Is human vision actually better suited to dark text on light, or is that just what we are used to?

There is real science here and it mostly favours your colleague, with important caveats.

The mechanism is the pupil. In bright conditions the pupil constricts, and a smaller aperture gives greater depth of field and reduces the effect of optical imperfections in the eye — much like stopping down a camera lens. A light background constricts the pupil; a dark one opens it, which lets in more aberration and softens the image slightly.

This shows up in measurements. Studies of what is called polarity generally find a small advantage for dark-on-light in reading speed, in proofreading accuracy, and in the ability to resolve fine detail. The effect is consistent enough to be considered established, and it is also small.

So it is not a rationalisation. It is a genuine optical consequence of how the eye adjusts to light level.

I would push back gently on the evolutionary framing though, because it proves less than it appears to.

Humans have functioning night vision and spent plenty of time in low light. The visual system handles both, and the specific advantage above comes from pupil optics in the moment, not from an ancestral commitment to daylight. "We evolved under the sun" is doing rhetorical work rather than explanatory work — the pupil argument stands on its own and does not need it.

Also worth noting that reading itself is evolutionarily brand new. Nothing about our visual system was shaped by text of any polarity, so appeals to what our eyes were built for run out very quickly once the question is about interfaces.

The caveats are where this gets practical, because the small average advantage does not settle what anyone should use.

Ambient light matters more than polarity. A bright screen in a dark room is uncomfortable regardless of scheme, and matching the display to the surroundings does more for comfort than the choice between them.

Some people genuinely do better with light-on-dark. People with certain visual conditions, and many with light sensitivity, report substantial benefit. That is not preference, it is a real difference in requirement.

Halation. On a dark background, light text can appear to bleed at the edges, particularly for people with astigmatism. That effect is individual and it can outweigh the general finding entirely.

Which is why the sensible design position is to support both properly rather than to pick a winner from the average result.

One claim worth being sceptical about since it circulates a lot: that either scheme meaningfully reduces eye strain or protects your vision long-term.

The evidence for that is much weaker than the confident advice suggests. What has better support is unglamorous — blink rate drops when staring at screens, and taking regular breaks and looking at something distant helps. That applies identically in both colour schemes.

Choose based on comfort in your actual lighting. It is one of the few areas where the honest answer really is that personal preference is a legitimate deciding factor.