Even if you’re not familiar with former U.S. Defense Secretary Donald Rumsfeld, you’ve likely heard the three categories he laid out in a 2002 press conference: “known knowns,” “known unknowns” and “unknown unknowns.” While much attention has been paid to the last category — the risks we cannot foresee — an equally intriguing and often overlooked concept is a fourth category that might be called the “unknown knowns.”
These are truths, assumptions or pieces of knowledge that exist within a system or organization, but are not consciously recognized, articulated or fully understood. In aerospace, these “unknown knowns” can play a decisive role in both innovation and failure.
In aerospace, we develop and operate the most intricate machines humans have ever created. Our work transports people and cargo around the globe, to the moon and to the farthest corners of the solar system. Modern aircraft and spacecraft integrate aerodynamics, propulsion, material science, control systems, software and much more into highly coupled architectures.
We rely upon models and simulations to manage these systems. Yet, within these processes lie embedded assumptions included in the design or practices and rarely questioned or even noticed. These are the “unknown knowns.”
While Rumsfeld did not include this fourth term in his original address (now often referred to as the Rumsfeld Matrix), some have suggested this might be the most significant category of all. Social scientist Steve Rayner argued these are “the things we don’t admit that we know”— a form of what he called uncomfortable knowledge.
Philosopher and psychologist William James observed people cannot easily absorb every complex idea, which leads to overlooking their basis and influence in our decisions. “By far the most usual way of handling phenomena so novel that they would make for a serious rearrangement of our preconceptions is to ignore them altogether,” he wrote in his 1904 lecture series, “What Pragmatism Means.”
One way unknown knowns manifest is through legacy knowledge. Aerospace builds upon decades of prior work, so design standards, empirical formulas and safety margins are often inherited from earlier generations of best practices. Engineers working on one part of a system may not fully understand how their changes interact with other parts, and often no single engineering team has a complete mental model of the system’s behavior due to the complexity. This is particularly common in today’s systems, which are dependent upon millions of lines of computer code. This creates a gap between what is technically known and what is cognitively understood.
Historical aerospace incidents illustrate the dangers of unknown knowns. Failures are rarely traced to a single cause; more often, they result from misinterpreted or overlooked information that was available but not properly understood. Data may have indicated a potential problem, but its significance was not appreciated due to assumptions, normalization of deviance or fragmented communication. In hindsight, these failures often appear preventable, precisely because the necessary knowledge was present.
Take the 1986 space shuttle Challenger disaster. There were concerns about launching given that morning’s low temperature, but information about the problem was discarded. Thirteen years later, the Mars Climate Orbiter failed to reach Martian orbit because of an overlooked conversion error as it was assumed all teams were using the same units.
At the same time, unknown knowns are not purely negative. They can also be a source of innovation. Engineers often rely on intuition — patterns recognized through experience, but not fully articulated. This inferred knowledge can guide design decisions in ways that formal analysis cannot.
For instance, an experienced technologist might sense that a particular configuration “feels wrong” based on subtle cues, even before simulations confirm it. This intuition represents a form of unknown knowing: knowledge that is real and valuable, but not fully explicit.
The challenge is how to manage unknown knowns. One approach is to emphasize reflective practice.
Engineers, designers and managers should be encouraged not only to apply established methods but also to question their origins and assumptions. Documentation can play a role by capturing not just what decisions were made, but why they were made. This helps transform unknown knowns into known knowns.
Another strategy is to foster interdisciplinary communication. Bringing together experts from different domains or organizations can surface hidden assumptions and reveal gaps in understanding. For example, a materials scientist and a software engineer might interpret the same data differently, and their dialogue can uncover insights neither would reach alone.
Similarly, robust peer review and independent verification processes can help identify knowledge that is present but unrecognized. Again, documentation and broad communication could assist others seeing similar assumptions in their work.
The use of artificial intelligence could offer new tools for addressing unknown knowns. Machine learning can detect patterns and anomalies in large data sets that elude human attention. However, these tools introduce their own challenges as they create a new layer of opacity. Engineers and designers must be careful that they are not shifting unknown knowns into a different domain by relying on automated systems. This care must be particularly employed as we incorporate new technologies, materials and processes that may invalidate some of our unknown knowns.
Ultimately, the concept of unknown knowns highlights a fundamental tension in aerospace: the gap between what a system actually is and what we humans understand about it. As the industry continues to evolve toward more autonomous, adaptive and interconnected systems, this gap is likely to widen. Managing it will require not only technical expertise but also epistemological awareness: an understanding of how we know what we know — and what we might be failing to recognize.
The Rumsfeld Matrix was originally framed in the context of national security, but its relevance extends far beyond. In aerospace, the unknown knowns remind us that knowledge is not simply a matter of possession, but of awareness.
The most dangerous risks may not be those that are hidden, but those that are hiding in plain sight, embedded in our systems, practices and assumptions. By striving to uncover and understand these hidden layers of knowledge, aerospace professionals can build not only more advanced systems, but also more resilient and self-aware ones.
