Signifier

A signifier is the part of a design that communicates. Where an affordance is what you can do with something, a signifier is the cue that tells you so: the flat plate on a door that says push, the handle that says pull, the underline on a link, the grip dots on a draggable card. Don Norman introduced the term to clear up years of confusion around "affordance."

Where it comes from

In his 2008 Interactions column Signifiers, not affordances, Norman wrote that "affordance" had been widely used and misused since he brought it into design, and argued that what people need, and what design must provide, are signifiers. He described a signifier as "some sort of indicator, some signal in the physical or social world that can be interpreted meaningfully." He made signifiers a core concept of the 2013 revised edition of The Design of Everyday Things, the one addition he made to the science of the original book.

Signifiers can be deliberate or accidental. Norman's examples include a crowded or empty train platform (a hint about whether the train has left), a bookmark showing how far you have read, the position of a scrollbar, a worn path across a lawn, and a flag showing wind direction.

Why it matters

On screen, almost everything is technically possible: you can tap anywhere. What people actually do depends on what they can see and interpret. Weak signifiers lead to missed features and guessing. Misleading signifiers, like decorative text styled as a button, lead to errors. Nielsen Norman Group notes that flat design often removes one of the strongest clickability signifiers, the 3D effect, making interfaces harder to use (Loranger, 2015).

How to apply it

Do:

• Give every interactive element a consistent visual treatment: color, shape, underline, icon, cursor change.
• Use placeholder text, labels and icons that state what goes where ("Ask about your documents" rather than "Type here").
• Show state: selected tabs, disabled buttons, unread counts. State is a signifier too.

Don't:

• Rely on hover to reveal that something is clickable.
• Reuse the same style for clickable and non-clickable elements.
• Let one signifier mean two things in different places, such as blue text that is sometimes a link and sometimes just emphasis.

Common mistakes

• Confusing signifier and affordance. The affordance is the action; the signifier is the cue. A hidden swipe gesture is an affordance without a signifier.
• Too many signifiers. Badges, arrows and pulsing dots on everything compete with each other and stop meaning anything.
• Ignoring social signifiers. Ratings, "most people choose this" labels and activity counts steer behavior strongly. If they are fake or inflated they become deceptive.

In AI products

The blank prompt box is the clearest case. It affords nearly any request but signals almost nothing about what works. Signifiers fill that gap: prompt suggestions, example cards in the empty state, a paperclip icon that shows files are accepted, a model or mode label that shows what is running, and a visible "stop" button while a response streams. Signifiers also matter for output: a small source marker that can be clicked, or a distinct style for AI-generated text, tells people what they can check and what came from the model.