Vellum: IELTS Academic Reading practice test

For a thousand years the European book was made of animal skin. Parchment is not leather: leather is tanned, which alters the protein chemically, whereas parchment is not. The skin of a calf, a sheep or a goat was steeped for days in a bath of lime, which loosened the hair and the fat so that both could be pushed off with a blunt blade. It was then lashed to a wooden frame and dried under tension, which forces the fibres into a flat sheet rather than the random tangle they would otherwise form. While it dried it was shaved repeatedly with a curved knife called a lunellum, until the sheet was thin enough to fold. The result is astonishingly durable. Manuscripts made in this way in the eighth century can still be turned by hand today.

It was also extremely expensive, because each animal yielded only a few usable sheets. A large Bible could consume the skins of more than two hundred beasts, which is to say that a single book represented a small herd, and the herd had to be raised, killed and processed before a word could be written. Nobody threw such a thing away. When a text ceased to interest its owners, the writing was scraped off and the sheets were bound into something else, producing what is called a palimpsest. The most celebrated of these carried a prayer book written in the thirteenth century over a much older copy of the works of Archimedes, including a treatise that survives nowhere else; the underlying text was recovered only in our own time, by photographing the pages in wavelengths the eye cannot see.

Paper came from elsewhere, and it came slowly. It was being made in China by the second century of our era, and it travelled west along the trade routes of central Asia, reaching Samarkand in the eighth century, Baghdad soon afterwards, then Egypt, then Spain, and Italy in the thirteenth. Each stage added something. The Italian mills at Fabriano beat their pulp with hammers driven by a waterwheel instead of by hand, which produced a more uniform sheet, and they learned to dip the finished sheet in gelatine so that ink would sit on the surface instead of spreading into the fibres like ink on blotting paper.

European paper was not made from trees. It was made from rags, chiefly worn-out linen and hemp, which were sorted, soaked until they began to rot, and beaten to a pulp in water. A workman dipped a wire mould into the vat and lifted out a thin film of fibre, which was pressed, dried and sized. Because the raw material was other people's discarded clothing, output was limited by something no mill could control: several countries forbade the export of rags altogether, and rag collecting became a recognised trade with its own routes and its own middlemen.

Paper and printing arrived in Europe as complements. A press requires a surface that is cheap enough to be bought by the ream, uniform enough to feed through a machine, and absorbent enough to take ink without smudging but not so absorbent that the impression spreads. Parchment meets none of these conditions well and costs a fortune besides. Gutenberg nevertheless printed a portion of his Bible on parchment, for buyers who felt that a book of such importance should not be made of old shirts, and the parchment copies were priced accordingly.

The rag shortage ended in the nineteenth century, when papermakers learned to reduce wood to fibre, first by grinding it mechanically and then by dissolving away the material that binds the fibres together. Paper became cheap enough for a daily newspaper to be affordable by a labourer, and the consequences for literacy and politics were very large. There was, however, a chemical price. To stop ink spreading, mills of the period sized their paper with a mixture of alum and rosin, which leaves acid in the sheet, and acid attacks the cellulose slowly from within. Books printed between about 1850 and 1990 are therefore turning brown and breaking along the fold, while the manuscripts of the eighth century are not. Librarians call the phenomenon the slow fire, and treating it costs more per volume than most libraries can find.

The industry has since changed its chemistry. Paper made to the permanence standards adopted from the 1980s is manufactured under alkaline conditions and carries a reserve of calcium carbonate that neutralises acid as it forms, and it should last several centuries. That is an achievement, but it is also a reminder of what the preceding century and a half cost. An entire period of printed culture, the period in which printing reached almost everybody, was recorded on a material that was destroying itself from the moment it left the mill.

There is a lesson in this for anyone who assumes that what survives from the past is a fair sample of what existed in it. What reaches us is filtered twice, once by what people chose to keep and once by what the material would tolerate. Skin outlasts acid paper; paper outlasts the magnetic tape of the 1970s; and none of us can yet say how long a file written this morning will remain readable. The question of what to write on has never been a merely technical one, because it decides, quietly and long afterwards, which parts of a civilisation get a vote in its own history.

Ask anyone what they would change about their memory and the answer comes back in one direction: they would like to lose less of it. Forgetting is the standard example of the mind letting us down, the thing that mislays a name at the wrong moment and empties a year of schooling into nothing. It is striking, then, how rarely the opposite case is put. A memory system that retained everything would not be an improved version of the one we have. It would be a system unable to do the thing memory is chiefly for, which is not to hold the past but to make it usable.

The clearest illustration comes from the few people whose memories do not behave normally. The Russian psychologist Alexander Luria spent thirty years studying a newspaperman who could reproduce lists of hundreds of items years after a single hearing, and who was, by any ordinary measure, a prodigy. He was also unable to read a poem, because every word summoned images so vivid that the sense of the line was lost among them. He could not summarise a story. He struggled to recognise a face he had seen in a different light, since to him it was a different face. What he could not do was the very thing an ordinary memory does constantly, which is to throw away the particulars and keep what several occasions have in common.

A more recent line of research points the same way. A small number of people have what is now called highly superior autobiographical memory: given a date from any year of their adult lives, they can say what day of the week it fell on, what they did and what the weather was. It is a genuinely extraordinary capacity, and it is also strangely narrow. On standard laboratory tests of memory these individuals score no better than anyone else, and several of them describe the recall as intrusive, arriving unbidden and refusing to be put down. Remembering more, it turns out, is not the same thing as remembering better.

The theoretical shift of the past decade has been to stop treating forgetting as decay. On the older picture, a memory faded because the trace degraded, in the way a photograph left in a window fades; the brain did nothing, and time did everything. On the newer picture forgetting is something the brain does, with dedicated machinery and its own regulation. The neuroscientist Blake Richards has argued that the goal of the system is not to maximise what is stored but to optimise decision-making, and that transience and persistence are therefore two settings of one mechanism rather than success and failure. In laboratory animals, interfering with particular molecular pathways slows forgetting measurably, and the animals become worse, not better, at tasks that require them to abandon a rule that has stopped working.

Why would a system be built this way? Two answers are usually given, and both are simple. The first is that generalisation requires discarding: a creature that remembers every individual route through a wood in full detail has a catalogue of routes, whereas a creature that has lost the particulars has the idea of a route, and the idea is what transfers to a wood it has not visited. The second is that the world moves. A precise memory of where the fruit was last spring is not merely useless if the tree has gone; it is actively misleading, and an animal that cannot let go of it will keep returning to an empty place.

Not everyone finds this persuasive. Noor Farrell, who works on the philosophy of the cognitive sciences, argues that the functional story has been extended far past its evidence. Most everyday forgetting, she points out, has no plausible benefit whatsoever: nothing is optimised when a competent adult cannot retrieve the word for a common object, or loses a set of keys, or forgets an appointment that mattered. These are failures of a system working under constraints of energy and space, and to describe them as adaptive is to fall into a habit that makes every observed feature of an organism look like a solution to something. A theory that can accommodate any outcome, she notes, has stopped making predictions.

The objection is a fair one, and it is not fatal. There is a difference between the claim that forgetting has functions and the claim that each instance of it serves one, and only the second is vulnerable to the keys and the missing word. Digestion has a function; indigestion does not thereby become useful. What the functional account does predict, and what Farrell's alternative does not, is that interfering with forgetting should impair performance, which is what the animal work reports. The clinical picture supports the same distinction from the other end. The therapist Sameera Vance observes that a large part of her caseload consists of people whose difficulty is not that memories have gone but that a particular memory will not recede, returning with its original force years afterwards, and that the treatment of such conditions amounts to helping a stalled process resume.

What I take from all this is a change in the question. If memory were a store, forgetting would be leakage and the only sensible engineering goal would be a tighter seal. If memory is instead an apparatus for predicting what is about to happen, then the past is worth keeping only to the extent that it is still informative, and discarding what has stopped being informative is not a tax on the system but part of how it works. That does not make any particular act of forgetting welcome, and it will not console anyone who has just walked into a room and cannot think why. It does suggest that the complaint is aimed at the wrong target.