In his classic book, Thinking, Fast and Slow, Nobel laureate Daniel Kahneman poses an interesting question:

How many animals of each kind did Moses take into the ark?

Kahneman explains:

The number of people who detect what is wrong with this question is so small that it has been dubbed the “Moses illusion.” Moses took no animals into the ark; Noah did.

I don’t know about you, but the Moses illusion fooled me. So what’s going on here?

Kahneman offers the question as an example of “norm theory,” the idea that we have internal models for the sets of relationships (between concepts, things, people, etc.) that we expect to experience in the world — i.e., the “norm.”

We don’t encounter concepts on their own but in relation to other concepts. Relations create contexts that change how we understand concepts, even leading us to misread them. (As you likely did with the question above.) Kahneman explains:

The idea of animals going into the ark sets up a biblical context, and Moses is not abnormal in that context. You did not positively expect him, but the mention of his name is not surprising.

But it’s not just a matter of context; the structure of the individual words matters. Kahneman elaborates:

It also helps that Moses and Noah have the same vowel sound and number of syllables… Replace Moses with George W. Bush in this sentence and you will have a poor political joke but no illusion.

When we design a navigation menu or bar, we’re creating a set of concepts. The clarity of each concept matters. But people don’t understand concepts in isolation: we understand them in relation to other concepts around them.

Changing just one concept can change the context. Consider the following sets of terms:

  • baseball, bat
  • bat, cave

Both sets have a word in common, “bat.” But even though it’s the exact string of three characters, it’s not the same word at all. The label’s meaning is different in each case; the set is carrying the conceptual load.

Information architecture isn’t just about clarifying individual terms but about creating clear distinctions. Distinctions manifest in labels but also in relations between labels. Effective IA calls for intentional context-making through language.