IELTS Academic Reading Practice Test 197

In the first half of the twentieth century, a radio was a piece of furniture. It stood in the corner of a living room, drew its power from the mains and was switched on for the evening news or a concert that the whole family listened to together. The sets of the 1930s and 1940s depended on vacuum tubes — glass valves containing a heated filament, a grid and a metal plate, from which electrons were drawn off and controlled. These tubes worked well enough, but they were bulky, ran hot and required voltages of a hundred volts or more.

Making a radio portable with such components was difficult. A valve set small enough to carry needed a heavy battery to supply its high voltages, and the valves themselves were fragile and consumed a great deal of current. Portable radios did exist, and were marketed for picnics and trips to the beach, but they were heavy and expensive, and often needed two separate batteries, one for the filaments and one for the rest of the circuit. The idea of a radio that a teenager could slip into a pocket and run on a few cheap cells was, in 1945, a fantasy.

The device that made this possible emerged from an industrial laboratory in New Jersey. On 23 December 1947, two physicists at Bell Telephone Laboratories, John Bardeen and Walter Brattain, demonstrated an arrangement of gold contacts pressed against a block of germanium, a semiconducting element that conducts electricity better than an insulator but not as well as a metal. Their device could amplify a signal while using only a fraction of the power a valve required. A colleague, William Shockley, went on to conceive a more robust version, the junction transistor, early the following year. The three men shared the Nobel Prize in Physics in 1956.

The advantages were obvious, but the first transistors were not. They were assembled by hand, produced in tiny numbers and cost far more than the valves they replaced. Bell made licences available to other manufacturers, and the earliest customers were the armed forces and the makers of hearing aids, for whom small size and low power consumption mattered more than price. Among the firms that obtained a licence was a small Japanese company, Tokyo Tsushin Kogyo, founded after the war by Masaru Ibuka and Akio Morita to make tape recorders.

The first transistor radio to reach the shops came from the United States. In 1954 a firm based in Dallas, Texas Instruments, which had begun making transistors for military equipment, joined with a small Indianapolis company to build a set around four of them. The Regency TR-1 went on sale in the autumn of that year, in a plastic case about the size of a paperback book. Its price — nearly fifty dollars — was high, and its sound was thin, but it proved that a radio could run on an ordinary battery.

Japanese engineers took the idea further. Tokyo Tsushin Kogyo produced its first transistor radio in 1955, and two years later introduced a model small enough to fit into a shirt pocket. Morita ordered shirts with slightly enlarged pockets so that his salesmen could carry the set and show how neatly it fitted. In 1958 the firm adopted a new name, Sony, and began selling radios in the United States in growing numbers.

Meanwhile the components inside were improving. Silicon, though harder to purify than germanium, produced transistors that worked at higher temperatures and leaked less current, and by the 1960s it had displaced germanium in most applications. Printed circuit boards replaced hand-soldered wiring, and prices fell year by year as production expanded.

The social effects were considerable. A radio that belonged to one person rather than to a household, that could be taken to a beach or carried in a car, and that could be tuned to stations playing the music its owner liked, arrived at the moment a distinct youth culture was forming around rock and roll. In Britain, offshore stations that broadcast pop music from ships were heard largely on transistor sets. Within a decade the devices were being made in millions.

Elsewhere the device had a different significance. In parts of Africa and Asia where few homes had electricity, a cheap radio running on dry cells connected isolated villages to news, weather forecasts and agricultural advice, and shortwave broadcasts in local languages reached audiences that no newspaper could. International broadcasters and development agencies distributed sets for exactly this purpose.

Competition was fierce. American firms had dominated the market for valves, but Japanese manufacturers, with lower labour costs and a determination to export, captured a growing share of world sales of portable radios during the late 1950s and 1960s. European makers, protected at first by national broadcasting monopolies that ordered sets from local factories, followed more slowly.

The transistor radio's importance in the history of technology is therefore double. It was the first mass-market product to depend on the transistor, and it showed manufacturers that the new device could be made cheaply, reliably and in enormous quantities. The same technology, miniaturised further and packed by the thousand onto a single chip, would eventually produce the pocket calculator, the personal computer and the smartphone. For anyone who used one, though, the appeal was simpler: it put broadcast sound in a private pocket, and it changed where, and with whom, people listened.

In a growing number of fields, a crop and a power station now occupy the same ground. Rows of solar panels stand two or three metres above the soil on steel posts, and beneath them vegetables, grasses or young fruit trees are grown, while sheep sometimes graze between the posts. The idea is known as agrivoltaics — a merging of agriculture and photovoltaics, the technology that converts sunlight into electricity — and its supporters argue that land used in this way can produce food and power at the same time, rather than having to choose between them.

Land is the reason the idea has become interesting. A large solar farm occupies hundreds of hectares, and in countries where farmland is scarce or expensive, covering good soil with panels is controversial. Planning authorities in several European countries have begun to restrict solar development on the most productive agricultural land, on the grounds that food production should not be sacrificed for electricity. If the same field can do both jobs, that objection loses much of its force. Solar farms also need connections to the electricity grid, which are cheapest where the network is already strong.

A photovoltaic panel is a sandwich of silicon layers. When a particle of light, a photon, strikes the silicon, it can knock an electron loose, and the resulting flow of electrons through an external circuit is a direct current. An inverter converts this into the alternating current that the grid uses. Commercial panels convert only about a fifth of the sunlight that falls on them into electricity, typically between 15 and 22 per cent, so a large share of the incoming energy becomes heat instead.

That heat is a problem, because a hot panel is an inefficient panel. For every degree Celsius above about 25 °C, the output of a typical module falls by somewhere between 0.3 and 0.5 per cent, so a panel baking at 60 °C in still air may lose more than a tenth of its capacity. Plants, by contrast, cool themselves: water evaporating from their leaves carries heat away, in the same way that sweating cools a person. A crop growing under the panels therefore lowers the temperature of the air around them, and in dry, hot regions the gain in the panels' output can be worthwhile.

The benefit runs in the other direction too. Shade reduces the evaporation of water from the soil, so a field of panels above a crop loses less moisture than an open field does. Trials in hot, dry places have found that lettuce, spinach, tomatoes and peppers grown under panels often need less irrigation and can produce comparable or even larger harvests than crops in full sun, because the plants suffer less heat stress in the middle of the day. In parts of the American southwest and in the Sahel, where water is the limiting factor rather than light, that trade can matter more than the loss of sunshine.

Not every crop can tolerate shade, however. Wheat, maize and rice, which evolved in open landscapes and are grown at high densities, generally yield less beneath panels, and for these crops the arrangement makes little sense unless the electricity is unusually valuable. The structures themselves are a further obstacle: panels mounted three metres up need stronger frames and deeper foundations than ground-mounted arrays, and they must be spaced widely enough for tractors to pass, which reduces the density of power generation.

For livestock farmers the calculation is simpler. Sheep are cheap lawnmowers: they keep the grass short around the posts, removing the need for mowing or chemical weed control, and they gain from the shade in summer. Solar grazing, as the practice is called, has become common in Britain, where sheep graze beneath panels on many farms, and in parts of Australia and the United States. Goats and poultry are used in some projects, though goats will eat almost anything, including the insulation on cables, which has to be protected.

Japan was an early pioneer. With limited flat land and high prices for agricultural produce, Japanese farmers began erecting panels over small plots in the early 2000s, often with narrow arrays and hand tools rather than tractors; the approach is sometimes called solar sharing. Germany, where subsidies have encouraged the technology, has since published a technical standard setting out what may count as genuine agrivoltaics, largely to stop developers from claiming the label for conventional solar farms that happen to stand on fields.

Economists see a straightforward attraction for the farmer, who receives rent or a share of the electricity revenue while continuing to farm. The obstacle is capital: an elevated array costs appreciably more to build than a conventional one, and only where land is expensive or grid connections are cheap does the arithmetic work. Researchers are now testing semi-transparent panels, which let a controlled amount of light through, and vertical arrays arranged in rows, which allow machines to pass between them and generate most power in the morning and evening, when electricity prices are higher.

None of this means that agrivoltaics will spread everywhere. The science is still young, and results from one region transfer poorly to another: a variety that thrives in the heat of Arizona may fail in Denmark, and the shade that protects a lettuce seedling in July may slow its growth in a dull September. What the trials do suggest is that the question is not whether crops and panels can share a field, but which combinations of crop and panel suit a particular climate, soil and market.

Suppose a tenant moves out of a flat and the landlord keeps the deposit, claiming that a wall was damaged. The sum in dispute is perhaps £600. Taking the matter to court would mean fees, forms and a hearing on a weekday, and legal advice would cost more than the money at stake. Most claims of this size are simply abandoned, and rarely because they are weak; the system is too slow, too costly and too confusing to be worth using.

Into this gap have stepped machines. Online dispute resolution is not a novelty: the auction site eBay has for years settled disagreements between buyers and sellers through automated processes that require no courtroom, and the volume of such cases dwarfs anything a court could handle. British Columbia's Civil Resolution Tribunal, which opened in 2016 and handles small claims of up to a few thousand dollars, guides a claimant through a series of questions, offers free legal information, encourages negotiation and only then passes the file to a human adjudicator. England and Wales have piloted an online service for money claims of up to £10,000.

The arguments for going further, and letting software decide rather than merely prepare the case, are not foolish. A machine can read a claim at three in the morning, ask for the documents that are missing, translate them, and set out the strengths and weaknesses of each side in plain language. It costs almost nothing per case, so the £600 dispute becomes worth pursuing. It is also consistent: two similar claims will be treated alike, whereas two judges may reach opposite conclusions on identical facts, a variability that looks to the losing party like luck.

Yet small claims are rarely about the law. They are about what happened. Did the tenant's chair mark the wall, or was the mark already there? Was the roof repaired properly, or does the damp come from somewhere else? Such questions turn on credibility: on who is telling the truth, who has kept the receipts, and whose account is consistent with the photographs. Judging that requires listening to people, weighing hesitation as well as sentences, and recognising when a party is frightened or ashamed. Software can compare documents, but it cannot see a witness's face.

Then there is the question of how such a system is built. A model trained on past decisions will reproduce the patterns in those decisions, including any unfairness they contained, and it will do so with an air of mathematical neutrality. Nor can most such systems explain themselves: a party told that the claim has failed, with no reasons beyond a percentage, has been treated unfairly even if the outcome happens to be correct. European data protection law recognises this, giving individuals the right not to be subject to a decision based solely on automated processing when it has a legal effect on them.

The greater danger, though, is not error but inequality. If businesses and wealthy individuals continue to use courts, with their hearings, advocacy and judges, while everyone else is processed by software, we shall have created two kinds of justice: one slow, expensive and human, the other quick, cheap and mechanical. In the long run that would undermine the legitimacy of the legal system itself, because a decision that nobody has explained, reached in a process nobody can attend, is unlikely to be accepted as just by the person who loses.

Cost is the counter-argument. A hearing with a judge, a clerk and a courtroom costs far more than most small claims are worth, and that is precisely why so few of them are heard. If insisting on human adjudication means that the £600 claim is heard by nobody, then the principled defence of human judgement leaves the claimant with nothing. Any serious proposal must therefore be judged by what it delivers in practice, not by how comfortable it makes lawyers feel.

What, then, should be done? The sensible division of labour is to let software do the work it does well and keep judgement with people. Machines can triage claims, tell a claimant whether the law is on their side, fill in forms, translate documents and propose a settlement. Many disputes would end there, settled voluntarily by the parties themselves, which is generally the best outcome available. What a machine should not do is impose a binding decision on a person who disputes the facts.

Where an adjudicator does decide, the process should carry the safeguards that any court would recognise: a hearing, even a short one by video link; the right to see the evidence on which the decision rests; reasons that address the parties' arguments; and an appeal to a person. Human involvement need not be elaborate. A judge who spends ten minutes reviewing a file that software has summarised is still in a very different position from a claimant who has been outargued by an algorithm.

The promise of cheap, fast justice for small claims is therefore real, and it would be a mistake to reject it out of suspicion of novelty. But justice is not merely the efficient allocation of money between two parties; it is a public act that carries the authority of the community behind it. Software may serve that authority. It cannot supply it.