He examines at least 12 proposed corrections, including the stylobate’s roughly 6 cm rise and the columns’ roughly 18 mm entasis. The proposed visual errors they supposedly counteract are, he argues, either not real or practically imperceptible at normal distances and viewpoints. 35

A fixed geometric correction would also be implausible as a general solution: visual perception varies among observers and with distance, angle, lighting, head orientation, and other conditions. The premise of a single architectural adjustment that corrects one illusion for everyone is therefore suspect. 5

The curves and departures from strict regularity may still have been intentional, but their purpose remains unknown. Goriely’s case is against confidently assigning them the specific purpose of optical compensation, not against their craftsmanship or aesthetic importance. 5

The intellectual source is Vitruvius, whose architectural prescriptions explicitly say that corner columns should be thickened because they appear more slender when isolated against the surrounding air. 5

In the nineteenth century, Francis Penrose’s precise measurements demonstrated that the Parthenon really did possess subtle departures from straightness and uniformity. His explanation of entasis treated the central swelling as a response to a supposedly disagreeable illusion in which a straight-sided shaft looks attenuated or concave. 2

Thus, real and remarkable measurements became coupled to an inherited explanatory story—one that Goriely says lacks direct historical evidence of the builders’ intentions and scientific evidence that the alleged perceptual problem needed correcting. 35

By the “supreme knowledge fallacy,” Goriely means the tendency to idealize ancient Greeks as possessing near-magical, comprehensive mastery—as simultaneously unsurpassed mathematicians, engineers, artists, and psychologists—and then to infer ingenious hidden intentions from every anomaly. 5

The claim is also self-reinforcing: popular accounts, teaching, and now AI-generated answers can repeat a familiar, elegant narrative with confidence, giving it apparent authority without supplying primary evidence. The broader caution about chatbots is well grounded: LLM outputs do not by themselves provide epistemic justification. 7

He examines at least 12 proposed corrections, including the stylobate’s roughly 6 cm rise and the columns’ roughly 18 mm entasis. The proposed visual errors they supposedly counteract are, he argues, either not real or practically imperceptible at normal distances and viewpoints. 35

A fixed geometric correction would also be implausible as a general solution: visual perception varies among observers and with distance, angle, lighting, head orientation, and other conditions. The premise of a single architectural adjustment that corrects one illusion for everyone is therefore suspect. 5

The curves and departures from strict regularity may still have been intentional, but their purpose remains unknown. Goriely’s case is against confidently assigning them the specific purpose of optical compensation, not against their craftsmanship or aesthetic importance. 5

The intellectual source is Vitruvius, whose architectural prescriptions explicitly say that corner columns should be thickened because they appear more slender when isolated against the surrounding air. 5

In the nineteenth century, Francis Penrose’s precise measurements demonstrated that the Parthenon really did possess subtle departures from straightness and uniformity. His explanation of entasis treated the central swelling as a response to a supposedly disagreeable illusion in which a straight-sided shaft looks attenuated or concave. 2

Thus, real and remarkable measurements became coupled to an inherited explanatory story—one that Goriely says lacks direct historical evidence of the builders’ intentions and scientific evidence that the alleged perceptual problem needed correcting. 35

By the “supreme knowledge fallacy,” Goriely means the tendency to idealize ancient Greeks as possessing near-magical, comprehensive mastery—as simultaneously unsurpassed mathematicians, engineers, artists, and psychologists—and then to infer ingenious hidden intentions from every anomaly. 5

The claim is also self-reinforcing: popular accounts, teaching, and now AI-generated answers can repeat a familiar, elegant narrative with confidence, giving it apparent authority without supplying primary evidence. The broader caution about chatbots is well grounded: LLM outputs do not by themselves provide epistemic justification. 7