Science and Pseudoscience

Science and pseudoscience have always had an ambiguous relationship. We usually know how to recognize, at a glance, what is scientific and what is not. In today's world, immersed in technological progress, this distinction seems almost obvious: science produces medicines, machines, computers and etc.; pseudoscience, on the other hand, seems to be a set of confused and uncertain beliefs. Yet, when we try to transform this intuition into a clear definition, the boundary becomes blurred: what is science? And on what basis can we say that one practice is scientific and another is not?

This is where the problem of demarcation arises: in order to draw a non-arbitrary line between science and non-science, we need to clarify at least two fundamental things:

• how science produces knowledge (the method);
• how and why this knowledge improves over time (progress).

Without a theory of method and progress, we risk defining science in a circular way ("science is what scientists do") or purely sociologically ("science is what the state calls science"), and then it becomes difficult to justify why some practices should be excluded and others accepted.

In this sense, demarcation is not only a descriptive issue ("what scientists do"), but also a normative one: it implies criteria on how they should work to produce reliable knowledge. It is, therefore, inevitable to reflect on the scientific method: not as a rigid¹, but as a set of constraints aimed at reducing error and making our knowledge of the world possible and public.

Scientific Method and the Concept of Progress

There is a saying that it is often more important to ask the right questions than to have the right answers. In fact, science almost always born from a question: about how nature works, about the causes of a phenomenon and about the patterns we observe. Since ancient times, humans have sought answers to questions such as why the sun rises and sets, why the stars move, why the sky is blue and etc.. For long periods of history, explanations relied on myths, legends, or supernatural agents: not necessarily irrational in the cultural context of the time, but often difficult to control and test.

Science, on the other hand, is characterized by a systematic approach: it is not enough to tell a plausible story; explanations must be constructed that can be publicly evaluated and, at least in principle, disproved by facts.² Schematically, a typical cycle includes:

1. Formulation of the problem: they are guided by prior knowledge, conceptual tools (e.g. math), and often by anomalies (something that does not add up).
2. Hypotheses: a clear and binding hypothesis is proposed; it must say not only "what could be true", but also what we expect to observe if it were true.
3. Empirical tests (observations, experiments, measurements): data is collected using transparent, controlled, and replicable procedures (or at least replicable under comparable conditions).
4. Critical evaluation and review: if the data contradict the hypothesis, the hypothesis must be modified or abandoned; if they support it, it does not become "true once and for all", but gains credibility and is subjected to more rigorous testing.
5. Sharing and external control:³ paper publication, peer review, discussion with the community, attempts at replication: science is also a social process, but with rules aimed at reducing errors, bias, and self-deception.

Without a clear idea of method, any practice could present itself as scientific, making the demarcation inevitably arbitrary. Method serves not only to describe what science does, but also to establish constraints: what kind of claims are admissible, how they must be justified, and how they can be challenged.

Similarly, scientific progress does not consist of a simple linear accumulation of facts or notions, but of a continuous improvement of knowledge over time: greater empirical accuracy, more predictive power, better control of errors, and the ability to integrate previously separate phenomena. Theories are not preserved because they are "true once and for all", but because they are, at present, the best available to explain and predict an phenomena.

A central point is that science never claims absolute certainty. At most, a theory is considered highly reliable within a certain domain and with a certain degree of confidence. This fallibility attitude is a strength: admitting the possibility of error does not mean "knowing nothing", but knowing responsibly, indicating limits, conditions, and margins of uncertainty. Pseudosciences, on the contrary, often tend to present dogmatic or refutable theses, avoiding what could disprove them.

Inductivism: Generalization of Facts

Inductivism is one of the responses for the demarcation problem, by which the science forms the general laws basing on the specific cases. A research process starts with observation of phenomena in the reality, by repetition of observations and discovery of regularities, more generalization is developed. For example, we see that, day after day, the Sun rises and sets, of a long succession of such cases, we conclude there is a kind of law of the kind: "the Sun rises and sets every day".

But, here the weakness of induction is obvious: the fact that it is always occurred is not logically bound with the fact of its occurring forever. A vast number of observations do not draw to a generalized conclusion: there is always the possibility, however small, that an exception will be made in some future, or that some other case will be studied to invalidate the generalization. This is not to say that induction is useless: it simply cannot give absolute logical certainty, but only probabilistic and pragmatic support.

This is also significant to demarcation. Whether the science is based on observation and experiments, inductivism appears to provide an easy criterion: what is verified with numerous facts is scientific. This is however not enough as a demarcation criterion. A pseudoscience may pose itself as empirically grounded through the cherry-picking of only positive examples and ignoring the negative ones. This is where cherry-picking comes into the game: it is not hard to find confirmations when one is free to arbitrarily select what observations to pay attention to. This is why it is not the quantity of evidence one (it is still important), but the manner in which the evidence chosen and tested: protocols, controls, replicability, error analysis, comparison with possibly undesirably data.

Another flaw is that observations are not absolutely neutral pure data. They are also restricted to reality and they are always aggregated and interpreted in a theoretical framework: the choice of what to measure, what instruments to apply, what categories to apply, how to deal with noise, is subject to assumptions and models.

In a few words, induction does reveal something true, the empirical concretion of science, but it cannot, of itself, rationally prove universal laws, or decisively distinguish science and pseudoscience.

Popper: Falsifiability

Another possible solution is the principle of falsifiability by Karl Popper. This is because Popper asserts that a theory is only scientific, when it puts itself to the danger of being disproved by observations or experiments: it has to make statements that are clear, to be disproven in the light of potential contrary evidence. In this view, we can go back to the concept, that no theory can be conclusively verified. Scientific advancement is not piling up evidences to the verified side, but subjecting our knowledge to constant and intense stress, to find out how far it will be held and whether there are superior models.

Science is falsifiable, while pseudoscience systematically takes positions against the danger of being disproved. Pseudosciences also have a tendency to come up with vague theories which are difficult to refute, by the expedient of retroactive explanations, semantic ambiguities, or ad hoc hypotheses that have no testable consequences. To the contrary, a good theory explicitly contains what is supposed to happen in the event that it is true and what is supposed to refute it in the event that it is not.

The merits of the Popperian criterion are evident: it is concerned with the importance of empirical testing and why science can only develop through criticism and not through confirmation. Nonetheless, it has significant limitations as well. To begin with, the thesis of Duhem-Quine demonstrates that falsifications are not unambiguous; an empirical test does not impact one disconnected hypothesis, but a system (auxiliary hypotheses, initial conditions, measuring instruments, theoretical models). When an experiment disproves a prediction, it is not logically decided what aspect of the theoretical system to give up.

Second, one can hardly make a distinction between legitimate and ad hoc adjustments made to a theory with the aim to either enhance the predictive power of that theory, or, to avoid its disproving due to the lack of the new empirical material.

In summary, the criterion of falsifiability clarifies the role of empirical risk and criticism in the scientific method and offers a more robust demarcation than inductivism. However, in order to fully explain the actual work of science and its historical dynamics, it must be integrated with considerations of the theoretical context, underlying assumptions, and actual practices of scientific communities.⁴

Kuhn: Science as a Communal Activity

Thomas Kuhn proposes a change of perspective: science is a communal and historical activity. Scientific knowledge is not the product of isolated genius, but the result of collective work involving communities of researchers spread across time and space. In this sense, science is a global job of humanity, rather than of the individual.

Kuhn observes that scientific development alternates between two phases:

• Normal science: stable periods in which a community works within a shared paradigm. The paradigm provides ideal models, conceptual tools, methods, evaluation standards, and criteria for acceptability. In this phase, scientists do not question the foundations of the paradigm, but solve problems within it.
• Scientific revolutions: phases of crisis in which anomalies accumulate, the dominant paradigm loses its explanatory power, and a new paradigm may emerge that radically reorganizes the field.

In this perspective, the demarcation is not given by a single logical criterion (induction or falsifiability), but by the fact that a practice operates within a paradigm recognized by a scientific community. Today, for example, in order to publish an article, it is necessary to comply with shared rules and standards: methodological clarity, comparison with the literature, verifiable and replicable data, transparent use of tools, formal language, and explicit limits. These rules are self-imposed by the community to make knowledge communicable and open to criticism. Without shared rules, rational discussion and progress would become impossible.

Pseudosciences, on the contrary, often lack a stable paradigm: they do not generate problems that can be solved systematically, they do not accumulate comparable results, and they do not have common criteria for evaluating successes and failures.

Kuhn's merit is to describe scientific practice and the dynamics of revolutions realistically, highlighting the importance of the community. However, the main limitation concerns the comparison between paradigms: if each paradigm defines its own standards, how can we say that one is "better" than another without falling into an epistemological relativism? How can we say that relativity is "better" than Newtonian mechanics, or that quantum mechanics is superior to classical physics, without falling into a form of relativism? Kuhn's emphasis on the historical-social dimension aspects, making it more difficult to establish a clear normative demarcation.

Lakatos: Research Programmes

Imre Lakatos attempts to mediate between Popper's falsificationism and Kuhn's historicism. On the one hand, he rejects Popper's idea of immediate and decisive falsifications; on the other, he distances himself from the relativistic risk inherent in Kuhn's notion of paradigm. His proposal is that of scientific research programs, interpreted as historical theoretical structures that develop over time and guide the activity of a scientific community.

A research program includes:

• a hard core, consisting of fundamental assumptions that are not abandoned by its supporters;
• a protective belt, consisting of auxiliary hypotheses, models, and initial conditions, which can be modified to respond to empirical difficulties.

When anomalies emerge, they generally affect the protective belt and not the hard core: this explains why theories are not immediately abandoned and provides an clear answer to the problem raised by the Duhem--Quine thesis: the failure of a prediction does not automatically imply the abandonment of the entire theoretical framework, since it is often attributes the error to correctable auxiliary hypotheses, rather than to the core of the program.

A program is scientifically valid if it produces new empirical predictions, and at least some of these predictions are actually confirmed. Instead, it becomes degenerating when it merely accommodates already known facts retroactively, introducing ad hoc adjustments to the protective belt without increasing the empirical content or predictive power.

Thanks to this, pseudoscience typically takes the form of degenerating research programmes: they do not produce new and risky predictions, but react to refutations only with retroactive adjustments, often without independent empirical consequences. Unlike progressive scientific programs, pseudosciences do not show growth in empirical content over time, nor do they show improvement in explanatory or predictive capabilities.

However, Lakatos' approach also has its limitations. In many cases, the assessment of a program as progressive or degenerating is only clear in hindsight, when the historical outcome is already known. In real time, it is often difficult to determine whether a program is going through a temporary phase of difficulty or whether it is actually degenerating. Furthermore, the line between science and pseudoscience remains unclear: some pseudoscientific programs can survive for a long time by promising future confirmations without providing them, making immediate and definitive judgment problematic.

Feyerabend: Criticism of the Universal Method

Paul Feyerabend criticizes the idea of a universal scientific methodology. There is no methodological criterion that is valid in every era and in every context and, consequently, there can be no universal criterion about the problem of demarcation. The history of science shows that many great innovations have violated methodological rules considered rational and correct in their time. It is not an invitation to total chaos or arbitrariness, but a critique to the idea that there is a fixed methodological recipe capable of guiding and justifying all scientific progress.

From this perspective, scientific progress is not the result of the mechanical application of immutable rational rules, but also depends on historical, social, and cultural contingencies. Decisions about which theories to accept or reject are not always determined exclusively by logical or empirical criteria, but also involve pragmatic, psychological, and institutional factors. The main merit of Feyerabend's position is that it destroys the myth of a rigid science and highlights the role of theoretical plurality, creativity, and even the conscious violation of rules as possible drivers of scientific innovation.

However, it is precisely this radical pluralism that raises a serious problem in terms of demarcation. If there are no shared methodological criteria, it becomes difficult to explain why some practices should be considered scientifically reliable, and others not. In particular, Feyerabend's position risks weakening the distinction between science and pseudoscience: if there is no method and if even breaking the rules can be valid, then a pseudoscientific practice could claim legitimacy as a simple alternative to official science.

In the absence of shared evaluation criteria, the danger is that pluralism will turn into epistemological relativism, making it impossible to distinguish between legitimate scientific criticism and simple rejection of evidence.

In summary, Feyerabend offers a powerful critique of methodological dogmatism and highlights the historical complexity of real science. However, the price of this position is a significant weakening of the normative demarcation between science and pseudoscience.

Case Study: COVID-19 as a Demarcation Case

The COVID-19 pandemic has made visible to everyone, how science operates in conditions of urgency and uncertainty, bringing aspects of the scientific method that are not generally familiar to the non-specialist public.

Throughout the introduction of mandatory vaccination, the "no-vax" movement has gained considerable visibility. One of the most common criticism concerned the possible side effects. Like all drugs, vaccines could cause side effects; however, the vast majority of effects are light and temporary, while serious reactions are extremely rare. The philosophically relevant point, however, is more general: in modern medicine, decisions are never made in terms of zero risk. The "zero risk" option does not exist; every healthcare decision is a choice between different risks, weighed against the expected benefits.

The issue of long-term effects is a particularly clear example of scientific reasoning in conditions of uncertainty. At the start of the vaccination campaign, long-term data were not available; however, this was not equivalent to "gambling", but rather making a decision based on the best available evidence. In this context, science shows its fallibilist character: it does not promise absolute certainties, but quantifies uncertainty and manages it in a rational and transparent manner.

Finally, the debate on vaccines shows why no single criterion of demarcation is sufficient on its own. Induction is vulnerable to arbitrary selection of evidence; falsifiability is not always immediate due to auxiliary assumptions; paradigms evolve historically; research programs are evaluated over time; and Feyerabend reminds us that actual scientific practice is more complex than any rigid methodological scheme. However, taken together, these tools allow us to understand why the scientific response to the pandemic, while imperfect, is epistemically justified: it was based on controlled data, public reviews, critical comparison of alternatives, and transparent management of uncertainty.

Conclusion

Pseudosciences have historically contributed to the formation of epistemic bubbles within societies, separated from the rest of the public debate. These bubbles tend to develop in opposition to shared scientific knowledge, a bit like salmon swimming upstream: a current that, in most cases, represents not an arbitrary imposition, but a collective attempt to protect the common good.

The consequences of such attitudes are not limited to the individual sphere. Vaccine refusal, for example, not only poses a personal risk, but also undermines collective protection, encouraging the possible reappearance of diseases that we considered defeated. Similarly, climate change denial is not a neutral position: it has potentially harmful effects on a global scale, influencing political and economic decisions that affect the entire planet. The examples could be multiple, but the central point is that pseudoscience produces negative externalities, affecting people who do not share those beliefs.

However, it would be simplistic and presumptuous to dismiss these phenomena by attributing them solely to the selfishness or bad faith of the individuals involved. Before judging, it is necessary to understand the conditions that make these dynamics possible in order to address them effectively.

In the end, it is essential to be able to reason in a structured way, not intuitively or randomly, but following the methodological principles that characterize scientific thinking: attention to evidence, awareness of limitations, openness to revision and etc. Only by strengthening these skills can we reduce the influence of pseudoscience and encourage more responsible and rational participation in social and political life.