Our Gemara on Amud Aleph describes how every human being was endowed by the Creator with unique features, stating that no two people look alike. It can sometimes be challenging to discern whether the Gemara is presenting an absolute rule or a general observation. The Gemara serves as a legal record but also contains parables and allegories. For example, if the Gemara praises God for creating beautiful roses, is this merely a general statement of appreciation, or could every word have halachic significance, such as whether an unattractive species of roses should still be classified as roses?

The statement that no two people look alike has inspired halachic discussions about doppelgängers and eyewitness testimony (Noda BeYehuda EH I:65, EH II:62). The Noda BeYehuda grapples with the subjective experience of the possibility that two people could look alike while maintaining the Torah’s trust in verified eyewitness testimony. Despite the possibility of similar appearances, he asserts that when it truly matters, such as for testimony, fine differences will be noticed.

We have all encountered someone who looks strikingly like someone else, even when they are not related. One question to consider is whether most perceived doppelgängers are due to cognitive bias or genuinely occurring similar appearances. Just as people might see others from a different ethnicity as “all alike,” specific features that stand out to an observer may evoke a sense of familiarity, leading to perceived similarity. Our brains often take perceptual shortcuts, filtering out differentiating details in favor of processing overall patterns and large amounts of visual data. Modern facial recognition software is improving our understanding of facial uniqueness, with significant implications for legal identity verification.

Scientifically, it is impossible to definitively assert that no two people will ever look alike because, given infinite possibilities, a perfect match could occur. If facial uniqueness depends on 3,000 different factors—bone structure, complexion, feature spacing, etc.—while highly unlikely, repetition of these factors cannot be ruled out entirely. This concept mirrors the thought experiment where an infinite number of monkeys typing on infinite keyboards will eventually produce not only the Bible but also countless other texts, even backwards. While impossible to say, we can declare it highly improbable.

Researchers have investigated the necessary variables for facial recognition to ensure such a low possibility of repetition that duplication becomes virtually impossible. Lucas and Henneberg explored this in their study (“Are human faces unique? A metric approach to finding single individuals without duplicates in large samples,” Forensic Science International, 2015). They found:

“Singularity was consistently achieved with a combination of up to seven traits. The larger the traits in dimension, the faster singularity was achieved. In samples of 200, 500, etc., it was determined that about one trait needs to be added for every 1,000 individuals to maintain uniqueness. With four facial dimensions, the probability of finding a duplicate is about 10⁻⁷, while eight traits reduce this to 10⁻¹⁴, or less than one in a trillion.”

These findings support the rabbinic tradition that facial recognition relies on so many variables that duplication is virtually impossible, and therefore and impossible enough so as to validate legal rulings based on identification, even though, statistically, anything remains possible.

Translations Courtesy of Sefaria, except when, sometimes, I disagree with the translation

Do you like what you see? Please subscribe and also forward any articles you enjoy to your friends, (enemies too, why not?)