We take it for granted that, at least in the West, literacy is the norm. But of course our ability to read is relatively new. For our distant ancestors, talking was no doubt the means of communication. They were gossiping, having arguments, trying to explain where the best hunting places were and sometimes, perhaps, recounting tales they’d heard from their parents – maybe tales telling the listeners where unexplained events had happened - the beginning of superstition.

Those who know about such things tell us that literacy became a part of human culture only about 5000 years ago. But it was not the sort of literacy we are used to now. It looks as though it came about in order to keep records: an administrative record of life in those days.

One of the most famous texts we know about, the Code of Hammurabi, was inscribed in cuneiform on a black diorite stele in Babylon in around 1750 BC. It’s a list of 282 laws governing debt, property, wages and, usually, very bloody punishments. Not that the man in the street would have been able to read it. And no doubt, those who could read it would have been the governing caste, charged with ensuring that the king’s will was both known and obeyed.

Incising marks on stone or on wax cylinders was done to record what had been produced and by whom, so that the rulers could levy taxes accordingly.

But writing anything we would now describe as literature came a lot later. This was encouraged by different alphabets and writing methods which enabled, at first, the educated classes, and then people as a whole to become involved. Writing itself, however, has been a source of controversy in many ways.

The most famous enemy of writing was Socrates, although, ironically, we only know about his view because it was written down by Plato as part of one of the 36 dialogues in Phaedrus.

But what was his objection? It sounds quite modern. He thought that, by offloading our memories onto the written parchment, our ability to remember would atrophy. And, with that, our ability to reason and think creatively. Why, after all, should we take the trouble to remember things if we can simply consult a written record?

But then he had little concept of the mass of information we now have – information no one person could hold in his memory.

It took the invention of printing, however, to bring writing to the masses – and, at the same time, far more controversy.

As we know, the printing of the bible in the vernacular language was anathema to the Catholic Church. It meant that the priest no longer had to be the intermediate in conveying God’s word to the illiterate masses. It also meant a series of disputes between people upholding different interpretations of the same words, so creating more and more religious sects.

And then there were the pamphlets which viciously attacked the politics of the day. But literacy was still largely not part of the skill-set of the working man or woman. Even in 1820, only one in 10 people over the age of 15 could read. in the UK, however, by 1900 literacy surpassed 90%. Much of the growth in the emphasis on reading came from the efforts of people to educate themselves. Those autodidacts, at least in the UK were often encouraged in this by their chapels and their trade associations. They saw it as a means of getting on in life, of at last breaking through class boundaries or of understanding God’s will – or both.

In my own family, literacy was seen principally as a means of reading God’s word. In previous generations I’m not sure that much novel reading went on. Indeed there weren’t as many novels as there are now.

Now some commentators tell us that the proportion of people actually reading stories has diminished. However, the world’s population has increased and so I suspect that, in absolute terms, the number of people reading is still going up. The existence of a supposedly ‘post-literate’ world, is belied by the sheer quantity of literature on the market.

But there is no doubt that the proliferation of podcasts and streaming services is changing things. With story-telling on an industrial scale via the television and streaming services, the world of fiction is redefining itself. But I’m not sure that it’s true to say that reading of stories is diminishing. Every time there is an interview with a celebrity, they have either just published a book or are about to. Whether using ghost writers or not. And the most popular genre seems to be murder mysteries. They’re all the rage.

And, personally, as far as obtaining knowledge of current affairs is concerned, I do not find podcasts to be my preferred means. It’s a very slow way of acquiring knowledge as compared to reading, and simply reflects the views of the people creating the podcasts, with little attempt at balance. Much quicker to read articles from different sources and then come to a conclusion.

In his excellent 1979 book ‘Objective knowledge – an evolutionary approach’, Karl Popper divided things into three worlds -

World 1 – the physical world;

World 2 - our mental and psychological dimension - the world of pain, pleasure, thoughts, ideas, fears, and hopes. Our teeth are part of World 1, but a toothache is a personal experience and so belongs to World 2.

World 3 - is the great library of human knowledge and culture.

Socrates’ objections had little or no validity, but Popper’s three worlds model tells us that there is a growing problem with our new form of word technology – Large Language Models.

As we know, they do not produce anything new, but predict the most likely subsequent words in a sequence. They are not guided by experiences, emotions or the imagination of Popper’s ‘World 2’. They possess none of these qualities. Our ‘World 2’, however, gives rise to our ‘World 3’ – the archived memories to which LLMs now have seemingly unlimited access. Yet they do not enrich it in any way.

On the contrary, they extract phrases from sources that are often unrelated to one another in order to provide an answer – any answer. Hence the hallucinatory nature of this technology. It is not merely that they do not verify their own results to ensure they are correct. They are incapable of doing so; that is not how they work. They can only assess which answers are likely, based on the data used to train them. Nevertheless, the results they produce still become part of our World 3.

And it has recently emerged that there is another problem: many of the humans now actually employed by algorithm builders to check the factual accuracy of the output are in fact checking it with other LLMs - it’s quicker and therefore pays better! And so the quality of the output being fed back into our world 3 is diminished.

But many of our fellow-citizens think that chatbots do indeed create new material. The plausibility of the output is such that we can easily imagine that we are actually dealing with an entity having a World 2 interior.

Because it is basing its output on what we have created over the centuries, what it provides has an eerie familiarity. That it produces its output at a stellar rate also gives the impression of intelligence and, when it gets it wrong, it is effusive with its apologies and, using its pet name, we correct the things we notice it has got wrong and it tries again until we have an answer which satisfies us.

A peer-reviewed study published by the American Psychological Association found that amongst 1923 adults completing simulated work tasks with AI assistance, 58% agreed that AI had done most of the thinking. This led many of the users to have a reduced level of confidence in their own abilities.

And many students – not me of course - rely on AI to write their essays, with a corresponding failure to learn.

It seems that the atrophy of our minds feared by Socrates is beginning to come to pass.

26 July 2026

Paul Buckingham