Hossein Jorjani
First publication: 2026 – 07 – 16
Every influential theory undergoes reinterpretation and revision. Later generations refine its concepts, introduce new terminology, extend its applications, solve new problems, modify empirical claims, and sometimes abandon assumptions accepted by its original proponents. The question therefore arises: At what point does a revised theory become a different theory?
The problem cannot be resolved simply by examining whether a theory retains its original name, terminology, or individual propositions. Nor can it be resolved by asking whether its later proponents acknowledge their intellectual debt to its founder.
The identity of a theory must be distinguished from the historical continuity of an intellectual tradition. A theory may undergo substantial reconstruction while remaining within the same tradition. Conversely, theories belonging to different traditions may employ similar concepts or explanatory principles. The question concerns what makes a theory identifiable as a particular explanatory system, and what must remain sufficiently continuous for that identity to survive revision.
A theory should not be understood merely as a collection of propositions or opinions. Nor should it be identified by the recurrence of characteristic terminology. Words such as class, evolution, dialectic, democracy, and liberty may persist while their meanings change substantially. Conversely, different terminology may be used to express closely related concepts. Even the persistence of a concept does not necessarily establish the continuity of its explanatory function.
Consider inbreeding. The term may designate a process, namely mating between individuals related by ancestry. It may also be used to describe a population pattern, such as an excess of homozygosity relative to a specified reference. The process and the pattern are related, but they are not identical. Moreover, the observation of a pattern does not, by itself, establish the process that produced it.
A concept may therefore retain its name while changing its referent, its explanatory role, or the relationship between what is observed and what is proposed as its cause. Similarly, individual propositions may survive after the explanatory framework that originally generated or supported them has disappeared. The same conclusion may be reached through different theoretical routes. The identity of a theory must consequently be sought beyond its vocabulary and isolated conclusions.
A theory possesses an explanatory architecture: an organization of concepts, principles, assumptions, and relationships through which explanations are generated and constrained. Some principles occupy more fundamental positions than others. Some concepts depend upon others for their definition or explanatory function. Certain assumptions establish the conditions under which explanations are considered applicable. The identity of a theory therefore depends not merely on the presence of particular components, but on their positions and relationships within its explanatory structure.
Every sufficiently developed explanatory theory possesses one or more organizing principles. These principles are not simply additional propositions alongside the others. They contribute to determining how the theory operates, which explanations it permits, and how its various components are related. A theory may have one principal explanatory foundation or several relatively autonomous principal contributions. In either case, the relationships among these components are important to its identity.
Consequently, the survival of most of a theory’s propositions does not necessarily establish the survival of its architecture. Nor does the replacement of a proposition necessarily imply that the architecture has changed. The decisive question is what role the modified component performed within the explanatory system.
A theory is constructed to explain particular phenomena or classes of phenomena. Its explanatory domain is therefore relevant to its identity. Two theories may address related subject matters without having the same explanandum. A development in one need not constitute a revision of the other. For example, Darwin’s theory of natural selection was primarily concerned with evolutionary change in organisms, particularly the development of adaptations, forms, and functions. Kimura’s neutral theory was developed principally to explain patterns and rates of molecular evolution. These domains overlap, but they are not identical. Evidence that neutral processes account for much molecular variation does not, by itself, invalidate natural selection as an explanation of adaptation. Before concluding that one theory has displaced another, we must therefore establish whether they address the same explanatory problem.
A further distinction concerns the mechanisms upon which a theory depends. Darwinian natural selection requires heritable variation. However, its basic explanatory mechanism does not require a particular molecular account of inheritance. The mechanisms of inheritance can, within limits, be treated as a black box. This does not mean that every conceivable theory of inheritance is compatible with natural selection. The required variation must be transmitted with sufficient fidelity for cumulative selection to operate. Nevertheless, different accounts of inheritance may satisfy that requirement. A revision of the theory explaining inheritance does not therefore necessarily constitute a revision of the theory explaining adaptation through natural selection.
We must distinguish dependence on a process from dependence on a particular theory of that process.
This distinction helps identify a theory’s indispensable commitments. A theory may require that certain conditions or processes exist without requiring a particular explanation of their underlying mechanisms. Theoretical identity depends partly on which commitments are indispensable to the theory’s explanatory architecture and which can be modified or replaced without changing its characteristic mode of explanation.
Theories naturally develop. They may incorporate new evidence, refine concepts, extend applications, modify subsidiary assumptions, and correct empirical claims without necessarily losing their identity. However, revisions differ in their consequences for the explanatory architecture.
In The Structure of Evolutionary Theory (2002), Stephen Jay Gould introduced a useful distinction among three kinds of changes to a theory, described metaphorically as cuts to the branches of a theoretical structure:
- Subsidiary cuts (S-cuts): Changes affecting secondary components without substantially altering the principal explanatory architecture.
- Revisionary cuts (R-cuts): Changes that substantially reconstruct principal components while retaining sufficient continuity for the theory to remain recognizable.
- Killing cuts (K-cuts): Changes that eliminate commitments indispensable to the theory’s characteristic explanatory structure.
Gould employed these distinctions in examining what he regarded as the principal supporting components of Darwinian theory.
The usefulness of his classification, however, does not depend on accepting his identification of those components or his assessment of the changes they underwent. The distinction between subsidiary, revisionary, and killing changes can be retained as a methodological instrument while its application to particular theories remains open to examination. Such classification cannot be determined simply by counting the number of propositions or components that have changed.
A theory may retain many propositions while losing an indispensable organizing principle. Conversely, extensive modifications of subsidiary mechanisms may leave its principal explanatory architecture intact. Nor can the classification be established independently of the theory’s explanatory domain. A change that is fundamental to one theory may be subsidiary or irrelevant to another. The assessment must therefore consider the explanatory function of the modified component, its relationships with other components, and the consequences of its modification for the explanations the theory is intended to generate.
The identity of a theory must also be distinguished from the continuity of an intellectual tradition. A tradition may preserve characteristic problems, concepts, methods, historical references, and institutional affiliations while accommodating theories with different explanatory architectures. This distinction is particularly important for schools of thought named after individuals, as discussed in Excursus 130, On Eponymous Schools of Thought. An eponymous designation may express intellectual inheritance, identification with a founder, or participation in a historical tradition. It does not necessarily establish that all theories associated with that designation possess the same explanatory architecture.
Similarly, the existence of a historical relationship between two theories does not establish their theoretical identity. The history of evolutionary biology provides an illustration. Darwin’s original theory and the later formulations associated with Neo-Darwinism and the Modern Synthesis are historically related, although their explanatory structures are not necessarily identical. Their relationship cannot be adequately described either by treating them as entirely independent theories or by assuming that their shared historical designation guarantees architectural identity.
Three questions must therefore be distinguished:
- What makes a particular theory identifiable?
- What degree and kind of continuity allow it to remain recognizable through revision?
- What makes different theories members of the same intellectual tradition?
The answers may overlap, but they need not coincide.
Historical continuity deserves recognition, but conceptual clarity also requires attention to differences in explanatory architecture. When a theory has undergone substantial reconstruction, a new designation may help distinguish the revised theory from its predecessor. However, naming conventions are not reliable criteria of theoretical identity. A theory may retain its historical name despite major revisions. Another may acquire a new name while preserving much of its earlier explanatory structure. The adoption or retention of a name is therefore a matter to be distinguished from the assessment of theoretical continuity.
Expressions such as neo-, post-, or revised may indicate different relationships to an earlier theory or tradition, but their use does not establish the nature or extent of architectural change. Theoretical identity must be examined before the appropriateness of a designation can be assessed. It cannot be inferred from the designation itself.
The identity of a scientific or philosophical theory should be sought primarily in its explanatory architecture: the principles that organize its explanations, the relationships among those principles, the domain within which they operate, and the commitments indispensable to their explanatory function. The persistence of terminology, concepts, individual propositions, or historical names does not, by itself, establish theoretical identity. Nor does every change in concepts, assumptions, or explanatory mechanisms necessarily constitute a change of theory.
To assess whether a theory has retained its identity, we must determine what has changed, what explanatory role the changed component performed, and whether the organizing relationships that characterize the theory have survived. The distinction between subsidiary, revisionary, and killing changes is useful precisely because it directs attention away from the mere extent of revision and toward its architectural consequences.
The boundaries between these categories may not always be sharp. Nevertheless, their application requires explicit identification of the theory under examination, its explanatory domain, and its indispensable commitments.
A theory is not defined simply by what it says, but by how its explanations are organized and generated.
Concepts may survive while architectures disappear. Vocabulary is inherited more easily than explanatory systems.
Recognizing these distinctions helps preserve both historical continuity and conceptual precision, without confusing the development of a theory, the reconstruction of its architecture, and the continuation of the intellectual tradition to which it belongs.
Reference
Gould, Stephen Jay. 2002. The Structure of Evolutionary Theory. Cambridge, MA: Belknap Press of Harvard University Press.
