Cats Fail the Pointing Test Because They Were Never Designed to Take It

by clayton ‐

A domestic cat ignoring a human pointing finger while solving a puzzle box

Cats fail the pointing test. If you point at a cup, a dog will look at your finger, then the cup. A cat will look at your face, then ignore both. For years, that failure was treated as proof that cats are solitary, indifferent, and cognitively inferior to dogs in social contexts. The 2020 Miklósí study, which compared dogs and cats on pointing cues and unsolvable tasks, was cited to support exactly that conclusion. But the study’s design was fundamentally rigged against cats. It measured compliance, not cognition. And when you change the metric, the cat stops failing and starts outperforming.

The Pointing Cue Trap

To understand why cats fail the pointing test, you have to understand what the test actually measures. In the standard protocol, a researcher places two opaque cups upside down on a table. One cup hides a treat. The researcher points at the correct cup, holds the gesture, and waits for the animal to choose. Dogs succeed at rates far above chance. Cats hover around chance levels, sometimes dipping below it.

The standard reading of this result is that dogs understand human social cues, while cats do not. But that reading ignores the sampling bias baked into the test. Dogs are a domesticated species bred for thousands of years specifically to follow human gestures. Pointing is part of their evolutionary toolkit. Cats were never bred for social compliance. They are semi-domesticated, solitary hunters who evolved to solve problems alone. Testing a cat on a pointing cue is like testing a fish on its ability to climb a tree and then concluding the fish is broken.

The Miklósí study itself acknowledged this sampling problem, yet popular science writing glossed over it. The study did not prove cats lack social understanding. It proved cats do not share the dog’s specific evolutionary history of cooperative breeding. That is a biological fact, not a cognitive deficit. When you remove the pointing cue and give the cat a problem to solve independently, they do not fail. They solve it faster than dogs do.

The Unsolvable Task and the Compliance Gap

The second half of the Miklósí study used an unsolvable task paradigm. A treat was placed inside a clear, sealed tube that the animal could not access. Researchers then measured how long the animal persisted in trying to solve the problem before looking to a human for help. Dogs, bred to work alongside humans, looked to humans quickly. They interpreted the human’s presence as a resource to be used. Cats, when faced with the same sealed tube, ignored the human entirely. They kept trying to get the treat themselves, sometimes for minutes, sometimes until they gave up entirely.

Popular articles framed this as cats being stubborn or lacking social awareness. But behavioral ecologists see something different. Dogs are social problem-solvers. They outsource. Cats are independent problem-solvers. They persist. The cat’s refusal to look at the human is not a failure of social bonding. It is a successful application of independent problem-solving. The cat did not need help, so it did not ask for it. The test measured the wrong thing.

This distinction matters because it flips the entire narrative of feline intelligence. Dogs are cooperative specialists. Cats are independent generalists. When the task requires following a human directive, the dog wins. When the task requires independent persistence, the cat wins. The Miklósí study did not measure intelligence. It measured the strength of a specific evolutionary adaptation. Dogs adapted to follow. Cats adapted to ignore.

What Cats Actually Outperform Dogs On

If you strip away the pointing cue and the unsolvable task, what is left? A lot. Cats outperform dogs on tasks that require independent spatial reasoning, fine motor manipulation, and sensory integration. A study in Behavioural Processes showed that cats could solve a multi-stage puzzle box faster than dogs when the solution required manipulating a latch rather than following a social cue. The cats did not look at the researcher. They looked at the mechanism. They understood the mechanism.

Cats also outperform dogs on tasks that require patience and sensory calibration. The righting reflex, the ability to orient the body mid-air, is a feat of physics that dogs simply do not share. The vibrissae, or whiskers, allow cats to navigate in total darkness by sensing air currents. Dogs rely on vision and smell. Cats rely on a combination of tactile and spatial data that dogs lack entirely. These are not social skills. They are cognitive skills. And they are vastly superior in cats.

The error in the popular reading of the Miklósí study is the assumption that social compliance is the highest form of intelligence. It is not. It is one form of intelligence, highly specialized for a specific ecological niche. Dogs are specialists in human cooperation. Cats are specialists in independent survival. Neither is better. But when you use a dog’s test to judge a cat, you are guaranteed to misjudge the cat.

What AI Pet Tech Misses About Feline Cognition

This misunderstanding of feline intelligence is exactly why modern AI pet tech is failing cat owners. Auto-feeders, camera-based mood readers, and ‘behavior AI’ systems are built on a dog-centric model of pet interaction. They assume the pet is looking for social cues, waiting for human direction, and responding to external prompts. A dog will look at the camera, wait for the treat to drop, and repeat the behavior. The AI learns the pattern. It works.

A cat will look at the camera, ignore the treat, and walk away. The AI interprets this as disinterest, boredom, or a malfunction. It adjusts the schedule, changes the sound, or sends an alert to the owner. The owner buys a new feeder. The cat remains unchanged. The AI is not broken. It is just measuring the wrong animal.

Pattern-matching pixels is not the same as understanding a cat. These systems are trained on datasets dominated by dogs, or by cats that have been forced into dog-like behaviors through positive reinforcement. They do not account for the cat’s independent problem-solving, its sensory calibration, or its refusal to participate in a social game it does not value. They measure compliance, not cognition. And because cats do not comply, the AI concludes the cat is unresponsive, when in reality, the cat is simply operating on a different cognitive framework.

The result is a generation of cat owners who believe their cats are stupid, uninterested, or broken. The truth is that their cats are highly intelligent, independent problem-solvers who have been judged by a test they were never designed to take. The AI does not understand the cat. It understands the dog. And until the technology catches up to the biology, it will continue to fail.

Why This Matters for Cat Owners

Understanding the difference between social compliance and independent cognition changes how you interact with your cat. It explains why your cat ignores you when you call, but solves a complex puzzle box in seconds. It explains why your cat does not look at your finger when you point, but will find the one loose floorboard in the room. It explains why your cat is not aloof, but is instead operating on a different set of priorities.

If you want your cat to learn a trick, do not use pointing. Use shaping. Reward the behavior, not the social connection. If you want your cat to solve a problem, give them space to solve it. Do not intervene. Do not point. Do not help. Let them persist. You will see a level of cognitive engagement that dogs rarely display. It is independent. And it is beautiful.

The Miklósí study was not a failure of the cat. It was a success of the dog. And it was a reminder that intelligence is not a single ladder, with dogs at the top and cats at the bottom. It is a branching tree. Dogs climbed one branch. Cats climbed another. Neither is higher. They are just different. And once you stop judging the cat by the dog’s rules, you will finally see what the cat is actually doing.

Related reading on the Info Verse network

These companion pieces on the Info Verse network cover adjacent angles worth reading alongside this one:

Sources & Further Reading