Speculative Evolution Wiki
Part of a series on speculative biology

Champion speciators
Development of intelligence
Intelligence on Earth
Offense and defense
Speculative bioenergetics
Speculative biomechanics
Speculative physiology

Also see: Development of intelligence, Intelligence on Earth, Intelligent aliens

Intelligence is hard to define exactly: it can be interpreted as a mental feature, or a set of features, that involve reasoning, planning, goal-oriented behaviour, learning from experience, abstract thought and projection, problem-solving and intentional self-adaptation to the circumstances.

A great number of speculative-biology projects, especially those about extraterrestrial life, try to include high, human-like intelligence for at least one species, for it's necessary to derive from biological interactions the extremely complex structures and behaviours that form a civilization such as ours.

Features of intelligence

Of course, living organisms cannot simply be divided as intelligent and not-intelligent. Being as complex as it is, it can be broken up in many characteristics, degrees and specializations. Xenology mentions two basic, not-too-anthropocentric sets of criteria to analyze intelligence. Edward Wilson's "behavioral hierarchy" divides it into three general levels of complexity, the last two of which are limited to animals, since they need a nervous structure to carry information:

  1. Stimulus → reaction process: simple biochemical mechanisms produce an automatic reaction to each relevant stimulus. Bacteria and protozoa can orient themselves in the environment and look for nourishment by following chemical and luminous signals; animals at this level include sponges, coelenterates and flatworms.
  2. Directed learning: arthropods, cephalopods and most vertebrates have a complex nervous system that allows the individual development of sterotyped behaviours as a response to familiar situations; however, they still rely often on genetic-based inherited behaviours (instincts).
  3. Generalized learning: primates, canids, corvids and few other animals (see "Intelligence on Earth") retain a wide range of memories, allowing the formulation of complex behaviours that can be modified and adapted to new situations and the generalization of patterns. Only a few basic biological functions are still committed to instinct.

The aerospace computer scientist Roger MacGowan has produced a list of five basic criteria that he considers both necessary and sufficient to determine intelligence itself in any conceivable lifeform:

  1. Input: the sensory input from the environment, deeded as raw material for intelligence to work on;
  2. Storage: conservation of information as memories for a future need, necessary for learning;
  3. Deduction: the ability to insert currently received information in learned categories based on memories;
  4. Induction: the extension of past experience to expected future events;
  5. Output: physical or mental activity as response to processed information.

Among the numerous capabilities that make up intelligence, these seem to be the most important:


A capability of inductive categorization is necessary to learn about the environment: an animal stung by a wasp learns to avoid everything that looks yellow and black. More complex is the formation of categories based on abstract concepts: in the Vaughan experiment, some pigeons exposed to images randomly distributed in two arbitrary sets learned to peck, rewarded with food, only the images of a particular set, considering as relevant property the membership to it, and not any perceptible feature.

The California sea lion Rio (1986-) has shown a great ability to create categories, associating pictures of animals and objects according to similarity (e.g. different crabs), abstract qualities (e.g. "food"), logical transitivity (if A→B and B→C, then A→C) and even distinguishing letters from numbers.


Memory is in turn a combination of many different features. The most known is the spatial memory, well expressed by animals that stockpile food, such as Clark's nutcracker, tits, jays and squirrels. Despite their extremely tiny nervous system, bees can remember for days data met only once, and for all life data met at least three times.

Spatial cognition

The classic labyrinth test is solved in different species with different abilities: ants and bees mark with chemicals the places where they've already been, pigeons commit to memory the environmental features, rats get to the treat at the centre of a radial labyrinth (a cetral platform with a number of arms, generally 8, only one of which contains food; it forces the rat to start anew from the centre each time) visualizing its geometric structure. Even the slime mold Physarum polycephalum, a colonial amoeba, is able to find the shortest path to exit from a labyrinth by scouting it with tendrils, mantening the colony cohesive.

Reasoning and problem solving

Over the course of several experiments, chimpanzees proved to be able to understand the operation of a structure and use it to their own advantage, for example to get food, not simply through trial-an-error or previous training, but through preemptive reasoning; New Caledonian crows appear to be able to comprehend and implement cause-and-effect relationships, and even combine more tools (also see below).


Also see: Consciousness

A baboon being subjected to the mirror test.

Self-consciousness is generally understood as the capability of an individual to distinguish itself from the environment (including other individuals) without a direct intervention of sense organs. It's very difficult to test this capability in animals: the mirror test, in which it's controlled whether an animal is able to see in a mirror a mark on its body (and therefore recognize the image in the mirror as its own image, and not another individual) has been passed by chimpanzees, gorillas, european magpies, some species of cetaceans and an elephant, but not monkeys. While of some significance, the mirror test is thought to be incomplete and biased towards sight.

Another factor currently under study is metacognition ("knowing to know"), the perception of one's own knowledge: in a 2009 experiment, some rhesus monkeys were able to evaluate the tests' difficulty by expressing uncertainty for the most complex ones, instead of trying a random answer.


Many animals are able to distinguish different quantities: angelfish seem to be able to recognize the large set, provided that it is at least twice as large as the other; pigeons and other birds can sort sets of objects from the smallest to the biggest; rhesus monkeys can  recognize the smaller between sets whose elements look different, and the same number of visual and auditive stimuli; african elephants are able to perform simple additions, computing the total number of apples left in buckets in different times.

Symbol interpretation

The capability for abstraction is extremely rare in the animal kingdom: most species reacts to an object's depictions only if realistic enough. It's strictly related to the abilities of categorization and prototype formation; for example, a human can see a face in a simple : ), identifying two dots and a curved line as the basic elements of a human face. The recognising of abstract symbols, that do not resemble at all to the symbolised object, has probably been developed with the association between animals and their footprints. Also see categorization.

Using complex languages requires symbolic abstraction to associate concepts with words (see below).

Mind theory

It's called "mind theory" the capability to attribute mental states (knowledge, beliefs, intentions, desires...) to other individuals: it's likely the single biggest incentive to intelligence in a social species. A baboon does not warn others about a danger they can't see, since it ascribes to everyone the same knowledge it has, as 3-years old children do. Chimpanzees, however, often take food for themselves only when they know individuals of higher rank can't see them, and they can deceive others giving them false notions: for example, pretending to be hurt or sick to get more food (see here, pages 41-47).

Many of the abilities above described exist, even well-developed, in organisms that are not thought to be especially intelligent, such as spatial memory in migratory birds and seed-stockpilers, or the orientation and complex language of bees. They're, however, highly specialized skills; we can give a new definition of "intelligence" as the capability of integrating and coordinate these skills together. This capability belongs, above all, to primates, cetaceans and some groups of birds (see Intelligence on Earth).

Encephalization quotient

Encephalization quotient, or EQ, is a way to measure the brain's development in an animal (or, more properly, a vertebrate) species. It could be thought that mere brain size would be indicative of intelligence, but let's consider this list of 38 animal species, ordered by brain mass (data from here, here and here):

Animal Brain mass Body mass
Sperm whale 8000 g 56000 kg
Killer whale 7200 g 7200 kg
African elephant 5000 g 6000 kg
Manta ray 2000 g 1500 kg
Bottlenose dolphin 1700 g 160 kg
Homo sapiens sapiens 1500 g 75 kg
Homo neanderthalensis 1400 g 70 kg
Walrus 1100 g 610 kg
Homo erectus 900 g 55 kg
Giraffe 700 g 1200 kg
Gorilla 600 g 250 kg
Horse 600 g 400 kg
Hippopotamus 580 g 1600 kg
Zebra 530 g 450 kg
Chimpanzee 500 g 65 kg
Cattle 430 g 470 kg
Tegu lizard 400 g 40 kg
Lion 240 g 200 kg
Monitor lizard 230 g 23 kg
Wolverine 185 g 15 kg
Pig 180 g 70 kg
Baboon 140 g 30 kg
Dog (husky) 140 g 32 kg
Sulcata tortoise 100 g 50 kg
Rhesus macaque 90 g 9 kg
Giant octopus 70 g 14 kg
Giant cuttlefish 70 g 11 kg
Cat 30 g 3 kg
Crow 13 g 0.4 kg
Rabbit 12 g 2.5 kg
Nile crocodile 9.4 g 760 kg
Saltwater crocodile 9.1 g 800 kg
African grey parrot 9 g 0.4 kg
Elephant fish 3 g 0.1 kg
Hummingbird 3 g 0.08 kg
Rat 2 g 0.4 kg
Anole 1 g 0.1 kg
Mouse 0.45 g 0.02 kg

This is not entirely useful. Obviously bigger animals need bigger brains to operate at the same level of cognition: due to its sheer mass, a hippopotamus needs a bigger brain than a cat. On the other hand, a man and several other animals have roughly similar brain mass but wildly different body sizes. We need a more refined approach.

This list gives the same species ordered according to a new feature, the brain-to-body mass ratio, that is, the brain mass divided by the body mass:

Animal Brain/body ratio
Hummingbird 1:27
Crow 1:31
Elephant fish 1:33
African grey parrot 1:44
Mouse 1:44
Homo neanderthalensis 1:50
Homo sapiens sapiens 1:53
Homo erectus 1:61
Wolverine 1:81
Bottlenose dolphin 1:94
Rhesus macaque 1:100
Cat 1:100
Tegu lizard 1:100
Monitor lizard 1:100
Anole 1:100
Chimpanzee 1:130
Rat 1:200
Octopus 1:200
Giant cuttlefish 1:200
Rabbit 1:210
Baboon 1:210
Dog (husky) 1:230
Pig 1:390
Giraffe 1:420
Sulcata tortoise 1:500
Walrus 1:550
Horse 1:670
Gorilla 1:670
Manta ray 1:750
Lion 1:830
Zebra 1:850
Killer whale 1:1000
Cattle 1:1100
African elephant 1:1200
Hippopotamus 1:2800
Sperm whale 1:7000
Nile crocodile 1:81 000
Saltwater crocodile 1:88 000

That's better, but there are still many issues. Clearly this method does not work, either.

This is due to allometry, a relationship between organs in different-sized body expressed with an exponential function. Since changing the size of an object with a constant shape doesn't leave unchanged its properties, to keep the same functions, the brain has to be scaled differently from the body: the exponent can vary, according to different estimates, from 0.3 to 0.75, with a likely average of 0.66 for mammals (for other animals it should be somewhat lower or higher depending on their metabolism, and for invertebrates a wholly different method might be needed due to their fundamental differences in the nervous system relative to vertebrates except, perhaps, in the case of cephalopods). That means that, if the body becomes n times heavier, the mammalian brain needs to become n0.66 heavier to perform effectively the same work.

Let's thus define encephalization quotient as a relationship between the brain mass and the 0.66 power of the body mass: Q = 100·E/S0.66, where E is the brain mass and S the body mass.

Animal EQ
Homo neanderthalensis 8.5
Homo sapiens sapiens 8.1
Homo erectus 6.4
Bottlenose dolphin 5.9
Chimpanzee 4.2
Crow 4.1
Tegu lizard 4.0
Monitor lizard 4.0
Anole 4.0
Killer whale 3.98
African grey parrot 3.96
African elephant 3.95
Rhesus macaque 1.74
Wolverine 1.70
Manta ray 1.60
Walrus 1.60
Hummingbird 1.59
Giraffe 1.57
Baboon 1.48
Cat 1.45
Giant cuttlefish 1.44
Dog (husky) 1.42
Elephant fish 1.37
Giant octopus 1.23
Horse 1.15
Gorilla 1.15
Pig 1.10
Zebra 0.94
Sulcata tortoise 0.76
Cattle 0.74
Lion 0.73
Rabbit 0.66
Mouse 0.59
Sperm whale 0.59
Hippopotamus 0.44
Rat 0.37
Nile crocodile 0.32
Saltwater crocodile 0.31

Much, much better. We get a reasonable-looking ranking, that still holds many surprises (but, then again, EQ doesn't perfectly correlate with actual intelligence). Still, the 0.66-method is generally well supported by known data. Generally, amphibians rank too low to be at the list, fish hardly go beyond 0.5 EQ (with some notable exceptions), while the most prodigious reptiles, birds and mammals usually go from 1.5 to 2.5 EQ.

As an average, herbivores (with the notable exception of the elephants) and insectivores stay under 1.0 EQ, while carnivores (especially cetaceans) and omnivores (especially primates) are above (see also here); social animals rank higher than solitary animals (dogs higher than cats, horses and lions higher than rats). Only a few species approach, reach or exceed 3 EQ, with both extinct and modern hominids at the very top.

Tool use

It could be said, if "tool" means simply an object extraneous to the body that an organism uses to extend its influence of the environment, that instinctive tool use is extremely widespread throughout the animal kingdom. Archerfishes spit water on their preys, reduviid bugs camouflage with remnants of killed prey, termites build their nests with mud and detritus, striated herons bait fishes with leaves and feathers.

There are also many learned and seemingly reasoned form of object use: octopi protect themselves with coconut and mollusk shells; dolphins use shells to trap fish and sponges to protect their snout; an alaskan brown bear has been observed using a stone to get rid of patches of moulting fur; elephants use branches to eliminate parasites, open passings in electric fences with rocks and cover water pools with bark to prevent evaporation; gorillas and orangutans measure with sticks the depths of the water streams they wade, and at least one orangutan has been seen trying to fish with a spear. The Galapagos woodpecker finch extracts larvae from bark using cactus spine, often breaking them to make them more manageable, obtaining up to half of their food with this method.

Using rocks to get food from hard-shelled objects seems to be an especially common skill: chimpanzees and capuchin monkeys do this to nuts, sea otters to the mollusks and sea urchins they carry on their chest, wrasses to bivalves; egyptian vultures and seagulls let respectively bones and oysters fall to the ground.

Chimpanzees, after man, may be the most skillful tool users. They extract termites from their nests and honey from beehives with sticks, use moss and chewed leaves as sponges to carry water, rocks and branches to fend off predators, and even wooden spears to hunt bushbabies. New Caledonian crows manage not only to manipulate food, but also to build more tools, for example bending wire to make a hook, to use tools to safely examine possibly dangerous items, and even to get other secondary tools, something that not even chimpanzees are able to do.


Human language has a number of features that distinguish it from most other forms of animal communication:

  • Arbitrariness: symbolic words are associated to concepts they have nothing in common with, and in fact different languages can identify any concept with any possible word (though there is a psychological relationship between some sounds and concepts: for example, o and u sound "bigger" than i and e).
  • Cultural transmission: language is taught to an individual by another; any individual can learn many languages very different from each other.
  • Discreteness: words are composed by a limited number of signals (phonemes, in the case of spoken language; characters, in written language) that can be combined in virtually infinite ways.
  • Displacement: language can be used to talk about objects or events that are not close in space and/or time to the speaker.
  • Duality: the language has a surface, mechanical level (the cluster of discrete elements) and a semantic level (the expressed concepts).
  • Metalinguistics: the language can talk about itself.
  • Grammar: the order in which expressed concept appear in a sentence, and often particles that modify the words (but meaningless on their own), change their meaning.

Yet, most of these features appear in one or in another animal language. Bee dance has a complex set of rules to indicate distant locations through the orientation and speed of movements; arbitrariness appears in the danger calls of meerkats and prairie dogs; the songs of finches and humpback whales show a great complexity, rudiments of grammar and cultural transmission between generations and populations. Even more complex abilities appear in bottlenose dolphins, that use greeting calls when they meet other groups and associate to each individual a specific call that works as a name; baboons can be taught to recognize english words from meaningless strings of text by identifying common groups of letters.

Washoe, a chimpanzee female taught to speak with sign language.

The gorilla female Koko (1971-) and the chimpanzees Nim Chimpsky (1973-2000) and Washoe (1965-2007) learned to convey simple concepts using American Sign Language, showing they're able to associate a meaning to abstract symbols. More specifically, some of them were able to combine the meaning of different words to create new terms: for example, identifying a swan, never seen before, as "water-bird" for Washoe and a ring as "finger-bracelet" for Koko. Moreover, Koko can use words as proper names (therefore as pure symbols), calling, for example, one of her pet kittens "All Ball". The bonobo male Kanzi (1980-), who communicates through graphic symbols (lexigrams), learnt some of them by watching a video of Koko, and is believed to fully understand human spoken sentences.

There's even the case of a bird, the grey parrot Alex (1976-2007), to whom a vocabulary of about 150 words was taught, was able to combine them as Washoe and Koko did, and used them to accurately describe shape, colour and number of the objects shown to him, or to ask water and food. Just before his death, he was apparently beginning to understand the concept of "nothing" or zero. What is more, he didn't need to use sign language: being a parrot, he was able to speak.

See here a more complete catalogue of animal communication.


An interesting reflection proposed in Xenology (here and following pages) is that intelligence, as a form of data storing and elaboration, could not necessarily require an individual memory, or a developed nervous system inside each body. There are three known ways in which biological systems can store information: the genome (sub-individual) the nervous memory (individual) and culture (super-individual). Each of these might possibly lead to intelligence, and to self-consciousness, that is, perception of the self as a coherent entity distinct from the rest of the universe, receiving information from the environment, with voluntary control over (part of) its own body.

Genetic sentience

In this form of consciousness, information is stored in the genetic pool of a population, and it's modified over many generations by natural selection. A rudimentary form of this kind of consciousness is found in social insects such as ants, termites and bees, and in other colonial organisms like slime molds. Ants and termites have only a very simple nervous system, slime molds have none, but they can develop complex behaviours in an instinctive and ereditary form: the architecture of anthills, termite nests, beehives and spider webs is not designed from time to time, but "hardwired" in the genome (Richard Dakwins gave the name "extended phenotype" to gene-dictated features external to the body).

An extremely simple form of genetic consciousness exists as adaptation in every organism, but it has its most high expression in social insects. A hypothetical advanced genetic intelligence would likely have its self-awareness extended to the entire colony, rather then each body (single worker ants and even queens don't show any self-regard, but colonies as a whole appear fully aware of their distinct identity and even fiercely territorial). Since genetic "memories" are innate in all the individuals, these would have a specific social role based on a specialized morphology (workers, queens, drones, soldiers...); such a society could have a complex technology.

Individual sentience

This is the only kind of consciousness of which we have a concrete example of advanced development (us). In it, data are stored (also) in a centralized nervous system that manages actions and sensorial information of a single individual, leading to a strong self-awareness (and, from this, empathy for others, at least in social species). The social structure, if it exists at all, is only partially innate, and roles in the community are learnt by each individual; there is no morphological specialization, and each of them can fulfill any role, though they have to undergo a later psychological and social specialization, that usually still allows mobility.

It can arise in solitary organisms too, but the development of individual intelligence seems linked to the complexity of societies (see development of intelligence). Given the high self-regard of its members, this is the only kind of consciousness whose society requires a government to mantain order.

Communal sentience

This is the most speculative kind of consciousness, of which we don't have any example. Here, data are stored in the common cultural heritage, allowing a much more extensive accmulation and faster elaboration. By extrapolating from the other two kinds, we can suppose that members of such a society wouldn't have a grerat specialization either morphological or social, and could fulfill indifferently any social role; they'd be able to perceive sociopolitical processes of their community with the same immediacy as their individual biological functions.

Examples in speculative biology