Distracted Boyfriend is built on a single fixed arrangement: a man walking with a woman turns his head to look at a second woman passing by, while the first woman looks at him with visible disapproval. The image never changes. What changes is the caption applied to each of the three figures, and that substitution is the entire mechanism of the joke.

What each figure stands for

Each figure is a labelled role, and the labels are what the poster supplies:

• The boyfriend is whoever is doing the choosing. His head is turned, so he carries the temptation or the lapse of attention.
• The woman he is looking at is the new, more appealing option. She is usually something trivial, distracting or ill-advised next to what he is meant to be committed to.
• The girlfriend is the thing being neglected. Her expression is doing the moral work of the joke: she is not surprised, she is caught out watching it happen.

Why the structure carries the joke

The humour depends on all three roles being filled with something that maps sensibly onto attraction, distraction and disapproval. Swap in a task someone should be doing, a shinier alternative, and the obligation being dropped, and the picture explains the relationship between the three captions without needing a punchline underneath. The format works because it is a diagram of a choice, and the joke collapses if only two of the three labels are given real content.

The photograph behind this format is a stock image, licensed for general commercial use. That's true of a fair few formats: someone poses for a photographer, the shot sits in a stock library available for anyone to license, and years later an unrelated crowd picks it up for a reason that has nothing to do with why it was taken.

Working out when it first got used as a meme, and by whom, is harder than it looks. A stock photograph can sit online for years before someone screenshots it and adds a caption, and the captioned version that first caught on is not always the first one anyone made. A claim about a precise date and account needs a source a reader can check for themselves, and none is available for this entry.

The format carries a joke only because the three figures pull against each other in a specific way: an established attachment, a competing pull that shouldn't be tempting but is, and someone caught looking. Swap in labels that don't set up that pull and the picture still shows a man staring at a passing woman, but there is nothing left for a reader to recognise themselves in.

What breaks it

The most common version drops the tension. A caption pair only works if the two options are genuinely competing for the same attention at the same moment. Label the girlfriend "finishing an essay" and the passing woman "a five-minute break" and there's a real pull. Label her "an urgent deadline" and the other woman "a mildly interesting email" and there's no contest, so no one reads it as a choice at all.

A second version removes the boyfriend's role. If the figure looking isn't given anything to be caught doing, the panel is just three people standing near each other. The joke needs a viewer to place themselves in his position, however briefly, and a caption that doesn't ask that of the reader tends to land as a photo with words on it.