Lantern: IELTS Academic Reading practice test

For most of human history, city streets fell dark at sunset, and anyone going out at night carried a light of their own. The first attempts at public lighting placed the burden on householders: early in the fifteenth century, London’s residents were ordered to hang a lantern outside their doors on winter evenings. The results were feeble: a candle behind a pane of horn or glass lit little more than the doorway beside it, and the rules were widely ignored.

In the second half of the seventeenth century, Paris began hanging candle lanterns above the middle of its streets, paid for by a special tax, and during the eighteenth century candles gave way to oil lamps, which burned longer and more steadily. In the 1780s the Swiss inventor Aimé Argand produced a burner with a hollow, circular wick that drew air up through its centre and gave a far brighter flame. Even so, an oil-lit street was dim by modern standards, and the lamps needed constant attention: their wicks had to be trimmed, and their oil, often obtained from whales, was expensive.

The real transformation began with gas. In the 1790s the Scottish engineer William Murdoch lit his own house with gas produced by heating coal, and in 1807 Pall Mall in London became one of the first streets anywhere to be lit by gas. Within decades, gas pipes ran beneath the streets of cities across Europe and North America. Gas gave a brighter light than oil and, once the pipes were laid, it was cheaper to run. And because the fuel arrived by pipe, nobody had to refill each lamp by hand.

Lighting the lamps, however, still required people. Every evening a lamplighter walked a set round with a long pole, used to turn on the gas and light it, and returned at dawn to put each lamp out; in a large city, lamplighters formed a small army until clockwork timers slowly took over their work in the twentieth century. Gas lighting also became far brighter in the 1880s and 1890s, when the Austrian chemist Carl Auer von Welsbach developed the gas mantle, a delicate mesh that glowed white when heated by the flame. The mantle extended the life of gas lighting at the very moment that it faced a serious rival.

That rival was electricity. In an arc lamp, a brilliant light is produced by a current leaping across a gap between two carbon rods. Arc lamps appeared on some Paris streets in the late 1870s. The light was so intense that several American cities mounted arc lamps on iron towers tall enough to illuminate many blocks at once; some of these ‘moonlight towers’ still stand in Austin, Texas. But the glare that suited broad avenues was harsh in narrow streets, and the carbon rods burned away and needed frequent replacement.

The incandescent bulb, developed independently by Joseph Swan in England and Thomas Edison in the United States around 1880, gave a softer light suited to ordinary streets. Early bulbs used fragile carbon filaments, but once filaments of tungsten were introduced early in the twentieth century, electric lamps became efficient and reliable enough to replace gas, and over the following decades most cities made the change.

From the 1930s, a new kind of lamp began to change the colour of the night. The low-pressure sodium lamp is remarkably efficient, producing far more light for each unit of electricity than a filament bulb. Its light, however, is an intense orange-yellow under which colours are almost impossible to tell apart: a red car and a blue one look much the same. High-pressure sodium lamps, common from the 1970s, gave a pinkish-gold light in which colours were easier to see, and together the two types gave the world’s cities their familiar orange glow.

Since around 2010, that glow has been fading. Light-emitting diodes, or LEDs, have replaced sodium lamps in city after city, and the arguments for them are strong. They often use around half the electricity of the lamps they replace, can run for many years without attention, and can be dimmed late at night or aimed precisely at the road instead of spilling light into bedroom windows. Their white light also makes colours look natural.

Yet the change has raised new concerns. Many early LED street lamps gave a cold, bluish-white light, and blue light is scattered more readily by the atmosphere, adding to the ‘skyglow’ that hides the stars from most city dwellers. Blue-rich light also disturbs the body’s internal clock and confuses wildlife, from insects drawn to lamps in huge numbers to newly hatched sea turtles that crawl towards lit roads instead of the sea. Many cities now choose warmer LEDs as a compromise. Even the claim that lighting makes people safer is less straightforward than it seems. ‘People have always felt safer in lit streets,’ says Dr Helen Carrow, a historian of urban technology, ‘but the evidence that lighting reduces crime is surprisingly mixed.’ Some studies have recorded falls in crime after lighting was improved; others have found no effect, or suggest that the benefit came from the attention a neighbourhood received rather than from the lamps themselves.

For much of the twentieth century, bilingualism was regarded with suspicion. Children who grew up with two languages were thought to be at risk of confusion, and early intelligence tests seemed to confirm that they lagged behind their monolingual classmates. We can now see that those tests measured little more than a child’s command of the language they happened to be given in, which was frequently the weaker of the two. By the 1960s the tide had begun to turn, and by the end of the century something close to the opposite view had taken hold. Bilinguals, it was argued, were not merely equal to monolinguals but in one important respect better equipped. Because they must constantly suppress whichever language they are not using, they receive a lifelong workout in what psychologists call executive function: the set of mental processes that allow us to focus attention, ignore distractions and switch between tasks.

The early evidence for this ‘bilingual advantage’ looked compelling. In a series of experiments in the 2000s, the psychologist Margaret Ellery and her colleagues found that bilingual children and adults responded faster than monolinguals on tasks that required them to ignore misleading information, such as pressing a left-hand button when an arrow pointed left but appeared on the right of the screen. Later work from the same laboratory suggested that among patients with dementia, those who had used two languages throughout their lives had developed symptoms several years later than those who had used one. It is difficult to overstate how appealing these findings were. They offered a rare piece of good news about a common experience, and they were taken up with enthusiasm by newspapers, by parents and by anyone who wished to promote language learning in schools.

Then the results began to falter. When other laboratories repeated the experiments, many found no difference at all between bilinguals and monolinguals, and the larger and more carefully designed the study, the less likely it was to find one. A team led by Tomas Veldhuis then uncovered a more uncomfortable pattern. They traced studies of the bilingual advantage that had been presented at scientific conferences and checked which were later published in journals. Of those that had reported an advantage, most reached print; of those that had found none, or found mixed results, only a small minority did. The published record, in other words, had been filtered. Readers of journals saw the successes, the failures stayed in filing cabinets, and a small or perhaps non-existent effect was made to look robust.

Why had the early studies found an effect at all? Part of the answer may lie in whom they compared. Bilinguals and monolinguals are rarely alike in every respect except language. In many of the original studies, the bilingual participants were immigrants or the children of immigrants and the monolinguals were not, yet migration is itself selective: people who move to another country tend, on average, to be healthier and more determined than those who stay behind, and are sometimes better educated. Socioeconomic status, which is closely linked to performance on executive-function tasks, was often measured crudely or not at all. Even culture may play a part, since communities differ in how strongly children are encouraged to respond quickly when a task is timed. Researchers can try to match their groups statistically, but they can only ever adjust for the differences they have thought to measure, and no study can measure everything.

None of this proves that bilingualism has no effect on the mind; the absence of evidence is not evidence of absence. Defenders of the hypothesis, among them the linguist Rafael Ostrowski, argue that the standard tasks were simply too blunt, and that the benefits of managing two languages may appear only in particular situations, or among people who switch between their languages many times a day, rather than as a general advantage that can be detected with a stopwatch. That may be true, and I would not rule it out. But a claim that can survive only by becoming steadily narrower is a claim in retreat, and after two decades and hundreds of studies, the most reasonable conclusion is that any advantage in executive function is small, inconsistent and a long way from the transformation of the mind that was once promised. The dementia findings, meanwhile, have fared little better in studies that follow people forward in time rather than looking back through their medical records.

It would be a mistake, however, to conclude that the case for bilingualism has collapsed along with it. That case never depended on reaction times. A second language opens a door to other people, other literatures and other ways of seeing the world, and for millions of children it is the language of their grandparents, of a community and of a family history that would otherwise be lost. It brings practical benefits too, in employment and travel, that nobody disputes. If anything, the episode should worry us for a different reason. By tying the value of bilingualism to a laboratory effect, its advocates struck an unwise bargain, making that value hostage to the effect’s survival. Parents who now hear that the ‘bilingual brain boost’ was a myth may conclude, wrongly, that there is nothing to be gained by passing on a language at home.

The story also carries a lesson about science itself. The bilingual advantage was not a fraud; as far as anyone can tell, the researchers who reported it were doing honest work with the methods of their time. What failed was a system that rewarded striking results, published them readily and left the dull ones unseen. That system is changing. A growing number of journals now agree to publish a study before its results are known, on the strength of its design, and large projects that pool data from many laboratories are becoming common. Such reforms will not, in themselves, tell us whether speaking two languages sharpens the mind. They will, however, make it much harder to believe that it does on the strength of evidence that is weaker than it looks.