Stories of serendipitous discovery are well known. We have learned many elegant episodes about serendipity; they are usually told as charming anecdotes about luck favouring the “prepared mind.” The list is long Fleming, Archimedes, Isaac Newton, X-rays, Velcro, Viagra, etc. The emphasis is placed on the moment of surprise — the fortunate accident that eventually leads to a breakthrough. They often appear as moments when something unexpected suddenly reveals a deeper pattern of reality.
During my research I have also encountered other, less frequently cited examples. One particularly striking case is Johannes Kepler’s long struggle with astronomical data, culminating in what I have started to call as his “Eight-Minute Epiphany.” After years of struggling with data that refused to fit the prevailing circular models of planetary motion, Kepler suddenly realized that Mars’s orbit must be elliptical. The insight arrived almost instantly — but only after a long period of wrestling with anomalies that refused to disappear.
Or take a much more contemporary example. For decades the aerospace industry treated rockets as disposable machines. The dominant assumption was simple: once launched, they were lost forever. Elon Musk and the engineers at SpaceX challenged this assumption with a radically different “What if…”question:
What if rockets could return, land, and fly again? What initially sounded unrealistic eventually reshaped the economics of spaceflight.
Both cases illustrate something important. Discovery did not begin with “increasing your luck surface”, or coicidensity hunting. It began with anomalies — observations that refused to fit existing explanations.
And this is precisely where the philosophical analysis of serendipity becomes interesting.
I am delighted to find new openings — the recent research of Matteo Costa and Selene Arfini invites us to reconsider what is actually happening in serendipity stories.
In their philosophical analysis of serendipitous discovery, they argue that serendipity provides two fundamentally different forms of explanation, operating at different moments in time. They call these Prospective Explanation and Retrospective Explanation.
The distinction may appear subtle at first, but it reveals something essential about how scientific discovery really unfolds.
Two faces of understanding
Retrospective explanation is the form of explanation science is most comfortable with. Once a discovery has been made and the phenomenon understood, we can reconstruct a coherent narrative describing what happened and why.
Within this retrospective perspective, discoveries often appear almost inevitable. The phenomenon fits into existing frameworks, the reasoning steps become visible, and the scientific method seems to have operated as expected.
Prospective explanation, however, describes a completely different moment.
Imagine Jocelyn Bell in 1967 examining chart paper stretching over a hundred meters. Among familiar signals and countless irregularities, a few millimeters of strange, regular pulses appear — signals that refuse to behave like interference, equipment malfunction, or human error.
At that moment she does not yet know she is observing a pulsar. What she has encountered is an anomaly.
Prospective explanation begins precisely here: in the uncertain effort to make sense of something that does not yet belong to any known framework.
The scientist does not possess a finished explanation. Instead, they search for what Costa and Arfini describe as a sense of understanding — a provisional coherence suggesting that the anomaly might represent something meaningful. This phase is exploratory, iterative, and deeply uncertain.
But it is also the moment when discovery becomes possible.
Where serendipity actually happens
This description resonates strongly with what I started to call Authentic Serendipity.
In my forthcoming book Serendipity Unleashed, (working title, out in few weeks) I argue that serendipity is not the pleasant surprise we recognize afterward, but a process that begins when an anomaly disrupts an existing explanation of reality.
From that moment onward, a demanding sequence unfolds:
What Costa and Arfini describe as Prospective Explanation corresponds almost perfectly to the abductive phase of this process — the moment when scientists must construct provisional explanations for something that does not yet fit established knowledge.
This is the point where discovery truly begins.
The curious absence of abduction
This also raises an interesting question. The reasoning process described by Costa and Arfini corresponds closely to what the American philosopher Charles Sanders Peirce called abduction — the form of reasoning through which new explanatory hypotheses are generated.
Curiously, the concept of abduction itself remains largely absent from their discussion.
This omission is surprising, because deduction and induction alone cannot produce genuinely new explanatory frameworks. Deduction applies existing rules. Induction generalizes patterns from repeated observations. Neither can explain how a scientist confronted with an anomaly is able to propose a radically new interpretation of reality.
Only abduction can do that. Abductive reasoning begins with the simple but powerful question Peirce identified as the engine of discovery:
What could be true here?
Discovery is fragile
At the moment an anomaly appears, discovery remains fragile.
The anomaly can still be ignored, dismissed as noise, or explained away by existing assumptions. Many anomalies disappear in precisely this way.
Transformative discoveries occur only when someone decides to take the anomaly seriously enough to challenge the prevailing explanation.
This is the critical moment where abductive reasoning becomes decisive.
Without it, anomalies remain curiosities rather than discoveries.
The social architecture of discovery
Costa and Arfini also emphasize something equally important: the broader context in which discoveries occur.
Retrospective explanations reveal that breakthroughs rarely belong to a single individual or moment. They emerge from networks of instruments, collaborators, conceptual frameworks, institutional settings, and even personal experiences.
Fleming required the Oxford research community to transform penicillin into a therapeutic revolution. Bell’s discovery depended on her intimate familiarity with the radio telescope she had helped construct. Penzias and Wilson needed Robert Dicke’s cosmological framework to interpret their mysterious background radiation.
Seen retrospectively, serendipity appears almost systemic. But this systemic perspective becomes visible only after the discovery has already occurred.
The urgency of anomalies
Where philosophical analyses of serendipity sometimes remain cautious is in recognizing the urgency that often accompanies anomalies.
When an anomaly appears, scientists face a choice. They can ignore it, postpone it, or pursue it despite uncertainty.
History shows that transformative discoveries frequently depend on this moment of commitment.
The anomaly must not only be noticed. It must be taken seriously enough to question the prevailing explanation.
In other words, the scientist must be willing to create a better story.
Serendipity as a real-time process
Seen from this perspective, serendipity is not the charming accident it is often portrayed to be. It is a demanding cognitive process that begins with anomaly detection and unfolds through abductive reasoning, provisional explanations, and eventual implementation. Only later, through retrospective explanation, does the discovery appear coherent and inevitable.
The philosopher’s metaphorical DeLorean may allow us to travel back and forth through the narrative of discovery.
But the scientist working in real time has no such luxury. They must construct understanding while the story is still unfolding.
And it is precisely in that uncertain moment that Authentic serendipity truly could start and lead to successful implementation.
The choice is yours – do you carry on and forget an anomaly – or do you stop and ask “What if something else is true?”
