The pessimistic meta-induction serves as a critical challenge within the philosophy of science, questioning the reliability of our best scientific theories. Often associated with the underdetermination of theory by evidence, this argument suggests that because past successful theories have been subsequently discarded or radically altered, we have no reason to believe that our current theories are approximately true. This line of reasoning casts a long shadow over scientific progress, implying that the epistemic justification for our most confident beliefs is fundamentally fragile.
Historical Roots and Core Structure
The argument finds its most famous articulation in the work of Larry Laudan, although its roots extend into the broader pessimistic induction literature concerning the history of science. The structure is deductive and hinges on a disjunction regarding the truth of a given theory T. Either T is approximately true, or T is not approximately true. If T is approximately true, then it must be empirically adequate, meaning it correctly describes the observable phenomena it was designed to explain. However, the historical record shows that empirically adequate theories are frequently replaced by ones that are empirically adequate but incompatible with the prior theory. This replacement often involves a reduction of the ontological commitments of the old theory or a shift in the conceptual framework. Therefore, the fact that a theory is empirically adequate now provides no assurance that it will be regarded as approximately true in the future, leading to the conclusion that we should not assume its approximate truth.
Distinguishing Variants: Underdetermination vs. Pessimistic Induction
It is essential to distinguish the pessimistic meta-induction from the general problem of underdetermination, despite their close relationship. Underdetermination argues that evidence alone does not necessitate a single rational conclusion, leaving multiple competing theories viable. The pessimistic induction is more specific and historical; it does not merely claim that we lack evidence for the truth of our theories, but that evidence actually suggests we are wrong to trust them. While underdetermination highlights a logical gap between evidence and theory choice, the pessimistic induction fills this gap with a historical premise: past theories that were confidently believed to be true or approximately true have been radically false or discarded. This historical premise is the linchpin, transforming a problem of logical possibility into an inductive argument about the trajectory of scientific change.
Responses and the Realism Debate
Philosophers of science have not been passive in the face of this challenge, developing a range of responses to defend scientific realism. One prominent strategy is to reject the historical premise by arguing that the theories we regard as failed or replaced were not genuinely successful approximations to the truth but were instead instrumental tools that worked within limited domains. On this view, Newtonian physics was not approximately true about the structure of space and time, but it was an extremely useful approximation for engineering and everyday scales. A second response involves shifting the target of realism away from truth toward empirical adequacy. Defenders of this position, like Bas van Fraassen, argue that we should only aim for theories that are empirically adequate, avoiding the metaphysical commitments associated with truth about unobservable entities. This dissolves the force of the induction because the argument presupposes a realist goal of truth that is neither necessary nor desirable.
Selective Pessimism and the Problem of Bias
A crucial counter-argument to the pessimistic meta-induction concerns its selection of historical examples. Critics point out that the history of science is littered with theories that were successful and retained, or at least significantly preserved, within new frameworks. The theory of conservation of energy, for instance, has persisted through immense theoretical upheaval, suggesting that not all past successes are discarded. Furthermore, the argument is accused of committing a selection bias by focusing on dramatic revolutionary cases while ignoring the cumulative and incremental progress of science. The mathematical tools of ancient astronomy, or the classificatory systems of early natural history, provided durable frameworks that were built upon rather than wholly erased. This suggests that there is a "residual" truth or empirical adequacy that carries over, challenging the stark dichotomy between complete success and total failure implied by the induction.
More perspective on Pessimistic meta-induction can make the topic easier to follow by connecting earlier points with a few simple takeaways.
