori tenets of science but also rejected Mach’s empiricism. Instead he proposed conventionalism. Poincaré agreed that the general laws of science were not empirically testable but argued that they were conventions for ordering phenomena. Such laws as do exist, argued Poincare, are not testable by experiment as they are not based upon fact, instead they are arbitrary conventions about how to use some words or expressions.


Thus, conventionalism sees the progress of science as the result of methodological decisions taken by scientists. The prevailing view was that science was getting nearer to a fuller understanding by constructing propositions that ever more closely reflected the ‘real’ or ‘true’ state of the (rational) world. Poincaré, however, argued that the persistence of Newtonian mechanics for example, was not because of its intrinsic relationship with reality but because of decisions taken by scientists. Poincaré suggested that scientists may decide that a theory, having established some success, shall not be allowed to be refuted. Once this decision has been made, any apparent anomalies are deflected by ‘conventionalist strategies’ such as the construction of auxiliary hypotheses. (This reflects Lakatos’s critique.)


Revolutionary conventionalism

Simplistic revolutionary conventionalism

This model is the exemplar of what is usually referred to as ‘conventionalism’ and can be seen in the work of Duhem. It takes the position of knowledge established by fiat but does not entrench it in incontrovertible methodological decisions.


On the contrary, Duhem suggests that science proceeds through established theories being supplanted by simpler explanations. An established theory becomes amended in the light of anomalies (rather than refuted, as the dogmatic falsificationists would have it) and surrounded by auxiliary hypotheses. This process, after a time, makes the theory cumbersome, and, more than likely inconsistent. Consequently, there is an intuitive need to provide a ‘simpler’ account and once this emerges it is adopted by scientists as a better match to reality. x


Simplistic revolutionary conventionalism thus presupposes that ‘reality’ is uncomplicated, that the physical world is a process of simple relationships. This is by no means established.


Duhem suggested that the decision as to which theory or component of a theoretical system was ‘correct’, was one guided by common sense, sagacity and metaphysical instinct.


Quine (1961) developed Duhem’s thesis and suggested that any theory of whatever complexity can be saved indefinitely from ‘refutation’ by imaginative readjustment of the background knowledge on which it is grounded. This leads to the possibility of the denial of any disproof of any particular components of a theory. Conversely, no statement is therefore immune from revision. In the imaginative reconstruction of a theory held by convention any part can be reconstituted, thus, that which is true by fiat may shift at the whim of the scientific community. This excludes any rational selection among alternatives.


Methodological falsificationism

Methodological falsificationism denies that science is established inductively. Nor does it see science as merely the result of convention. It argues, basically, that scientific propositions can be falsified.


There are a number of varieties of falsificationism. Lakatos refers to naïve methodological falsificationism and sophisticated methodological falsificationism, which he sees as conventionalist approaches and dogmatic falsificationism, which he sees as a justificationist approach.


See SMF1

analytical review

Hughes (undated) wrote:

Conventionalism states that the human brain creates logical structures and then devises experiments to fit into them. A scientist works like an artist, using creative imagination..

associated issues


Gaston Bachelard (1884–1962) was a major French conventionalist philosopher of science. For him, science, as an objective endeavour, constantly struggles against the 'reverie' of our poetic/artistic tendencies, which tend to be grounded at a common-sense level and take experiences at face value.

Thus scientific knowledge is not (as some empiricism would have it) opposite to ignorance but opposite to a tenacious web of error. Further, history of science is not a continuous accumulation of knowledge but a discontinuous revolutionary process, through which earlier conceptions are rejected, displaced, and replaced by new theoretical constructs.


Bachelard argues that epistemological obstacles inhibit science, but a revolution does not see the end of them. For Bachelard, rationalism, positivism, empirical realism are all such surviving obstacles. Such obstacles, while presenting themselves as universal philosophies of the scientific method are pertinent only to a certain phase in the history of each science. Thus a pertinent philosophy must be one that negates all previous ones. Science is characterised, not be fixed ontological presuppositions but by theoretical formulations that are susceptible to continuous transformation. For Bachelard, the dynamic of science comes from the dialogue between concept formation and their 'concretisation' in instrumentation, which produces, in the scientific experiment, the phenomena of scientific observation.


Mastin (2008) described Bachelard as a constructivist on the basis that he claimed (in 1934) that "Nothing proceeds from itself. Nothing is given. All is constructed" despite the fact that it was 1967 before Jean Piaget first used the expression "constructivist epistemology".

related areas

See also





Hughes, J., undated, What is Science? available at, accessed 21 January 2013, still available 16 December 2016.

Mastin, L., 2008, 'Constructivism', available at, accessed 9 March 2013, still available 16 December 2016.

copyright Lee Harvey 2012–2019


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