The dictionary defines imagination as
a) the formation of a mental image picture
b) the formation of an idea
c) the formation of a concept that is not real or present
d) the faculty permitting visionary or creative thought.
What I am concerned with is this faculty of imagination. For example, writing a musical composition or a novel, forming a scientific theory, or designing a car or a jumbo jet. These are creative acts.
Materialists since the Enlightenment have argued that imagination is the recombination of ideas. Ideas are abstracted from the senses.
David Hume (1711-1776) said:
“…this creative power of the mind amounts to no more than the faculty of compounding, transposing, augmenting or diminishing the materials afforded to us by the senses and experience. When we think of a golden mountain we only join two consistent ideas, gold and mountain, with which we were formerly acquainted. A virtuous horse we can conceive; because, from our own feeling, we can conceive virtue; and this we may unite to the figure and shape of a horse, which is an animal familiar to us. In short all the materials of thinking are derived either from our outward or inward sentiment: The mixture and composition of these belongs alone to the mind and will. Or, to express myself in philosophical language, all our ideas or more feeble perceptions are copies of our impressions or more lively ones.” [Hume 1748]
Hume did not say how he selected these two concepts out of the millions we have. He did not say whether the combination of two concepts is always meaningful. Nor how to select a useful combination.
‘Golden mountains’ and ‘virtuous horses’ seem to be only the stuff of novels.
Hume did not address how we can talk about horses and mountains in the first place. These things do not exist in our heads but in the world.
In the 19th Century Thomas Huxley (1825-1895) claimed that conscious choice played no part in this combining of two or more concepts.
“The consciousness of brutes would appear to be related to the mechanism of their body simply as a collateral product of its working, and to be as completely without any power of modifying that working as the steam-whistle which accompanies the work of a locomotive engine is without influence upon its machinery. Their volition, if they have any, is an emotion indicative of physical changes, not a cause of such changes. … the argumentation which applies to brutes holds equally good of men.” [Huxley 1912]
This view is called ‘Epiphenomenalism’ (Greek epi= upon; phainomai = to appear). It is contrary to the frequent notion that our thoughts and desires cause our bodies to move in certain ways and towards certain goals.
The epiphenomenalist says physical events in the brain cause these thoughts and desires . The idea is that physical things such as dropping a weight on your foot causes impulses in the brain which you experience as pain. But you are not able to originate any actions. The physical processes in your nerves and brain cause the screams and hops which follow. ‘You’ are an impotent spectator stuck inside an autonomous robot. There is nobody home.
The 20th century school of psychology known as behaviourism espoused epiphenomenalism. Behaviour is always the result of some reflex in response to some stimulus. The founder of this school, John B Watson (1878 –1958) held that
“… such things [as concepts and general ideas are] mere nonsense; … all of our responses are to definite and particular things. I never saw anyone reacting to tables in general but always to some particular representative… The question of meaning is an abstraction, a rationalisation and a speculation serving no useful scientific purpose.” [Watson 1920]
We have now got rid of meaning.
In the same vein the Oxford philosopher Gilbert Ryle (1900-1976) contended that we could paraphrase any statement involving ‘mind words’ such as ‘imagine’ without loss of meaning. A statement about behaviour would be as good or better:
“picturing, visualising, or ‘seeing’ is a proper and useful concept but its use does not entail the existence of pictures which we contemplate or a gallery in which such pictures are ephemerally suspended. Roughly imaging occurs, but images are not seen… True, a person picturing his nursery is, in a certain way, like that person seeing his nursery, but the similarity does not consist in his really looking at a real likeness of his nursery, but in his really seeming to see his nursery itself, when he is not really seeing it. He is not being a spectator of a resemblance of his nursery, but he is resembling a spectator of his nursery…. There is no answer to the spurious question, ‘where do the objects reside that we fancy we see?’ since there are no such objects.” [Ryle 1949,2000]
The mind has disappeared!
The psychologist B F Skinner (1904–1990) applied the behaviourist principle to political theory. He advocated the abolition of the idea of ‘Autonomous Man’. “Freedom and dignity illustrate the difficulty. They are the possessions of the autonomous man of traditional theory, and they are essential to practices in which a person is held responsible for his conduct and given credit for his achievements. A scientific analysis shifts both the responsibility and the achievement to the environment.” [Skinner 1971, 1973].
Now, even the brain has disappeared.
The self-negating nature of the above ideas does not seem to have bothered any of their advocates. The arguments in whatever way they are mistaken, are certainly imaginative.
One of the reasons that behaviourism has fallen somewhat from favour is the rise of ‘Artificial Intelligence’. Computers have internal states. So the idea that the internal states of human beings are irrelevant is no longer plausible (if it ever was).
No one has yet built a machine that displays ‘imagination’ as well as a human. Nor one that displays the versatility of a human.
Some computer programs can now make a passable attempt at music, novels etc . IBM’s ‘Deep Blue’ has beaten the human world champions at such complex games as go and chess. This has given rise to doomsday scenarios where computers will think for themselves. They may decide that humans are suitable merely as pets or, worse, that humans are dispensable all together.
This idea depends for its force on the asserted equivalence of the human brain and a computer. In the early days of AI it was claimed that humans functioned according to certain rules such as ‘if I am thirsty, have a drink”. It soon became clear that there were so many rules it was impossible to code them all. The end result of this technology was a few applications in narrow domains of expertise.
It also led to wide-spread disillusionment with AI.
The next technical advance which revived the hope for AI was the use of statistics to digest huge quantities of input data and ‘learn’ a task. This approach when applied to speech gave rise to automatic translation. You can now speak into a smart phone in one language and get back the ‘same’ meaning in another desired language.
But this technique has limits. The translation programs need huge quantities of text in the two languages to ‘understand’ how the two languages correspond. The programs don’t understand what the various sentences refer to. Humans can understand a far bigger vocabulary than the best translation apps. They can understand made up words. They also understand that words refer to things and events in the real world.
The same approach to ‘machine learning’ led to automatic face recognition. An accessible explanation of how such ‘neural networks’ function is on YouTube [Serrano 2017] These machines need billions of faces to become proficient. Humans can recognise a person after one sighting..
The technique which currently has the AI community buzzing is so-called ‘deep learning’. This depends on multiply-connected artificial ‘neural networks’. Each network learns aspects of the task to be accomplished by the whole machine. These are then integrated in successive layers into appropriate conclusions or actions. This is the kind of thing used in self-drive cars where many sensors feed into the top layers of the neural network.
Such computers may make fewer mistakes than a human on similar tasks, but they are not infallible. They do not tire but they are not conscious. They can’t extrapolate from experience in the same way that humans can.
Some computer scientists who have invested many years in the pursuit of ‘AI’ agree:
“Today’s AI, which we call weak AI, is an optimizer, based on a lot of data in one domain that they learn to do one thing extremely well. It’s a very vertical, single task robot, if you will, but it does only one thing. You cannot teach it many things. You cannot teach it multi-domain. You cannot teach it to have common sense. You cannot give it emotions. It has no self-awareness, and therefore no desire or even understanding” [Lee 2018]
Of course, the entities which display ‘imagination’ in the development of AI are the AI designers.
Imagination is Creation
My purpose is not to define ‘imagination’ exactly. I assert that human beings have such a faculty.
Human beings are able to think creatively to a greater or lesser degree. Jumbo jets now exist when a century ago they did not. By any criterion that is creation.
[Hume 1748] Hume D An Enquiry concerning the Human Understanding Sect II (modern version ed E Steinberg)
[Huxley 1912] Huxley T Method and Results Macmillan : p240, p243 available at
[Lee 2018] Lee K-F We are here to create available at https://www.edge.org/conversation/kai_fu_lee-we-are-here-to-create
Ryle G (1949,2000) The Concept of Mind Penguin p234,237
Skinner B F (1971,1973) Beyond Freedom and Dignity Penguin p30
[Watson 1920] Watson JB Is Thinking Merely The Action Of Language Mechanisms? British Journal of Psychology11, 87-104. available at
[Serrano 2017] Serrano L A Friendly Introduction to Neural networks and Image Recognition available at