Animal intelligence – the jury is still out

Today I looked at a YouTube video of a very clever Beagle.

Maybe he had been playing “101 things to do with a kitchen”. Who knows what his background is? But whatever it is, he can certainly be seen being extremely clever solving a problem. The key issue for researchers into animal intelligence is the way he moves a chair in order to be able to get up on a bench and then get to some food in the microwave.

As I said in the article that I have re-posted below, this is something that dogs are not supposed to be capable of. But issues of evaluating intelligence turn out to be not that simple. Read on…

Animal intelligence

A friend sent me a link to an interesting video about a Japanese man who had trained his fish to perform.  The link FYI is

You might also be interested in

I remember reading about a study of the intelligence of different species.

The test was: if the animal was conditioned to do a specific behaviour, how long did it take for it to change when circumstances changed. For example the animal was taught to find food in a specific location. Then the location was changed. I have done experiments like this with dogs. Individuals vary, but most dogs pick it up quite quickly. I can’t remember what the other species in the experiment were, but fish came last! Food was placed in the bottom left hand corner of the tank, and they learned to find it there. Then the food was placed in the bottom right hand side. The fish did 80 repetitions of going to the left side before they found it on the right side. Gloop, gloop, gloop.

It does raise the issue of what is intelligence and how do you measure it. I do remember visiting a friend’s place once. She had recently had a new front fence built. The old gate was at the end of the block, with a diagonal path leading to the front door. The new gate was in the middle of the block leading directly to the front door. As I pulled up outside the house I saw my friend’s husband who had just arrived before me. He went to the end of the block and pushed on the fence palings, where the old gate used to be, and then looked very perplexed when nothing happened. He was phenomenally intelligent but somewhat nerdy.

I think intelligence these days is regarded as being made of a range of different skill components, such as verbal, mathematical, spatial, musical, physical, social, emotional and so on, rather than being one monolithic quality.

Animal Cognition Display1

Animal Cognition Display1 (Photo credit: Pesky Library)

Animal studies continue to surprise people. Octopus have been studied and found to be more intelligent than expected. They easily learn to find their way through a maze, for example. This is an interesting one. There is nothing inherently crucial for intelligence in being able to negotiate and remember a maze. It is done as a standard test because it was used extensively in research with rats. What people forget is that the research was designed to study how rats learned. Going through the maze wasn’t important per se. They just needed a task that the rats were capable of. In the wild rats make a living by negotiating confined spaces and remembering the routes by reference to visual landmarks. That’s why they are good at mazes. Octopus make their living entirely differently, so being able to negotiate a maze is pretty impressive.

English: Photograph of a simple eight-arm radi...

A simple eight-arm radial maze with side walls.  (Photo credit: Wikipedia)

Octopus can also learn by observation, which is considered rare amongst non-human animals. It is said that cats can do it but dogs can’t, although I have anecdotal evidence of one particularly smart dog doing it. He learned to open the zip on a backpack after being shown by his owner. An octopus was recently captured from the wild, so it was totally untrained by humans. Another octopus was given training in how to open the lid of a jar in order to get (and eat) a crab inside. The octopus after some training was able to do this. Then the untrained octopus was allowed to watch, and then given the same task. It performed the task immediately.

An octopus escaping an aquarium through a thin...

An octopus escaping an aquarium through a thin crack. (Photo credit: Wikipedia)

Tool use is another skill that used to be considered a deal breaker. There is now a book about tool use by animals of a variety of species.

Have a look also at this remarkable video of a dingo.

My friend Silke, who grew up with dogs and used to train with me works at the Community Garden in Coburg. She said they are often visited by a Beagle who lives nearby. Apparently the dog gets out by moving a chair over to the fence then getting on the chair and over the fence. So-called anecdotal evidence. I meant to follow up and try to speak to the owners and see if I could confirm the story, but haven’t got around to it. The dingo video is the first documented evidence that I know of, of this sort of thing being done by canines. It is a task that chimpanzees have difficulty with.

“Behaviourists” are people who subscribe to the philosophy of Behaviourism, the view that it is not possible or necessary to know about mental states; all we can know about is observable behaviour. This is all very well, but what are we to make of the behaviour we observe? Maybe the Behaviourist would say there is no such thing as intelligence – just the ability to perform tasks. But what are we to make of observable animal behaviour? For example, there was a chimpanzee who was highly trained by a keeper in a zoo.

We have all seen cute little baby chimps dressed in clothes and holding their trainers’ hands, learning to do amazing tasks using sign language, and we marvel at the close relationship they animals seem to have with their handlers, and their almost human abilities. They are said to have the mental abilities of a human two or three year old. But these are young animals. What happens to them when they grow up? They become dangerous and aggressive. They have the strength of eight men, and the mental age of a toddler. What will they do if they don’t get their own way?

The chimpanzee in the zoo that I mentioned, highly trained and bonded with his keeper, one day, for no apparent reason, drew her arm into his cage and proceeded to eat it. It was very difficult for keepers to release the woman. She was seriously injured. Eventually she returned to work and went to visit the chimp. She said that he never again made eye contact with her. What are we to make of this? That he was embarrassed? That he knew he had done wrong?  Or like good Behaviourists, just that the animal did not emit eye contact behaviour?

Stanley Coren

The most publicised author on the intelligence of dogs is Stanley Coren.

Cover of "Intelligence of Dogs"

Cover of Intelligence of Dogs

English: portrait Stanley Coren

Stanley Coren (Photo credit: Wikipedia)

He claims that:

• Dogs’ mental abilities are close to a human child age 2 to 2.5 years.

• “There are three types of dog intelligence: instinctive (what the dog is bred to do), adaptive (how well the dog learns from its environment to solve problems) and working and obedience (the equivalent of ‘school learning’).”

• The average dog can learn 165 words, including signals, and the “super dogs” (those in the top 20 percent of dog intelligence) can learn 250 words (Note: recent reports claim that there is a Border Collie who knows 400 words)

• Dogs can also count up to four or five, said Coren. And they have a basic understanding of arithmetic and will notice errors in simple computations, such as 1+1=1 or 1+1=3.

• During play, dogs are capable of deliberately trying to deceive other dogs and people in order to get rewards (this is a quality only found in human children at the age of 3 to 4)

Ranking by breed

Coren’s most famous contribution is the list ranking the intelligence of dogs by breed – Border Collies came first.

The flaw in the list (however interesting it is) is that it is not actually a ranking of intelligence. It is based on the perceptions of various humans – obedience judges, veterinarians and so on. So it tends to be based on how dogs are seen to perform on a limited range of tasks.

Intelligence does not necessarily mean obedience, not does it mean compliance or being easy to manage. I once had a call from a client with a Border Collie who was chasing joggers in the park and being unruly at home, destroying the garden and causing various other problems. “I thought these dogs were supposed to be intelligent!” my client complained. “Yes” I said “your dog is intelligent – that’s the problem. He is a working dog. He is chasing joggers because he was bred to herd, and he is being destructive because he is a working dog cooped up in the backyard, and he needs more stimulation.” I believe that dog like this intellectual stimulation more so than just exercise. This of course assumes they have minds, not just bodies.

Measures of intelligence

• One measure is how many repetitions it takes a dog to learn a new command.

This is one that I have noticed in my carer as a dog trainer. My first German Shepherd, River, was definitely way up there. both in terms of quickness to learn and willingness to do what I asked. I remember once demonstrating to a class the method used to teach the dog a “new command” for an action the dog could already do in response to an existing command. For example River could lie down when I said the word “drop”. I asked the class to pick an entirely new word. They chose “music”. I asked how many repetitions would it take for her to be able to lie down in response to the word “music”. They estimated it would be a lot. She did it in three.

She also learnt to do scent discrimination in about three goes. This is considered to be an advanced exercise, but actually it is really easy for the dog. They can perceive different scents without any trouble, without having to think about it, so to speak, just as I can immediately see the difference between a blue ball and a red ball. The only issue is a training problem – how does the trainer communicate to the dog that the task is to pick out the article with the handler’s scent on it and leave the others?

The first two attempts River brought an article at random, and got the wrong one. I just quietly put it back without praising her, so she knew there was some problem. On her third attempt, she sniffed the unscented article, left it, then sniffed the scented article. She looked a little unsure about what to do next. I gave her a little encouragement, and she looked at me as if to say “you mean that’s what you want?” and romped back with the correct article. It was as if the penny had dropped. I praised her and she was overjoyed to understand the task and find a solution. From that day on, she knew what to do and always did it with joy. She never got it wrong.

OK, so River was a legend. But the average dog finds it difficult to learn a command (or what I would call a signal or cue). Or do they? Doesn’t it depend on the skill of the trainer?

I have seen trainers being so sloppy in the signals that they give that it’s no wonder the dog is confused. I have also seen trainers give many different cues at the same time (for example a voice command or word, a hand signal, unconscious body language, facial expressions such as raising the eyebrows, as well as contextual cues such as picking up their car keys or putting their running shoes on. Being verbal, people assume the dog knows and understand the voice command or cue. But as Dr Ian Dunbar’s famous exercise (“does your dog know what ‘sit’ means?”) demonstrates, if you strip away all the extra cues, the dog no longer knows what “sit” means. First, sit your dog any way you like. The average pet owner will say “sit”, pull up on the lead and push down on the bum. Most dogs sit. Now take the lead off: people say “sit” and push down on the bum. Some dogs drop out and the rest sit. Many people fade out the “push down on the bum” signal – almost. But they still say “sit” and lightly touch the dog’s hind quarters with one finger tip. Sit your dog without touching; the handler says “sit” and most dogs no longer know what “sit” means. Many people now teach sit by means of a lure hand movement which become an upward hand signal. Dogs are more responsive to visual signals than to voice. So stop giving the hand signal, and most dogs no longer know what “sit” means. Obedience competitors take up “training posture” (standing at attention with the dog at heel). Get the dog to face the handler, and the dog no longer knows what “sit” means. Ask the handler to sit down in a chair, and the no longer knows what “sit” means.

Once I had a class with some of my practical pet trainers combined with some obedience competitors from a local club. We did a group stay. My handlers wandered off at their leisure and sat down during the stay. The obedience competitors, used to standing at attention opposite their dogs, looked at the others, and gradually moved away and sat in chairs. Their dogs all immediately took this as a cue that the training session was over, and broke their stays. They obviously didn’t know what “stay” meant. Another of Ian Dunbar’s exercises was to ask the handler to turn his or her back on the dog, and ask the dog to sit. So no eye contact or facial expression cues. Very few dogs knew what “sit” meant under those conditions. One of the hardest ones was to lie down on the floor and ask your dog to sit. This would reveal something about the dog’s temperament. A “dominant” dog would stand over the owner. I saw a submissive dog who was horrified that her owner was lower than her, and she got quite distressed and would have burrowed through the floor boards if she could. An unruly, playful or opportunistic dog would jump all over the owner. Very few knew what “sit” meant.

I used to do an exercise in my advanced classes, in which I asked the people to stand on a chair and ask their dog to drop. Silke, who was in the class, was born in Germany where her parents ran a dog club and bred and trained German Shepherds. Silke could “read” a dog at three years of age. She was very intuitive, but at that stage did not know the theory, and could not articulate why she did something. Silke’s dog was the only dog in the class who dropped. I said “why did Silke”s dog drop?” Everyone looked puzzled. Silke said “I don’t know”. I said, “firstly, everyone said “drop” so none of the dogs know what “drop” means. Secondly everyone gave a hand signal. What was different about Silke’s signal?” Everyone looked baffled. “You all gave your normal signal, with your hand at waist height. When you are standing at ground level, the hand signal is in front of the dog’s eyes.When you are standing on a chair, it is above the dog’s head, so from the dog’s point of view, it is a completely different signal, and the dog doesn’t recognise it as meaning “drop”. Silke bent down and gave her signal at Rascal’s eye level. That’s why he dropped.”

So if the measure of intelligence is how long it takes for the dog to learn a command, surely we have to ask “what command (or cue) is the dog receiving and attending to?” Why does it take forever for most dogs to learn what “sit” means, and an instant to learn what the rustle of a plastic bag in the kitchen means? It’s not hard to teach a dog a cue, you just have to make it relevant, so that the dog attends to it in a “noisy” environment. But the other consideration is how do we know whether the dog “knows” what the command means? Presumably, if we are good Behaviourists, by observing whether or not the dog responds with the correct behaviour. But knowing is not necessarily doing. There is the issue of motivation, the “why the hell should I?” factor. I think there are three broad stages in training: 1. establish the behaviour, 2. put it on cue and 3. sustain the motivation. This is the issue raised by Coren’s next criterion.

• The other is the percentage of commands obeyed.

This is only partly a matter of intelligence. Sure, the dog has to understand the command to some extent in order to “obey” it. But performing the correct action on command (or cue, as modern trainers prefer to say) depends on the dog’s motivation as well as understanding. The dog might know “what” you ask, but may still ask “why should I?”. This can depend on various intangibles that dog trainers and owners recognise, such as “willingness”, “submissiveness”, “responsiveness” and so on as well as training and the dog knowing that there are consequences of doing it or not doing it. Responsiveness may be related to the dog’s breed-based drives, which in turn determine the dog’s interest in the most commonly used rewards, such as interaction with the handler, food and play with toys. It will also depend on the training methods used and the way the trainer has built a motivational foundation in the dog.

Cognitive tests

The most astonishing evidence of animal intelligence comes from Alex the African Gray Parrot. He learned (although this is disputed by some) to perform tasks requiring recursive logic, which nobody believed until his trainer documented it. An example of recursive logic would be doing a calculation on a calculator. For example 154 divided by 8 plus 27. Key in 154, press divide, key in 8 … oops, it won’t work that way. You have to do 8 plus 27 equals, put that into memory, clear, key in 154, press divided by, then press read memory to retrieve the 35. Go figure. Some people have trouble with that. Alex could do a task such as “pick up the long green object that is next to the object that is round but not blue.” This requires a similar sort of complex logic, getting the result of one operation, putting it aside, doing a different one, then bringing the first one back. A friend of mine once said “it’s not surprising that a parrot can distinguish between a long green object and a round red object, because they have to be able to tell the difference between a caterpillar and a berry in the wild.” However, that’s not the point. His skill was not just in recognising the objects, but in performing logical operations about them. You could say though that his intelligence was “instantiated” in a survival skill (recognising what to eat), just as the rats’ intelligence when negotiating the maze was instantiated in finding their was through their physical environment to forage in the wild. However, not everyone agreed on the interpretation of what Alex could do. David Premack, emeritus professor of psychology at the University of Pennsylvania, is known to psychologists for his work on language and primates. He said of Alex “There’s no evidence of recursive logic, and without that you can’t work with digital numbers or more complex human grammar.”

Premack is famous amongst dog trainers for another reason. He is the only psychologist to have been turned into a verb – as in “I would Premack the dog”. This refers to Premack’s earlier work on “The candy and pinball machine effect”. This was a study in which he showed that a reward should be regarded as the opportunity to engage in a behaviour, rather than as a thing. A preferred behaviour could act to reinforce a less preferred behaviour. Not only that, a child’s preferred behaviour could change according to different circumstances. Given a free choice some children would rather play with a pinball machine (or in modern times, this would mean to play video games). Other children would choose to eat candy. The first preference can be used to reinforce, or increase amount of time the child spends doing the other activity. Think of a practical situation. In my suburb there is a library on one side of the street and a swimming pool on the other. If I had to go for a swim before I was allowed into the library, this would increase  how often i went for a swim. My friend is the opposite. He likes exercise, but doesn’t like sitting in a library reading. In his case, if you didn’t let him into the pool until he had read a book, he would end up doing more reading. The behaviours have to happen in the other order.

Another example: your child is getting under your feet in the kitchen saying “what’s for dinner, Mum?” You say “go outside and play”. Think of another situation. Your child is outside playing, and you say “Come inside immediately – your dinner’s ready!’ Which behaviour do you want to reinforce, eating or playing? And which behaviour will act as a reward for the other? It all depends on what the child’s preference is at the time. You have to control the child’s access to the preferred reward, and give it only on the condition that the child does the less preferred behaviour. You don’t get dessert until you’ve eaten your greens. You don’t get to go out and play until you have eaten your dinner. You don’t get to play video games until you have spent some time playing outside. So when you are in the park, or any other rich environment, you have to “Premack your dog”. The preferred activity is to run around in the park. The less preferred activity is to sit patiently and wait to be released. You don’t get released to run in the park until you sit quietly. In the park, food is about number ten on your dog’s list of preferences, so it could actually act as a punishment, even though at home it might be a preferred reward. The best reward after your dog comes back to you in the park is to release him or her to go off and run around again. Forget about forcing fillet steak down the bloody animal’s throat.

That is a detour about the complexities of motivation. Unless you can sustain motivation, you won’t get the behaviour, no matter how intelligent the animal is.

One of my favourite stories about Alex is told by Irene Pepperberg in an article about how Alex was trained. She presented him with a tray full of different shaped and coloured objects and he had to pick the correct one, given particular criteria. She asked him to repeat the task many times, and recorded the results for her research. She must have asked him once too often. Even pleasant tasks that are rewarded become unpleasant if you have to repeat them too often. Alex reached “the point of refusal”. He cracked the shits and kicked the tray over. Irene commented that she was sure he knew the correct answer, he just got sick of doing it. She had to record it as an incorrect response in her research.

Here are some links about Alex:

And about animal minds in general:…/virginia-morell-text

My boarder Ariel says we should stop eating beef because it is not environmentally sustainable. I stopped eating pork when I read that pigs were in the top ten most intelligent animals. The top nine were either cetaceans (whales and dolphins) or primates. I don’t remember what the criterion of intelligence was in that study. I had a friend who had trained a variety of species for film and television work. She had a rare breed of black pigs, and she trained one of them. She said it was smarter and much quicker on the uptake than any dog she had ever trained. I wouldn’t eat my dog, so why would I eat pigs? Ariel said wryly “we should eat chicken, because they are more environmentally sustainable – and they aren’t intelligent.” That may be debatable. Anyone who has attended one of Bob Bailey’s chicken training workshops might disagree.

One of my dog training friends says “what is so special about intelligence?” She won’t eat anything that has a face, including fish. I don’t know about octopus. They are so alien it’s hard to tell which way is up. So, no more calamari?

My friend Valda alerted me to an article in the Weekend Australian earlier in the year about the intelligence of dogs.  It said that dogs are apparently much, much smarter than cats. I didn’t see the article but I will try to track it down. It’s a subject I could talk about for hours! Interesting issue though about how you measure intelligence. “Cat people” say that cats are just too dignified to perform tricks like dogs do, dogs are just obsequious not intelligent. How do you measure the intelligence of an animal that just keeps its thoughts to itself?

When I visited my friend Brett (who organised our high school reunion), we were talking about intelligence. He mentioned the boy who was the dux of the school, and on his academic record was the “brightest” of the boys – but he asked who did I think was the brightest of the girls? I hesitated to say, because these days intelligence is not regarded as a single entity. As I said at the beginning, psychologists talk nowadays about different types of intelligence – verbal intelligence, musical intelligence, spatial skills, mathematical skills and so on. I commented that Brett has “social intelligence”. He’s amazing the way he can make and keep contact with so many people. But I can think of about a dozen girls who were very bright (in academic terms). I also think of another school friend, who topped the state in needlework, but didn’t get any acknowledgment for her achievement, because it wasn’t academic, and therefore wasn’t valued. She has made a career of craft work, and does talks on the history of embroidery, and is going overseas to a conference about quilting. She is clearly very talented, so who says that is not classed as “intelligence”?

We still have unresolved issues about assessing animal intelligence, and what this means for whether we should eat fish. The Australian philosopher and ethicist Peter Singer said recently at a debate in the Melbourne Town Hall that she should avoid eating meat for three reasons:

  1. health: recent evidence shows that eating even a small amount of red meat increases our risk of cancer and cardio-vascular disease
  2. environmental: meat production on a large scale uses a lot of energy and land, which could otherwise be used for feeding many more people on grains and vegetables
  3. ethical: it is impossible to raise, transport and slaughter animals humanely on an industrial scale

Peter Singer said it was OK to eat oysters because they don’t have a brain and can’t feel pain. As for other species, we may have to re-assess what they are capable of. I’m still off pork.

Related articles

About transformational1

I have many interests and I have had a varied career. I am a semi-retired professional dog trainer, specialising in the use of positive reinforcement. I do some consultations, I run instructor workshops and I am setting up a Dogs and Psychotherapy Research Project. I have a law degree from Melbourne University (but have never practiced) and I am passionate about Human Rights. My first degree was in Sociology. I worked as a social researcher on issues such as low income housing, women's refuges and women in the workforce. I live in an inner suburb of Melbourne, Australia, and I have a German Shepherd called Chance.
This entry was posted in Dog training. Bookmark the permalink.

1 Response to Animal intelligence – the jury is still out

  1. Johannah says:

    I’d venutre that this article has saved me more time than any other.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s