The Serendipity Society - Two faces of serendipity

Two Faces of Serendipity: From Prospective to Retrospective Explanation

How chance discoveries both spark and explain scientific breakthroughs

March, 2026

This post is based on

Arfini, S., & Costa, M. (2025). What kind of explanations can serendipity provide? European Journal of Philosophy of Science, 15:42. https://doi.org/10.1007/s13194-025-00670-4

Serendipity explains: a return to the future… and back

Chances are you have heard at least one of these stories: Alexander Fleming discovered penicillin because a petri dish got contaminated during his holiday; Jocelyn Bell stumbled upon pulsars while hunting for quasars; Penzias and Wilson found cosmic background radiation when all they wanted was to eliminate annoying interference from their equipment.

These tales of scientific serendipity are often told as inspirational narratives about luck favouring the “prepared mind”. But there is something deeper happening here, something that challenges how we think about scientific discovery and, we argue in this paper, also and more importantly, scientific explanation. Explanation that addresses both the future – what scientists will learn during exploratory research – and the past – how epistemologists will later reconstruct and describe what happened. Indeed, we maintain that serendipity can serve as a theoretical DeLorean time machine, allowing metaphorical labcoat-wearing Marty McFly to dynamically look forward and backward into the past and the future of their scientific research.

Thus, to account for its time‑bending abilities, we should first stop treating serendipity the way everyone treated Doc Brown: as an eccentric footnote instead of the real engine driving the story.

Beyond “happy accidents”: serendipity as an explanatory process

Although recent scholarship has begun to take serendipity seriously as a subject of philosophical inquiry, indeed, this concept has historically been treated as science’s quirky occurrence – interesting to mention in Nobel Prize acceptance speeches, but only recently subject to the kind of rigorous systematic analysis we propose here. Yet when we examined these famous discoveries more carefully, we realized something crucial: serendipity is not just about what scientists discover by accident; it is also about how they come to understand what they have found, and why we can, nevertheless, explain these discoveries.

Our research asks two deceptively simple questions: What kind of explanations does serendipity provide? And explanations of what, exactly? The answer is that serendipity offers two fundamentally different types of explanation, operating at different moments in time and serving different purposes. We call them ‘Prospective Explanation’ or, the look into the future and ‘Retrospective Explanation’ or, the reshaping of the past.

The dynamic path forward: prosepective explanation

Imagine you are Jocelyn Bell in 1967, staring at a chart paper stretching over 100 meters. Among the expected signals and countless errors, there is something odd: just a few millimeters of strange, regular pulses. You do not yet know you are looking at a pulsar. You do not even know what a pulsar is. What you have is an anomaly that resists all your attempts to explain it as interference, equipment malfunction, or human error.

Jocelyn Bell (Burnell) pictured at the Mullard Radio Observatory in 1967 with the pulsar chart, left. Image courtesy of the Cavendish Laboratory. (University of Cambridge)

This is where Prospective Explanation begins. It is the messy, uncertain, dynamic process of trying to make sense of something unexpected while the discovery is still unfolding. Unlike standard scientific investigation, where you have a hypothesis and test it, serendipitous discovery forces you to construct explanations in real time, constantly revising them as new evidence emerges.

What makes this fascinating is that scientists in this position are not looking for complete understanding. Instead, they are searching for what we call a “sense of understanding”, a provisional coherence that tells them they are onto something worth pursuing further.

The Prospective Explanation is thus characterized by its lack of a predetermined method. This is not a flaw, it is a distinctive feature. Serendipitous discoveries cannot follow established protocols because, by definition, they emerge from the unexpected. The very absence of a guiding method allows scientists to notice anomalies that do not fit existing frameworks and to explore possibilities that were not on anyone’s research agenda. The first trip of our metaphorical DeLorean, then, is a few preliminary and tentative steps into a future that is still taking shape as the scientists’ hypotheses unfold.

Looking back to understand forward: retrospective explanation

Now fast-forward. The discoveries have been made. Pulsars have become an established celestial phenomenon. Penicillin is saving lives. The cosmic microwave background radiation is textbook material. Within this perspective, we can construct what we call the Retrospective Explanation.

This is the explanation that retrospective analyses of scientific discovery naturally gravitate toward, and for good reason. It meets all the criteria of proper scientific explanation: it achieves explanatory coherence, fits phenomena into a comprehensive scientific worldview, and provides genuine understanding (not just a “sense” of understanding) about both the discovered phenomenon and the discovery process itself.

But what makes serendipity philosophically intriguing: the Retrospective Explanation does not just explain what was discovered. It also explains why the discovery happened the way it did, including all those contingent factors that seem almost embarrassingly unscientific.

Why did other scientists not discover penicillin, even though they surely encountered contaminated plates? Why was Bell’s “imposter syndrome” (her own term) actually crucial to her discovery, making her scrutinize data more carefully than her more confident male colleagues might have? Why did Penzias and Wilson’s obsessive elimination of every possible source of interference turn an equipment problem into a cosmological breakthrough?

The Retrospective Explanation can address these questions because it recognizes that serendipitous discoveries emerge from complex networks of people, tools, institutional contexts, and even emotions. Fleming needed the Oxford research community to fully develop penicillin’s therapeutic applications. Bell needed her hands-on experience building the radio telescope, not just theoretical knowledge of how it worked. Penzias and Wilson needed Robert Dicke’s conceptual framework about Big Bang radiation to make sense of their “noise”. In this sense, the retrospective explanation necessarily encompasses a holistic and situated perspective: by looking into the past, we then actually reshape the contours of what was necessary for the discovery to happen. The DeLorean is then turning back, and while it is revisiting the past, it is also changing our present understanding of how a discovery process works.

Bridging discovery and justification

In fact, what then emerges from distinguishing these two types of explanation is philosophically significant: serendipity might help us rethink the classic distinction between the context of discovery and the context of justification. Traditionally, these are kept separate: discovery is the creative, messy process of generating new ideas; justification is the rigorous, rational process of proving those ideas valid. Philosophers have long struggled to find cases where these contexts genuinely overlap.

Serendipity provides exactly such cases. The Prospective and Retrospective Explanations are not just different perspectives on the same event: they are two interdependent parts of a single process. The dynamic, uncertain reasoning that characterizes Prospective Explanation is the discovery phase. The coherent, rigorous analysis that characterizes Retrospective Explanation is the justification. And neither can exist without the other, even from a temporal perspective. The prospective explanation blurs into the retrospective one because, without the exploratory, hypothesis-testing work of Prospective Explanation, there is no discovery to justify. Similarly, without the coherent, comprehensive framework of Retrospective Explanation, the initial serendipitous event remains just an unexplained accident, not a scientific breakthrough.

Why this matters

Recognizing serendipity’s dual explanatory nature has practical implications. If we want to foster more serendipitous discoveries – and given their outsized impact on scientific inquiry, we should – we need to create environments that support both types of explanation.

This means not just celebrating “prepared minds” (though expertise certainly matters), but also recognizing the social, institutional, historical, and even emotional or contingent factors that enable scientists to notice anomalies, pursue them despite uncertainty, and ultimately construct coherent explanations. It means supporting research communities and networks that can help scientists make sense of unexpected findings.

Most importantly, it means acknowledging that transformative scientific discoveries often emerge not from following the method and time linearity, but, as Doc Brown’s best invention, from knowing when to revise, suspend, or creatively transcend its constraints. Serendipity, far from being science’s embarrassing secret, might actually reveal something profound about how scientific discovery and understanding genuinely works: as an emergent, situated, irreducibly complex process where chance and wisdom interact.

Author bio

Matteo Costa is a Ph.D. Candidate at the Department of Philosophy, Sapienza University of Rome, Italy. His research interests include philosophy and history of science, the epistemology of discovery processes, and the role of chance in scientific inquiry.

Selene Arfini is a Researcher and Professor of Philosophy of Science at the Department of Humanities of the University of Pavia, Italy. Her research focuses on situated cognition, human-technology interaction, and the role of ignorance in cognitive and scientific processes.