"The question of whether a computer can think is no more interesting than the question of whether a submarine can swim"
About this Quote
Dijkstra’s line is a neat little demolition charge aimed at a popular category mistake: treating “thinking” as a natural kind that computers might either possess or fail to possess, like a soul hidden under the hood. By pairing “computer” with “submarine,” he forces the reader to hear how odd the original question sounds when you swap in a different engineered object. A submarine clearly moves through water, yet calling that “swimming” feels anthropomorphic, a metaphor smuggled in as a definition. The machine doesn’t fail at swimming; the word fails at precision.
The intent is less to dunk on AI than to police language and, by extension, intellectual standards. Dijkstra came out of a tradition in computer science that prized formal clarity over philosophical fog. In the mid-to-late 20th century, as “machine intelligence” became a public obsession (Turing tests, expert systems, sci-fi), the debate often slid into theater: grand pronouncements built on elastic verbs. Dijkstra’s quip warns that once you let metaphors do your thinking, you stop noticing what you’re not specifying - representation, reasoning, goals, error, explanation.
The subtext is a refusal to grant the debate its drama. Asking whether computers “think” invites a yes/no spectacle; asking what computations they perform, under what constraints, with what guarantees, is less glamorous but more real. Like much of Dijkstra, it’s an elitist kind of kindness: he’s saving you from an argument you can’t win because its terms were never stable to begin with.
The intent is less to dunk on AI than to police language and, by extension, intellectual standards. Dijkstra came out of a tradition in computer science that prized formal clarity over philosophical fog. In the mid-to-late 20th century, as “machine intelligence” became a public obsession (Turing tests, expert systems, sci-fi), the debate often slid into theater: grand pronouncements built on elastic verbs. Dijkstra’s quip warns that once you let metaphors do your thinking, you stop noticing what you’re not specifying - representation, reasoning, goals, error, explanation.
The subtext is a refusal to grant the debate its drama. Asking whether computers “think” invites a yes/no spectacle; asking what computations they perform, under what constraints, with what guarantees, is less glamorous but more real. Like much of Dijkstra, it’s an elitist kind of kindness: he’s saving you from an argument you can’t win because its terms were never stable to begin with.
Quote Details
| Topic | Artificial Intelligence |
|---|---|
| Source | Unverified source: The threats to computing science (EWD898 keynote text) (Edsger Dijkstra, 1984)
Evidence: Primary source is Dijkstra’s own keynote address text. The wording in the primary source is: “...the question of whether Machines Can Think, a question of which we now know that it is about as relevant as the question of whether Submarines Can Swim.” This is the origin of the commonly-quoted para... Other candidates (2) Scientific Cognition, Semiotics, and Computational Agents... (Selene Arfini, 2025) compilation96.2% ... Edsger Dijkstra, winner of the 1972 Turing Award, who said, “The question of whether a computer can think is no m... Computer science (Edsger Dijkstra) compilation36.3% n that we think computer science is about computers is pretty much the same reason that the egyptians thought geometr... |
More Quotes by Edsger
Add to List







