This blog post examines whether Lakatos’ research program can provide sufficient grounds for explaining scientists’ theoretical choices.
The philosophy of science seeks to answer questions such as: What constitutes science and what does not? Can we determine the methodology scientists follow when science develops, and if so, what is it? What mechanisms can explain the history of scientific development? In contemporary philosophy of science, representative attempts to answer these questions include Popper’s falsificationism, Lakatos’s research program, and Kuhn’s paradigm. These offer diverse perspectives on science; each perspective successfully answers some of the preceding questions, yet simultaneously contains sufficient points of criticism to become subjects of debate and revision. Lakatos’s research program builds upon falsificationism to some extent while attempting to resolve its problems. However, methodological limitations also exist. While examining explanations of Lakatos’s research program and the limitations pointed out by Chalmers in ‘The Philosophy of Modern Science’, I discovered that it does not include content related to the researcher’s trust in theories that have failed many falsification attempts. Therefore, this paper briefly explains the preceding content and, considering this, proposes another modification to Lakatos’s research program.
A core concept in Popper’s falsificationism is that while experience cannot prove a theory to be true, it can prove it to be false. That is, even if a theory adequately explains numerous observational results, there is no guarantee that future observations inconsistent with the theory will not emerge. Conversely, if observed results differ from those predicted by a theory, we can know that theory is false. The basis for scientists accepting a theory is the failure of attempts to falsify it; thus, such acceptance is always provisional. Evidence that something is false provides more information than simply reconfirming what is already known, and this leads to the advancement of science. For this reason, in falsificationism, the advancement of science occurs when a hypothesis predicted to be true is falsified, or when attempts to falsify a hypothesis predicted to be false fail.
However, observations capable of falsifying a theory also depend on the truth of several underlying assumptions (such as the accuracy of the experiment, the precision of the observational instruments, and the validity of the theory itself). Since these assumptions, like the theory itself, can be proven false, both acceptance and falsification of a theory must remain provisional. Furthermore, historically, even when observations occur that cannot be explained by any existing theory, they are often regarded as anomalous cases or observational errors rather than leading to the theory’s falsification. These are historical events difficult to explain from the falsificationist perspective that theories are discarded upon falsification.
To address this problem, Lakatos and Kuhn examined the historical process of scientific development. They both reached the conclusion that theories must be understood as structures. That is, parts of a theory are considered more fundamental, and scientists studying that theory fundamentally regard them as true. Even if observations inconsistent with the theory occur, the cause is thought to lie in factors other than the basic assumptions. The remaining parts are subject to modification as research progresses and observations that fail to match predictions emerge. However, these modifications must be sufficiently clear to allow independent verification. Taking the discovery of Neptune as an example, when Uranus deviated from the orbit predicted by Newtonian mechanics, researchers considered the possibility of another celestial body beyond Uranus influencing its motion, rather than questioning the fundamental principles of Newtonian mechanics.
In Lakatos’s research program, the various hypotheses constituting a theory can be understood as divided into a robust core and a protective belt. Hypotheses fundamentally considered true are termed the ‘robust core,’ while the other, modifiable hypotheses are called the ‘protective belt.’ Scientists regard the robust core as unfalsifiable; when falsifying observations occur, they modify the protective belt to shield the robust core. In this process, researchers follow two guidelines: the ‘negative discovery method’ (not targeting the core for falsification) and the ‘positive discovery method’ (modifying the protective belt to explain observations and predict new phenomena). Kuhn also proposed the concept of ‘paradigms,’ similar to research programs, describing developments within these programs as analogous to progress during periods of ‘normal science’.
The difference between Lakatos and Kuhn lies in the change of the mainstream research program or paradigm. Kuhn explained paradigm shifts by giving them a historical context and viewed the interaction between groups of scientists as an important factor. Lakatos, on the other hand, evaluated the value of research programs based on their success rate in making new predictions. That is, research programs capable of making new predictions and succeeding in them grow into the mainstream, while those that persistently fail to falsify are discarded. This can be thought of as a loose form of falsification. A single falsification does not immediately lead to the theory’s abandonment, but persistent failure to falsify results in the theory’s discard.
The problem is how long a research program must fail to make new predictions before it can be discarded. The annual parallax predicted by Copernicus in the 16th century could only be measured in the 19th century. Since the protective shield of a regressive research program can be modified and transformed into a progressive one, evaluating whether a specific research program is superior to competing programs requires looking back at the past.
In ‘The Philosophy of Modern Science’, Chalmers suggests that Lakatos’s research program has limitations when viewed as a methodology consciously followed by scientific researchers. Explaining historical scientific development as a rule differs from determining the norms actually followed by researchers at the time. While the former explains the process of theoretical change, the latter explains the choices made by researchers that caused such change. Lakatos’s research programs and negative/positive discovery methods were proposed as norms governing discovery, serving as a methodology to explain researchers’ choices. However, Chalmers points out the limitations of this explanation.
Chalmers’ critique is as follows: Lakatos failed to provide criteria for selecting among competing research programs. Since it is impossible for a researcher to immediately grasp the gradualism of a specific program, Lakatos’s theory cannot explain researchers’ choices. Furthermore, the changes scientists brought about before Lakatos’s theory was proposed cannot be seen as the result of following his recently devised methodology.
Chalmers proposed modifying the theory to explain only theoretical change, not researcher choice. While Lakatos did not separate theoretical change from researcher choice, he acknowledged the inability to specify the norms of choice and sought to separate the two, explaining only theoretical change. He objectively expressed the potential for a specific theory to develop through the concept of the degree of productivity. He argued that researchers need not grasp productive potential; it suffices to assume that science evolves over time toward programs with higher productive potential. Since researchers cannot know the norms governing theory selection, studying regressive theories is also rational.
However, I do not believe the Lakatos methodology must abandon explanations of scientists’ choices. The criticisms Chalmers presents are refutable. While it is true that Lakatos failed to specify a norm for choosing between competing theories, if theory selection and theory change are separated, his methodology can still function as a way to explain the common trajectory followed by researchers. This cannot be a guideline guaranteeing the success of individual scientists, but it can be a methodology that incorporates the thinking necessary for selecting successful theories.
As such a guideline, it can offer confidence in hypotheses that withstand more attempts at falsification. Within a specific theory, hypotheses that endure more attempts at falsification, are studied over a longer period, and successfully make predictions are more trustworthy than those that do not. For example, the decision by researchers to follow Newtonian mechanics at the time of Neptune’s discovery need not be explained as merely a methodological choice. This is because the fundamental hypotheses of Newtonian mechanics—the inverse-square law and the laws governing force and acceleration—had enabled numerous predictions to be made without contradiction. These fundamental hypotheses had undergone extensive historical verification, leading researchers to consider the possibility of the protective hypothesis being incorrect more seriously than the possibility of the fundamental hypothesis being incorrect. Consequently, to reject the fundamental hypothesis, an unresolved error must persist over a long period.
While this idea is difficult to apply to choices between all competing research programs, it can influence selection when hypotheses within each program are similar but their core choices differ. Researchers will tend to choose the research program that adopts a hypothesis as its core if that hypothesis underpins more predictions. While old theories may be falsified, this tendency does not pose a problem in explaining scientific progress.
This paper explained Lakatos’s theory, which adopts Popper’s falsificationism more loosely to describe historical events, and discussed Chalmers’s critique and modification of it. It was also suggested that accepting Chalmers’ modified research program can still explain the common thinking followed by researchers. This critiques Chalmers’ position, which, like Popper, acknowledges that scientists cannot precisely choose the most successful theory of their time, but also fails to propose a methodology that researchers consciously followed when scientific progress occurred. While the impossibility of successful theory selection may lead to the impossibility of presenting a methodology that guides individual scientists to success, it is possible to present the methodology scientists actually followed. As one guideline of this methodology, it proposed placing trust in hypotheses that withstand more falsification attempts. This discussion is productive as an attempt by philosophy of science to explain more, and it will be possible to discover further such guidelines to enhance our understanding of scientists’ choices.
