2.4 Some conclusions
Some naturalists and ethologists insist that animals are completely without thought, mentality or conscious awareness. A few others, such as Donald Griffin, believe that even quite simple creatures may have mental states comparable in kind to our own (see Griff in (2001), from which many of my examples have been taken). The majority do not want to commit themselves, but generally avoid questions about animals’ mental life with distaste, on the grounds that mental states cannot be observed directly, so talk of them is unscientific. Boakes writes:
... attributing conscious thought to animals should be strenuously avoided in any serious attempt to understand their behaviour, since it is untestable, empty, obstructionist and based on a false dichotomy ...
Fortunately, we do not need to take sides in this debate, which I think has been bedevilled by misunderstanding, prejudice and muddled thinking. Animals may, or may not, be consciously aware of what they are doing. The main point I’ve been trying to make in my choice of case studies is simpler. It is that behaviour we would be inclined to label ‘intelligent’ can be seen on a continuum: Einstein was intelligent; we would almost certainly want to say that the chimpanzee using a tool to get at a banana is acting intelligently. Clearly a plant isn’t intelligent – how could it have a mind in any conceivable sense? But what about migrating birds, nest-building wasps or foraging gulls? It seems irrational to draw a firm line somewhere in nature and say ‘here intelligence begins’. And it is surely anthropomorphic to draw that line under Homo sapiens and exclude all else.
Every organism in the natural world has problems to solve – what David Attenborough once called ‘the trials of life’: how to find food; how to avoid becoming food; how to find a mate; how to navigate safely around a hostile and dangerous world. All existing animals have devised workable, if imperfect, solutions to these problems. And so computer scientist Luc Steels has defined intelligence as ‘success in self-preservation’. Now it seems most unlikely that wasps reason about how to build their nests or consult blueprints; birds do not have printed maps to guide them; nor do ants have charts of their territory or control their armies with orders. No animals other than humans appear to have language in any truly developed sense; still less (I believe) does nature itself have a rational plan for its own development. But nature has found ingenious answers to the trials of life. It has its own forms of intelligence.
All of our case studies were chosen to suggest that this natural intelligence – if we understand the term in a suitably liberal sense – may be widespread in nature, and that anthropomorphic definitions of the idea might not tell the whole story. My aim now is to move away from, and to contrast, the general idea of ‘intelligence’ as being exclusively human intelligence.
Try to recall some of the characteristics of human intelligence that Al seeks to represent and replicate on machines.
We came up with the following list:
- rational thought (using symbols to reason with, as in logic and mathematics)
- language (using symbols to communicate in speech and writing)
- the ability to make plans, design and foresee
- the ability to learn specialised knowledge and skills.
You might have been able to come up with some other plausible ideas.
Now, let’s try to use this to draw some conclusions about natural intelligence from what we have learned so far. If we can agree that, in general terms, intelligence can be recognised in activities that appear goal-directed, systematic and ordered, what might this suggest about animal intelligence?
Look back at Case Studies 1 to 6. What mental or behavioural capacities do you think the various animals discussed there would need to produce the goal-directed, systematic behaviour we observed? Based on your general knowledge, can you think of any additional abilities that might be necessary, and that some animals may possess? Try to think clearly about the case studies and develop a broad conception of intelligence – one that doesn’t presuppose logic, language or formal, explicit thought.
This is probably the most challenging exercise I’ve asked of you so far. However, I think it is definitely worth carrying out. I arrived at six mental and behavioural characteristics that I thought were demonstrated, at some level or other, by most of the animals in the nine case studies:
You may have come out with different ideas, but let’s look at each of these six in a little more detail and extend our analysis beyond the human sphere.
Drives. At the root of every form of animal behaviour there seem to be basic drives. Living creatures have purposes in what they do. We are all familiar with the complexity of human motives. Why am I writing this course? Naturally, I want to earn money to eke out my paltry existence; but I also have other, less tangible reasons – creativity, professional pride, desire to convey ideas I think are interesting, and so on. Why are you studying M366? Presumably, your purposes are many and various. Animal drives tend to be simpler, of course. At the root of everything lies the battle for survival, which can be boiled down to what the philosopher Dan Dennett once memorably called the four Fs: feeding, fleeing, fighting and finding a mate. Without drives, intelligent behaviour – or behaviour of any sort other than randomness – simply wouldn’t happen.
Recognition. In the case studies, we noted that birds navigate with reference to stars and landmarks, lobsters detect prey in waterborne plumes, etc. Even the humblest creatures sense their environment: single-celled amoebae are sensitive to light, for instance. But with increasing sophistication comes an increasing ability to recognise and discriminate. A lobster can tell the difference between a plume given off by a predator and one indicating a possible mate. Herring gulls can tell the right type of dropping zone from the air. Chimps recognise each other as individuals. In all these cases, what the animal is identifying is a complex pattern of features – a special configuration of stars, or a suitable combination of flatness, hardness and physical position relative to the waterline in a rock. This ability to recognise patterns as a whole will be a topic of major importance in the course.
Classification. Along with this ability to recognise comes the related ability to classify or categorise the things that are sensed. We’ve seen in the case studies that migrating birds can distinguish some landmarks and constellations from others; lobsters can classify plumes into friend or foe. In a very well-known experiment, a plane towed the basic shape illustrated in Figure 9 across the sky above fields in which birds were feeding. If the object was towed in orientation (a), the feeding birds below took no notice – the moving shape resembled a harmless goose. However, when towed in orientation (b), the shape immediately came to resemble a predatory hawk, and the birds below all went into avoidance and hiding behaviour. Clearly they were able to classify shapes at least into the simple categories ‘dangerous’ and ‘harmless’.
Humans deal with the complexities of the world by means of an intricate system of concepts in which whole families of phenomena are classified under one label – for example, the tag ‘tree’ covers the entire vast collection of things we call trees. This ability to categorise seems to be related to our capacity for language. We use a semantic marker – a word – to stand for a whole mass of features and individuals. Whether animals can have a conceptual system of any sort is a matter of fierce debate among ethologists and behavioural psychologists. No animals have a language of any sophistication, so if they have any system of concepts at all it is bound to be limited. Various experimenters have reported work in which chimps can be taught to count numerical sequences. Pigeons can learn to discriminate between categories such as ‘picture with a person in it’ and ‘picture without a person in it’, and even between some geometric shapes presented in various sizes and orientations. But there is little agreement over how such results should be interpreted, and the whole question is clouded by anthropomorphism, as the case of ‘Clever Hans’, which I’ll discuss below in Case Study 7, indicates.
Response. By this I mean rather more than the fact that animals react appropriately to the stimuli they receive. A simple creature like a lobster or an ant seems to have a fairly limited range of responses it can make: the lobster, for example, when it detects a plume, can choose to move towards it or away from it. However, although such reactions may appear straightforward at first sight, they conceal deep complexity. As I stated, it is by no means understood how the lobster maintains itself in the centre of the plume and selects the right direction to move in. And consider the elementary action of catching a ball: we wouldn’t be inclined to call it intelligent – anyone can do it. Dogs can be trained to do it. But, although we aren’t aware of it, we are performing complex calculations in real time as we move to make that clean catch. Even actions to which we would not give a single thought, such as walking and handling objects, turn out to be massively complicated processes.
But the most important point here is that animals are active. They move around the world; they are intimately involved with it; they are constantly required to respond to its demands; and they shape it to suit their needs. It is in this activity that the roots of intelligence lie. In his science fiction novel of 1930, Last and First Men, Olaf Stapledon pictured a future time in which humanity decides to construct several huge brains, without bodies, suspended in vats, intended to solve all the problems of knowledge through pure intellect:
Man, they said, is a very noble organism. We have dealt with other organisms so as to enhance in each its noblest attributes. It is time to do the same with man. What is most distinctive in man is intelligent manipulation, brain and hand. Now hand is really outclassed by modern mechanisms, but brain will never be outclassed. Therefore we must breed strictly for brain ... We must produce an organism that will be no mere bundle of relics left over from its primitive ancestors precariously ruled by a glimmer of intelligence. We must produce a man who is nothing but man ...
The governing caste were strongly opposed to this policy ... Man, they said, was essentially an animal, though uniquely gifted. His whole nature must be developed, not just one faculty at the expense of others ...
So it proves. The great disembodied brains are built, but:
[the great brains] had a growing sense that though in a manner they knew almost everything, they really knew nothing.
The normal mind, when it experiences intellectual frustration, can seek recreation in companionship, or physical exercise, or art. But for [the brains] there was no such escape. These activities were impossible or meaningless to them ...
But actually it would be much worse than this. A brain without a body could never have evolved intelligence. It would not think at all, for it would have nothing to think about.
Communication. I didn’t stress this in my discussions of the case studies but it is certainly implied in them. Some of those animals work together collectively – the ants and, to some extent, the birds. We will look more deeply at such communal behaviour early in the next course, but it seems fairly obvious that cooperative action couldn’t be achieved without some form of communication between individuals. Many animals do have elementary ‘languages’ – systems of signalling to one another. But without syntax, subtle semantics or verbal association, these have nothing remotely resembling the power of our languages.
Learning. Again, this did not appear explicitly in the case studies. However, remember that migrating birds have been shown to learn the shapes of the constellations and landmarks by which they navigate; and oystercatchers discover their feeding preferences from their parents. Learning seems crucial for intelligence. An animal that is inflexible, that has no capacity to change in response to experience, will be a poor contender in the struggle for existence. We’ll return to this point in the next course. Learning will be a major issue in the remainder of the course.
Now let’s use an SAQ to consolidate these ideas.
Look back at the six case studies and for each of them note down which, among the above six characteristics, seem to be necessary for the behaviour portrayed.
My answer is as follows:
- Case Study 1: drives, recognition, response and communication – probably classification and learning;
- Case Study 2: drives, recognition, response and communication – classification and learning are possible but unlikely;
- Case Study 3: drives, recognition, classification, response and learning – communication is not implied by the particular case study, but we know that chimps, which are gregarious animals, have sophisticated forms of communication and can even be taught the rudiments of human language;
- Case Study 4: drives, recognition, classification, response and learning;
- Case Study 5: drives, recognition, classification and response;
- Case Study 6: drives, recognition, classification, response and communication.
You may have come up with rather different ideas; the whole area is difficult because we are having to infer mental properties (if animals have mental properties at all) from outward behaviour.