in the late 1980s, two American psychologists, Peter Salovey of Yale and John Mayer of the University of New Hampshire, were casting around for a pithy way to sum up human qualities such as empathy, self-awareness and emotional control. For a while, the phrase they hit on -- "emotional intelligence" -- languished in academic obscurity. Then Daniel Goleman, a writer with The New York Times, picked it up and nailed it to the mast of his best-seller Emotional Intelligence: why it can matter more than IQ.

Now the phrase is popping up everywhere. On the Oprah Winfrey show. In magazines that challenge you to "know your own emotional IQ". On Internet sites which offer to test emotional intelligence or "EQ" with scenarios such as: "You're on a plane that suddenly hits bad turbulence. Do you (a) continue to watch the movie (b) go on the alert for an emergency (c) do a little of a and b (d) not sure -- never noticed?" *Score 200 points and you're an emotional "genius", 25 and you're a "Neanderthal" in need of psychotherapy. Emotional intelligence may have started out as an academic catch phrase, but it is fast becoming the psychological mantra of the mid-1990s.

And mostly for some very obvious reasons. Blaming today's epidemics of violent crime, marital strife and teenage drug abuse on poor morals and a decline in national character sounds blimpish and defeatist. Blaming these ills on deficiencies in EQ, on the other hand, doesn't sound so bad. After all, it might be possible to improve levels of emotional intelligence in the young to equip them for life's trials. And there is much about emotional intelligence that sounds like plain common sense. Isn't it obvious that the ability to, say, control rage or develop empathy is likely to make a better indicator of future success than the kind of abstract intelligence measured by IQ?

But even if it is, there are two big questions. Can emotional intelligence really be measured in a meaningful way? And if so, can young children found wanting in it be taught the necessary skills? Can EQ become a tool for fixing failing education systems?

Crusaders for emotional intelligence seem unreservedly optimistic. Not because they believe they have the tools today to reduce the concept of EQ to a numerical yardstick in the manner of IQ (Internet questionnaires are for fun, not science), but because of the remarkable predictive power they claim for something called the marshmallow test.

Back in the 1960s Walter Mischel, a psychologist at Stanford University, gave marshmallows to groups of four-year-olds and then left the room, promising that any child who could postpone eating the marshmallow until he came back, some 15 to 20 minutes later, would be rewarded with a second marshmallow. Years later, Mischel discovered that the kids who triumphed over their desire had grown into teenagers who were socially, emotionally and academically more competent than the four-year-olds who ate the marshmallow at once. Self-control in the face of a marshmallow at four was shown to be "twice as powerful a predictor of later academic prowess as IQ", says Goleman in his book.

Goleman sees the ability to delay gratification as a master skill, a triumph of the reasoning brain over the impulsive one. But does it really provide a fundamental measure of emotional IQ? Unfortunately, the marshmallow test turns out to conceal some very complex mental behaviour.

Mischel went on to discover that the successful children were able to think of something else. Some would sing, tap their feet, tell themselves stories, imagine the marshmallow was a fluffy cloud--anything to avoid eating it. One held out by falling asleep. So putting off rewards is not a single skill that educationalists can easily pinpoint and work on, but depends on complex cognitive abilities. Not only that, but what is true of the ability to delay gratification may be even more true of other, subtler components of emotional intelligence, such as self-awareness, impulse control, self-motivation and empathy. Can we measure these?

Some psychologists have serious doubts. "The idea that you can measure emotional intelligence like IQ is very misleading," says Ross Buck, professor of communication sciences at the University of Connecticut. Emotional skills are slippery and relative in a way that IQ isn't, he concludes. "Your communicative ability with someone you know is different from your communicative ability with a stranger, and each relationship will have its own characteristic emotional communication."

Paul Harris, lecturer in experimental psychology at the University of Oxford and author of Children and Emotion, agrees. If you try to measure empathy, he says, your measurement will depend on who the child is being empathetic towards. In other words, every emotional response is embedded in its social context.

Behavioural research, aided by the advent of the video camera, shows how early in life this begins. Studies by Vasudevi Reddy, a psychologist at the University of Portsmouth, show how expressions of shyness in very young babies depend critically on social context. Young babies will sometimes turn their heads, avoid gaze or raise their arms to hide their faces--but precisely when and how often they do so depends on who they are with.

But even if we could measure such emotions in infants and the young, teaching their control to children, as Goleman suggests, would be no small task. Individual differences, coupled with the way emotions depend on social context, make it hard to imagine a training programme that could fit the needs of different children. "If someone has a tendency to be aggressive, you can train them to recognise and control their feelings," says Buck. "But training an extrovert will be different from training an introvert. And it's not the same as saying this person is or is not empathic."

And there is another stumbling block for would-be teachers of empathy or self-awareness. Talk to your average five-year-old about empathy and you won't get much response because expressing an emotion is not the same as understanding it. Harris, for example, concludes that four and five-year-olds have yet to discover that emotional lives are strongly influenced by a knowledge of other people's feelings. At this age, he says, children believe that happiness and sadness depend simply on whether or not people get what they themselves want. Only later does their conscious emotional universe expand to include such notions as pride, guilt and shame.

Despite these complications, some psychologists and educators in the US are having a determined crack at improving "emotional literacy" with the aid of specially designed teaching programmes. At New Haven, for example, the Augusta Lewis Troup Middle School provides lessons in impulse control in which children are taught to think of traffic lights. About to hit out in anger? "See" the red light, stop, calm down. Amber light means think through the problem. Green light, a positive, nonaggressive solution.

Goleman's book abounds with enthusiasm for such programmes. Yet their value to the majority of children remains unproven. According to Harris, the emotional analysis and training of children is certainly effective at the extremes of emotional illiteracy, for instance with children who've been severely maltreated or have specific needs. But Harris doubts such programmes could be extended effectively to the general population.

Even Goleman quotes the Nicomachean Ethics, in which Aristotle says: "Anyone can be angry--that is easy. But to be angry with the right person, to the right degree, at the right time, for the right purpose, and in the right way -- that is not easy." The concept of emotional intelligence may have helped people realise that the emotional skills are important to intellectual achievement, but we may not be much nearer understanding how to measure or develop it than Aristotle was more than 2000 years ago.

Research by Karen Gold, David Concar.