Why You Don't Notice Obvious Errors in What You Read
"How many animals of each kind did Moses take on the Ark?"
— Most people answer "two" without noticing it was Noah, not Moses
In 1981, psychologists Thomas Erickson and Mark Mattson discovered something troubling: people routinely answer questions containing obvious factual errors — even when they know the correct information.
When asked "How many animals of each kind did Moses take on the Ark?", over 50% of participants answer "two" without noticing the error. Everyone knows Moses didn't build the Ark — Noah did. But our brains don't catch the mistake.
This is the Moses Illusion: a demonstration that having knowledge and using knowledge are not the same thing. Your semantic memory is not automatically fact-checked against incoming information.
You'll see 12 questions. Some contain factual distortions, some don't.
Answer normally OR click "This is wrong!" if you spot an error.
Erickson and Mattson's groundbreaking 1981 study revealed that the Moses Illusion isn't about carelessness or ignorance — it's about how our brains process meaning.
The key is semantic similarity. Moses and Noah share many features in our mental representation: both are Biblical figures, both received divine messages, both are associated with water/flooding, both have similar phonological patterns (two syllables, same vowel stress).
When researchers replaced "Moses" with "Nixon" — a name with no semantic connection to biblical arks — the illusion disappeared entirely. Nobody tries to answer "How many animals did Nixon take on the Ark?"
Your brain matches incoming words to concepts based on features, not exact identity. "Moses" activates enough "Biblical-patriarch-water-story" features to pass.
We don't fully parse every sentence. We extract "good enough" meaning and move on. Detailed fact-checking is computationally expensive.
Having knowledge stored in memory doesn't mean it's automatically consulted. Recognition and recall are separate processes.
The brain predicts upcoming content. "Biblical ark story" primes related concepts, making Moses feel "right enough" to pass through.
False claims that share features with true claims slip past our mental fact-checkers. "Scientists at Harvard found..." activates trust even if the claim is fabricated.
Leading questions with subtle distortions can alter witness memories. "Did you see THE broken headlight?" vs "Did you see A broken headlight?" changes recall.
Textbook errors that use semantically similar terms often go unnoticed by students AND teachers, propagating misconceptions.
Politicians can make false claims using language patterns that match true statements. The semantic similarity helps falsehoods "feel right."
When reading important information, deliberately pause and ask: "Who specifically? What exactly? Where precisely?" Force your brain to match specifics, not features.
Assume the text contains errors. This mindset activates more careful processing. Proofreaders use this technique professionally.
When you encounter proper nouns or statistics, mentally (or vocally) repeat them. This forces deeper encoding and comparison with stored knowledge.
Experts catch more errors in their field — not perfectly, but better. Deep knowledge creates more precise semantic categories that reject impostors.
"The Moses Illusion demonstrates that we often do not notice incorrect information in what we read, even when we possess the relevant knowledge to detect the error."
— Park & Reder (2004), Cambridge Handbook of Thinking and Reasoning