Previous: Can AI Evolve?
Next: When and Where it Rains on the Sun | SciShow News



View count:294,206
Last sync:2022-12-03 02:15
Many people assume that fish are less intelligent than mammals, but it turns out that isn’t true at all - they are actually way smarter than you probably give them credit for.

See the Mirror Test crossover video on SciShow Psych here:

Hosted by: Stefan Chin and Brit Garner

SciShow has a spinoff podcast! It's called SciShow Tangents. Check it out at
Support SciShow by becoming a patron on Patreon:
Huge thanks go to the following Patreon supporters for helping us keep SciShow free for everyone forever:

Adam Brainard, Greg, Alex Hackman. Sam Lutfi, D.A. Noe, الخليفي سلطان, Piya Shedden, KatieMarie Magnone, Scott Satovsky Jr, Charles Southerland, Patrick D. Ashmore, charles george, Kevin Bealer, Chris Peters
Looking for SciShow elsewhere on the internet?

Image Sources:
[♪ INTRO].

Stefan: Hey Brit, did anyone ever tell you that goldfish only have a 3 second memory?

Brit: I don't think… wait what was the question again? < I see what you did there. Well it turns out that so-called fact is just plain wrong. The whole idea that fish don't have long-term memories and aren't very smart in general probably arose because their brains look very different from ours. And they're often thought of as primitive because they were among the earliest vertebrates—they split off from the rest of the gang several hundred million years ago.

But of course, since then, fish haven't just stood still—or tread water? They've continued to evolve, and their mental abilities have evolved along with them. I was hoping that, as a host of our sister channel SciShow Psych, you could help me explain how fish are actually pretty smart. < < I'd love to.

The main reason many scientists thought fish were less intelligent than mammals is that they don't have a neocortex. That's the outer part of the brain found in mammals which is particularly large and wrinkly in humans, and in us, it's responsible for higher cognitive functions like problem solving, abstract thinking and planning. So, it was long assumed that fish just couldn't do many of those things.

But in the past few decades, researchers have come up with clever ways to test their intelligence, and fish keep excelling at those tests. Take koi, those giant goldfish-like species that people keep in ponds. Many people who keep them say that they can recognize the person that feeds them and know when and where they'll be fed.

This is called time-place learning. And scientists have found those koi keepers are right. When they fed laboratory fish at one end of their tank in the morning and the other end in the evening, the fish picked up the pattern and would wait in the right place for their meals.

Fish are also capable of Pavlovian learning, just like dogs and many other vertebrates. That's when an animal learns to associate a stimulus—like food—with a completely unrelated cue—like a bell. For example, rainbowfish were able to associate a light turning on with food after about 14 repetitions.

It takes rats about 40 repetitions before they learn a similar association. And the memories fish form last a whole lot longer than 3 seconds. Lab experiments have shown that rainbowfish can remember the location of a hole in a net for a whole year—and given that they typically only live about 2 years in the wild, a year is a pretty long time.

And not only do their memories last—some fish are really quick learners. In a 2012 study, researchers gave cleaner wrasses, capuchin monkeys, chimps and orangutans the same learning test by placing equal amounts of food on two colored plates. One plate color was always removed when they ate from the other one, so the animals had to learn which color to eat from first.

The fish were actually the first to pick that up—most of the primates never got it. And they were also quickest to figure things out when the researchers switched which color plate stuck around, thereby demonstrating reversal learning—a cognitive feat usually associated with brain size, and therefore high intelligence. Fish can also remember complex location information by creating something called a cognitive map.

It's kind of like the map of your neighborhood you have in your head. And for frillfin gobies that live in tide pools in the Caribbean, having a good mental map is a matter of life and death. When scared, these gobies will leap out of their home pool into a neighboring one.

So they have to know where those neighboring pools are, otherwise they'll end up on the rocks—and become a tasty sea gull snack. In the 1950s, researchers showed that gobies removed from their home pools for up to 40 days were still able to remember the location of neighboring pools and jump into them when scared. And that's just one example—studies have found other species can remember the locations of food, predators, hiding places, and potential mates.

When it comes to navigation, fish seem to be every bit as skilled as their land-dwelling kin. But navigation doesn't necessarily mean intelligence, at least not in the way that humans define it. One of the high bars we set for intelligence in other animals is the ability to use tools—which scientists roughly define as using an inanimate object to do something you can't do with your body alone.

And—surprise!—there are fish that use tools! Several species of wrasse have been shown to use rocks to smash open sea urchins. And some freshwater fish that make nests—like cichlids and catfish—will glue their eggs to a leaf or rock so they can pick them up and relocate them easily if the nest is disturbed.

Now, there is even some evidence that fish can do math—sort of. They're not solving algebra problems, but they may be able to count, or at least tell less versus more. In scientific terms, this is called numerical competency.

This makes sense for schooling fish, since there's greater safety in bigger numbers. But even non-schooling fish seem to have this skill. A recent study used cards printed with dot arrays where the number of dots could be manipulated.

The researchers then trained goldfish to select the card with the most dots or the fewest. And they trained a lot - they did about 1200 practice sessions. In the end, these goldfish could distinguish between 10 and 15 dots.

And they could choose the most or fewest dots with about 91% accuracy. That's comparable to the performance of non-human primates with similar amounts of training! Some fish are socially adept, too.

They can tell individuals apart, remember who's who, and learn from their peers. And some fish use their social skills to work with other species. Perhaps the most impressive case of cooperation is the joint hunting done by groupers and eels in the Red Sea.

A grouper will wake its normally-nocturnal hunting buddy during the day by approaching a moray eel's cave and shaking its head in a particular way. Then, the duo will take to the reef—the grouper snacks on small fish that the moray spooks out of the reef into open water while the moray nabs the fish that dart into crevices in the reef to escape the grouper. Groupers will sometimes invite wrasses or octopuses to hunt with them, too.

And they can tell their partners where a fish is hiding by doing a special headstand! All of which suggests they've got pretty stellar social cognition—the cognitive processes related to understanding and interacting with others. Though it's clear that fish can recognize other fish, what remains contested is the idea that they might be able to recognize themselves.

In theory, if an animal can recognize itself in a mirror and thereby pass the quote-unquote mirror test, it has to have an abstract sense of self—a pretty heady concept that scientists used to think was exclusive to us and our relatives. But when exposed to a mirror, captive manta rays displayed unusual behaviors—they blew bubbles and rolled and unrolled the fins on the sides of their heads. They don't do this to other manta rays, so the researchers believe these behaviors are basically the fish's way of checking "is this me?" Most recently, cleaner wrasses were given mirrors, and they seemed to pass, too—if you want to learn more about that, I joined Brit on a whole episode about what happened and what it really means, which you can watch over on SciShow Psych.

Overall, it seems like fish are probably about as intelligent as other non-human vertebrates. Like mammals or birds, they can do some pretty smart stuff—especially species with really complex social lifestyles, like cleaner wrasses. A lot of this was probably overlooked because we humans had preconceived notions about which animals were smarter than others.

And a lot of the time, we still bias our tests towards our ideas of intelligence—though, we're getting better about that. Intelligence is subjective, and it turns out when we design better tests, we find that all sorts of species have cognitive abilities we never expected. The question becomes how fish actually perform these higher cognitive functions without a neocortex.

They do have forebrains—the part of the brain our neocortex comes from—and that seems to be where some or all of this cognition takes place. But their forebrains are smaller and less complex than the forebrains of mammals. So it's still not entirely clear how such a “primitive design” is able to pull everything off.

And figuring that out can tell scientists a lot about how brains work—even ours. < Thanks for watching this episode of SciShow! And thanks to Brit for helping me wrap my head around fish intelligence. << If you want to learn more about other species and the concept of a ‘sense of self', you should check out our episode on the mirror test over on SciShow Psych. < Yeah! And if you want learn more about how brains work or see more of Brit, be sure to subscribe to the channel while you're over there! [♪ OUTRO].