IELTS Reading: Multiple choice

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.