In philosophy, “qualia” refers to the subjective qualities of our experience—what it’s like to see blue or feel delight. Daniel Dennett called them “the ways things seem to us.” But for computer scientists, a different kind of introspection has long been at play: What is the essence of their field? Is it about the machines that compute, or about something more abstract?
The question has haunted the discipline since its inception. In the 1930s, mathematicians like Alan Turing and Alonzo Church laid the groundwork for a theory of computation—an abstract framework that didn’t require any physical device. A decade later, engineers built the first general-purpose electronic computers, and by the 1960s, computer science had emerged as a distinct academic field, blending these two very different traditions. Tensions were immediate.
In 1967, three towering figures—Allen Newell, Alan Perlis, and Herbert Simon—wrote a spirited letter to Science magazine. Their stance was blunt: “Wherever there are phenomena, there can be a science to describe and explain those phenomena. There are computers. Ergo, computer science is the study of computers.”
But not everyone agreed. A famous quip—often attributed to Edsger Dijkstra—claimed that computer science is no more about computers than astronomy is about telescopes. The analogy is catchy, but it may not have originated with Dijkstra, and it’s worth examining whether it captures the full picture.
As a science journalist covering the field, I’ve often leaned on that quote. It’s a flattering way to distinguish computer science from mere tech journalism—suggesting the discipline is about timeless mathematical truths, not just the latest gadgets. But I’ve also felt a nagging doubt. Is the analogy missing something crucial? After all, many of the most profound questions in computer science—from the limits of computation to the nature of algorithms—would never have been posed without the existence of actual computers.
To understand this, consider the field’s dual heritage. Its mathematical parent deals with abstract problems like the Komlós problem, which concerns the discrepancy of vectors and has implications for algorithm design. Its engineering parent builds the machines that run those algorithms. The two sides are intertwined: theory often emerges from practical challenges, and engineering often relies on theoretical insights.
Take the recent progress on AI-assisted proofs of Erdős’s problems. These breakthroughs came about because computers could explore vast search spaces, suggesting conjectures that human mathematicians might never have considered. The machines didn’t just serve as tools; they actively shaped the questions being asked.
Similarly, the digital proof revolution is forcing mathematicians to reconsider what counts as a proof. When a computer verifies a complex argument, does that enrich or constrain mathematical discovery? This is a philosophical question that only arises because computers exist.
Even in theoretical computer science, the physical machine leaves its fingerprints. The famous P vs. NP problem, for instance, is about the efficiency of algorithms—a notion that only makes sense when resources like time and memory are limited, which is precisely the reality of physical computers.
So, does computer science need computers? The answer is nuanced. The theoretical core of the field—the study of what can be computed and at what cost—can be pursued without any physical device, much like pure mathematics. But the field’s history, its driving questions, and even its identity are inseparable from the machines that brought it to life. As one researcher put it, “Computer science has two parents: a mathematical one and an engineering one, and it’s really a cross between those two.”
Perhaps the best answer is that computer science is a discipline that thrives on the tension between the abstract and the concrete. It doesn’t need computers to exist, but it would be a very different—and far poorer—field without them. The analogy to astronomy and telescopes is suggestive, but it falls short. Astronomers don’t build their own telescopes; computer scientists often do, and that hands-on engagement with the physical world is part of what makes the field so dynamic.
