How smart are dogs really? Canine cognition in comparative context
Updated: Apr 23, 2019
Stephen Lea, professor at Exeter University.
For an extra-terrestrial, unobtrusively observing Earth creatures, dogs must seem to be the cleverest of all. After living a precarious existence in the wild they ambushed humans with their canine charms and cute furry faces with the result that many now enjoy a luxurious lifestyle, with regular walks, good food, veterinary cover and plenty of fun and games, a far cry from their early years many thousands of years ago. With the valued monika of ‘Man’s Best Friend’ many people are convinced their dog’s skills and intelligence are of the highest order. But how smart are they? Stephen Lea presented at Sherborne Science Cafe seeking a convincing answer to this very question, namely whether the cognitive capacity of dogs is a special case as often claimed, or fairly average when put alongside appropriate comparison groups.
Stephen Lea came to this work as a Cambridge-trained psychologist. As an editor of the journal, Animal Cognition, he read many papers highlighting that dogs were uniquely clever but also other papers suggesting that other species could be as clever. This piqued Stephen’s interest with the result that he has spent several years attempting to put dog cognition in a comparative context with other animals. Some of this work has been conducted with Britta Osthaus and an open source paper (see Further reading) is recommended.
1.What is a dog?
Seemingly an easy question to answer but scientifically a little more complex than expected. A dog can be described (Fig.1.) using several perspectives, which essentially, in combination, define a dog. These perspectives are of value because they represent three great constraints on species cognition and are beneficial in identifying possible comparator species (see Section 4). These perspectives are:
Phylogenetic- dogs are members of the order Carnivora. There are 286 species of carnivora and the order is subdivided into cat-like feliforms (e.g. civets, hyenas, mongooses) and dog-like caniforms (e.g. bears, seals, raccoons). The planetary dog population is 400-1200 million, exceeding that of all other carnivorans combined. Whilst most carnivorans eat meat, some (e.g. the giant panda) do not. Most dogs are village dogs, scavenging and occasionally fed by humans. Domestic cats, in comparison represent a population of ~600 million.
Ecological- domestication is likely to have altered dogs’ ecological niche. Wolves, their likely direct ancestors, are social cursorial hunters, hunting in packs and able to run down (rather than ambushing) larger prey animals. Other cursorial hunters include the dhole, African wild dogs, spotted hyenas, chimpanzees, toothed whales and bottle nose dolphins.
Anthropogenic- dogs have been domesticated by humans and domesticated animals tend to share certain differences, often morphological, from their wild ancestors, (e.g. smaller size, reduced muzzle, smaller teeth) and also some behavioural differences (e.g. increased tolerance for humans, reduced dependence on foraging). Dogs were likely to have been domesticated 10—20 thousand years ago but were beaten by the pigeon as the first domesticated animal.
Fig. 1. A carnivoran, domesticated, cursorial hunter- but is his cognition exceptional?
Animal cognition refers to the mental capacities of non-human animals and the study of those capacities including animal conditioning and learning. Over the past four decades there has been a shift in approach and animal mental processes are considered by looking at behavioural consequences. Many experiments on animal cognition start by establishing some behaviour through associative learning (e.g. habitation, operant conditioning, classical conditioning) as building blocks for more complex behaviour. Classical conditioning relies on the association between stimulus and response whereas operant conditioning relies on reinforcement. The former is based on involuntary reflexive behaviour and the latter on voluntary behaviour. Habituation refers to the waning of an animal’s behavioural response to a stimulus as a result of a lack of reinforcement during continual exposure to that stimulus- a form of learning involving the elimination of behaviours not needed by the animal.
3.Extent of scientific literature and data
Whilst dog cognition is well researched, as is the cognition of some other animals to a lesser degree (e.g. chimpanzees, dolphins), data on many other species is somewhat scant. A general comparison of dogs with all species is therefore impossible to conduct. By understanding the cognition of a few species, a framework can be established which can be extended to other animals. Canine cognition is very well understood because unlike other animals they are obtainable in large numbers (good availability) and do not have to be studied in captivity (cost advantage).
4. Dogs in context- selecting comparator species
In selecting comparator species scientific literature and data must be available (see Section 3) and comparators should bear a strong resemblance with the definitional features of a dog. These perspectives (as mentioned) represent major constraints on a species’ cognitive abilities. Additionally, different species should be ideally raised in a similar fashion, though standardisation of this aspect is difficult.
A comparative approach with extant animals seems awkward rather than making a comparison to that of a single, idealised model organism. However, there is no such thing as a generalised, model animal, only particular animals and for this reason a principled selection of animal for comparison with dogs has been made.
Potential comparators are shown in Fig. 2.
Fig. 2. Phylogenetic, ecological and anthropogenic groupings of potential comparator species
Sensory Cognition: perceptual ability forms the raw material of cognition. Of importance is what information an animal can extract from perceptual input. In dogs, olfaction is excellent. They can discriminate between identical human twins and follow scent trails. Although no experimental practical data exists, big carnivorans have large nasal structures suggesting similar abilities with dog olfaction. In domestic animals, pigs are at least as good as dogs, if not better.
Visually, dogs can discriminate complex visual patterns, comparable with pigeons and primates. Dogs recognise individual human faces but so too can chimpanzees and sheep. In some tests, distinct differences are noted. Pigeons do not respond to partially occluded objects (one object partially hidden by another) but dogs, like humans, recognise the continuing solidarity of occluded objects.
As regards hearing, dogs’ abilities are excellent but not exceptional. They can distinguish different human speech sounds probably to a greater degree than other carnivorans, responding to up to 25 different commands. Cats, ferrets and horses can distinguish a few human sounds and dogs and cats can distinguish between two different humans. Recognition of individual conspecifics (members of the same species) has been demonstrated in hyenas, kittens, the dwarf mongoose and the pinniped and is highly developed in dolphins and sheep, where it is utilised very early in life. Chimpanzees can develop, unlike a dog, an extensive vocabulary.
Physical Cognition: refers to the ability of animals to operate efficiently in the world of objects. Much research has been influenced by investigations into human cognition by techniques originated by Jean Piaget (1896-1980) concerning the early development of children. Object permanence is one of the simplest tasks within this framework (if an animal knows an object which has disappeared from view continues to exist). Only great apes show evidence of understanding invisible displacement in object permanence tasks though many species cope well with visible displacement. Wolves perform as well as dogs, as do sea lions and sloth bears. Amongst social hunters, only chimpanzees have consistently shown evidence of object permanence in both visible and invisible displacement tasks. Dolphins, like dogs, perform well with visible but not invisible displacement tasks. Amongst domestic animals, object permanence in visible displacement tasks is also well developed in goats and pigs.
Animal are often tested on their abilities to perform practical ‘animal solving’ tasks. The key here is that problems should be solved spontaneously. Experiment shows dogs have little or no spontaneous insight into physical cognitive problems (e.g. pulling strings for a treat or opening a latch to escape from a box). Similarly, dogs don’t spontaneously use tools though they can be trained to use them. Amongst carnivorans, tool use has been demonstrated in giant pandas, lions, American badgers, some bears and African wild dogs (which use fences to trap large prey). Bottle nosed dolphins use tools despite a lack of limbs. Chimpanzees also show spontaneous tool use. In contrast, domestic animals demonstrate little physical cognition though some social hunters (hyenas, raptors) show more advanced physical cognition than dogs.
Whilst dogs do not excel in physical cognition, their performance is at least equalled by other members of two of the three comparison groups.
Spatial cognition: a contrast is made between small and large scale situations. Dogs learn characteristics of small areas well, but their working memory for locations is of low capacity though their long-term memory for places is excellent. Comparative literature is sparse but, so far, suggests no exceptionalism by dogs compared to all comparison species.
Social cognition: there are three aspects to social cognition:
i) Using another animal’s behaviour as a cue e.g. following a point or gaze. Whilst a dog’s ability to use another animal’s cue is impressive other carnivorans are better.
ii) Social learning (learning an adaptive behaviour from watching another animal behaving the same way)- dogs have impressive capacities for social learning which are better than other carnivorans, apart from wolves. Some social hunters, dolphins and chimpanzees, show better evidence of motor imitation than dogs. In contrast, domestic animals, with the exception of the goat and pig, show very poor social learning, especially the horse.
iii) Theory-of-the-mind (responding in a manner which suggests understanding of another animal’s cognitive processes)- an aspect of theory-of-the-mind is perspective taking, that is, can an animal understand what another animal can perceive and then predict what it will understand? If so, can it use that information either to mislead another animal (deception) or to enter into the same state of mind (empathy)? Some wolves, like dogs, show evidence of perspective taking but none is shown in other carnivorans. In domestic animals, pigs show signs of perspective taking.
Questions of whether dogs show empathy has been investigated with regard to humans rather than to conspecifics; for example, contagious yawning is taken as evidence of empathy. Dogs demonstrate this phenomenon and seem to show emotional contagion when they perceive an emotional reaction in humans.
There is scant comparative data for empathy with only dogs and chimpanzees being extensively researched. Chimpanzees are less likely than dogs to solve tasks requiring perspective taking or deception and more likely to show empathy.
Linked to theory-of-the-mind, is the concept of self-consciousness. This is evidenced in chimpanzees and dolphins by the mirror-mark test, a test which dogs fail. However, dogs rely more heavily on scent and olfaction than vision, and urine-marking tests tentatively suggest that dogs do show self-recognition.
6.Conclusion- no clear pattern of cognitive exceptionalism in dogs exists. Dog cognition is that which would be expected of a domesticated, socially hunting carnivoran. A key finding is that the cognition of dogs should be understood, not by regarding them as members of any one group of comparators, but that they exist at the intersection of all three comparative groups (see Fig. 2.) and consequently possess a unique versatility by virtue of this combination of influences. In answering the question in (section 1) ‘What is a dog?’, it is an animal belonging to all three comparator groups and all three of those qualities contribute to the overall cognitive aptitude of Man’s Best Friend.
Lea, S.E.G. & Osthaus, B. (2018) ‘In what sense are dogs special? Canine cognition in comparative context’, Learning and Behaviour, vol. 46, pp. 335-363.
DIO: 10.3758/s13420-018-0349-7 (Opensource)
Just published: ‘Mama’s Last Hug: Animal Emotions and What They Teach Us About Ourselves’, Frans de Waal.