The Fish In The Mirror Knows Who They Are

For a very long time, humans have had a convenient way of deciding how intelligent other animals are.

We invent a test. We design it around behaviours that make sense to us. Then, when another animal fails to perform exactly as expected, we conclude that they must lack whatever ability we were supposedly testing.

New research on cleaner wrasses is another reason to question that arrogance. Scientists studying bluestreak cleaner wrasses have found evidence not merely that they can recognise themselves in a mirror, but that they may possess a sense of self before ever seeing one. That is a fairly extraordinary thing to discover about animals routinely dismissed as “just fish”.

Who Is That?

The traditional mirror test involves placing a mark somewhere on an animal’s body where they cannot see it directly. If they look in a mirror and then investigate or attempt to remove that mark from their own body, researchers take this as evidence that they understand the reflection represents themselves. Cleaner wrasses have already passed this test. But researchers wanted to know something else. How quickly could they recognise themselves?

Nine fishes were marked on their throats before ever being shown a mirror. When the mirrors were uncovered, their behaviour followed a remarkable pattern. Seven initially behaved aggressively towards their reflections, apparently treating the stranger in the glass as another fish. Then something changed.

They began performing behaviours researchers describe as contingency-testing. They moved their bodies while watching what happened in the mirror, seemingly testing whether the movements of the mysterious fish matched their own. Within four hours, six had attempted to scrape the marks from their throats. The first did so after only around 20 minutes.

The researchers argue that this rapid recognition strongly suggests the fishes already possessed a sense of self. They did not spend days staring into a mirror and somehow learn to become self-aware. They already were.

Then They Started Experimenting

What happened afterwards may be even more fascinating. Once the fishes appeared to understand who they were looking at, their behaviour towards the mirror changed again. The aggression and contingency-testing disappeared. They began doing things researchers interpreted as exploring how the mirror itself worked. Three fishes were observed carrying pieces of shrimp towards the mirror, dropping them and watching their reflections as they sank.

Think about what is being suggested here. First:

“Who is that?”

Then:

“Why do they move when I move?”

Then:

“That’s me.”

And finally:

“What else does this strange thing do?”

Whatever precise language we eventually use to describe these cognitive processes, this is plainly not the behaviour of an unthinking biological machine. And cleaner wrasses are already known for other impressive cognitive abilities. But their importance goes beyond proving that one particular species of fish is clever. The research raises a much more uncomfortable possibility. Perhaps our tests have been failing the animals, rather than animals failing our tests.

Maybe The Test Was The Problem

Manta rays and pigs, for example, have demonstrated behaviours associated with testing mirrors or using them as tools, but failure to perform the traditional mark test has been used against the idea that they recognise themselves. But why should every animal demonstrate self-awareness in exactly the same way?

A mark that is meaningful to one species might be irrelevant to another. An animal might understand perfectly well that a reflection represents themselves while having absolutely no interest in touching whatever researchers have stuck onto their body. Humans have repeatedly confused an animal’s failure to perform our preferred behaviour with an absence of the underlying mental capacity.

The researchers behind this cleaner wrasse study suggest self-awareness could potentially be common across vertebrates. If so, we have underestimated an enormous number of animals. Again.

Meanwhile, We Use Fishes As Equipment

There is a particularly ugly context surrounding cleaner fishes.

Cleaner fish species have been deployed by the salmon farming industry as living tools for parasite control. Norway used more than 510 million wild-caught and farmed cleaner fishes between 1998 and 2024.

Use peaked at more than 61 million individuals in a single year. And what happened to them? Disease. Injuries. Emaciation. Conditions poorly suited to their needs.

A Norwegian Food Safety Authority survey found that half of farmers reported having no cleaner fishes left in cages by the time the salmons were sent to slaughter. In those cases, mortality had effectively reached 100%.

All this for a system whose effectiveness was far from overwhelming. In the largest study of lumpfishes cited in separate research, involving nearly 25,000 individuals across 80 farms, only around 3% had lice in their stomachs.

Hundreds of millions of animals were introduced into an industrial system as biological pest-control devices before sufficient evidence existed that the practice even worked particularly well.

To be clear, the cleaner wrasses in the mirror experiment are not the same species as every cleaner fish used by Norwegian salmon farms. The point is broader than that. We continue to discover extraordinary complexity in fishes while industries continue treating fishes as interchangeable units of production.

They Never Needed To Pass Our Test

There is also a danger here. We should not replace one hierarchy with another.

Cleaner wrasses do not deserve consideration because they passed an intelligence test. A fish who cannot recognise themselves in a mirror does not therefore become available for us to exploit. Neither should protection depend upon whether an animal can solve puzzles, remember faces, use tools, understand mirrors or impress scientists. The cleaner wrasse research matters because it demolishes another comfortable human assumption about other animals. But the moral case was already there. Animals do not need to prove that their minds resemble ours before their lives belong to them.

Perhaps the most revealing thing about the mirror test was never what animals saw when they looked into the glass.

It was what humans refused to see when we looked at them.