Placing Other Mammals by the Scale of Primate Intelligence

Just after I wrote of avian intelligence, and how well it matched expected primate intelligence if we measured by the animal’s forebrain neuron number, I realised that that narrow methodology wasn’t the one I would employ to estimate another primate. For primates, the data might imply two thirds of all interspecific variance is explained by raw brain size, but they also imply that much of the rest is explained by dietary breadth. Together, these two factors explain at least three quarters of all variance – a stunning result! Given that, I would never have left out dietary range when estimating an unknown monkey species. As such, I should have added 20 odd points to omnivorous birds such as crows and kea (the only omnivorous parrot), and subtracted about thirty points from any animal whose diet consists of a single plant. If I had, the figures given would have been even closer to those IQs I estimated by direct comparison. If I make any such adjustment below I will signal it like so: kea 87+20=107, or giraffe 105-15=90 (giraffes mostly eat acacia leaves, but sometimes include other leaves in their diets).

asian elephants

There are at least three other cognitively gifted genera that I can immediately place on the primate scale, namely the Asian and African elephants, and the bottlenose dolphin, which I would assign values of 155, 165 and 155 respectively. I came to these values by a direct comparison of how these animals performed on several tasks, compared to their closest matched primate (the gorilla). Unlike the birds, these mammals seem to match their closest primate over most cognitive tasks. Judging by these few, mammals seem much easier to rank by this primate scale than birds do. Given that, it is odd that few have tried to do so among the non-primate mammals. For that matter, it is just as odd that it wasn’t attempted for primates before the existence of a universal primate general intelligence factor was proved by analytic means. The benefit of using an absolute scale, as I have begun to construct here, is that it can be used to base more speculative extrapolations. As I hope to show, this has significant consequences. Now to continue to our story.

The world renown neuroscientist, Susana Herculano-Houzel, has hypothesised that any mammal’s intelligence may prove to parallel its neocortical neuron count more closely than any other physiological metric. Very few such counts have been done to date, and of the above three animals, we only have 2 sets of data, both for the African elephant. Stained surveys indicate that this number is 11 billion. The best fit from my primate data would match that to an expected primate intelligence of 214. By contrast, isotropic fraction puts it at 5,6 billion, with a best fit of an IQ of 173. If the second technique proves the better method (the jury is still out on it) then, what should be a ballpark figure, would turn out to be remarkably accurate for this species.

 

Mammals by the Primate Intelligence Scale

 

 primate intelligence scale  mammal
250  humans of 60,000 BC
195  orangutan
180  chimpanzee
165  African elephant
160  gorilla
155  bottlenose dolphin
155  Asian elephants
130  California sea lion
 110  macaque
 70  Capuchin
 50  baboon
 30  pig

 

A natural effect is to exaggerate the intelligence of any species we are closely associated with. Even though we may try and be impartial, we have a better understanding of their performance, and greater insight into their strengths. This is even true of academia, where it is often proposed that rats and mice might be more intelligent that an unstudied ‘average’ mammal. However, if this comparison is restricted to equally well studied animals, such as dogs, pigs, Capuchins, or crows, they always come out very poorly. This should come as no surprise since, if neocortical neuron counts are useful in ballparking rodent intelligence, the expected intelligence of rats and mice would be would be way below our zero mark at -144+20=-124 and -194+20=-174 respectively. My page on how I derived the scale is HERE, where, I aimed for zero approximating average mammalian intelligence. Rats are known to possess many valuable cognitive skills, but none of those would place them in the most able echelon of mammals, which is all that interests us here.

pig and piglet

Of all our domesticated animals, the pig is now thought to be most intelligent. Recently, these abilities have been touted, due to its role as a food animal, however, these same tests place it well below the Capuchins with consistency. This matches neurological data, since pigs have significantly fewer neurons in their forebrain than Capuchins, and only a third as many as kea or crows.

If pigs were primates, and we knew nothing of them other than that stained neocortical neuron count and diet, their expected intelligence on our scale would be 16+20=36, with an upper limit of around 50. This would place them at the level of baboons at best. From the studies I have read, baboons actually do have a distinct edge over pigs in most cognitive tasks, and 20 to 40 is likely their real level. Recently, isotropic fractionation has also derived one neuron count for the domestic pig. This corresponds to an IQ of -4+20=16 on our scale.

The next smartest domestic animal is probably the the dog. Canine intelligence seems to vary wildly by breed – far more than for other domesticated animals. If we measure by breed, then border collies likely displace pigs in this role, possibly reaching as high as 70 points.

California sea lion

A fourth mammalian order that is reputed for intelligence, is the carnivora. Recently, a test, known to correlate strongly with g, was applied to a wide group of them, and the animal that scored highest was the polar bear. This came despite evidence that their procedure was heavily biased against animals with larger bodies. Unfortunately the only animal among them of which I have sufficient data to place independently is the dog/wolf. As such, I can only place the polar bear within a very wide range of IQ 50 to 150.  There is some other useful data on bears, such as their persistent failure of the mirror test, and occasional tool use, but not enough to narrow that range. Also unfortunate, is that the carnivores with the largest brains in this order, the pinnipeds, were not part of this survey.

Of the pinepeds, only the California sea lion is sufficiently well studied for a direct estimate of its intelligence. Judging from their language and logic skills, they exceed the abilities of Capuchins, but not gorillas. Their closest well-studied match should be macaques, with an IQ of 110 on our scale, and which they seem to have an distinct edge over. This also raises questions over the intelligence of the walrus which has a brain three times larger, is held in captivity, and yet has not been subjected to useful cognitive testing.

Sanc0602

Returning from these fixed points, it is time to extrapolate out to whales, for which we only have neuron number. This subset constitutes just the pilot whale, fin whale, and false killer whale, all of which are known only from stained counts. These have neocortical neuron counts of 37.2 billion, 15 billion, and 10.5 billion respectively. As I have argued elsewhere, because these animals employ unihemispheric sleep, these raw numbers should be divide by a significant factor <2. Assuming that factor is exactly two we have.

 primate intelligence scale  mammal
250  humans
246*  pilot whale
191*  fin whale
180  chimpanzee
169*  false killer whale
155  bottlenose dolphin

*figures derived from neuron count only.

Extrapolating further through serial application two sets of data, each calibrated from only a few points, we can derive values of 335, 285, and 260 for sperm whales, killer whales, and humpback whales respectively, based on brain size and probable neural densities. Though we can’t take these figures too seriously given the current paucity of data, I still find it of interest in terms of potential for discovery. Here the sperm whale figure exceeds that of modern humans that are raised under ideal conditions with low bacterial loads (originally designed to be the 300 point IQ mark on the scale I am using.)

Alternatively, we could normalise those three most speculative cetacean figures on the bottlenose dolphin brain weight. For the bottlenose dolphin, the IQ value can be directly calculated due to its many cognitive similarities to the gorilla. If we take that approach, then those three figures would drop by 25 points. Even then, both killer whales and sperm whales would have expected species average figures that fall well within the normal range of humans raised under modern conditions. Yet more reason for further research on the largest cetaceans.