The problem is that infants learn to rely on their parents to decide when they have had enough to eat, rather than following their own internal cues.
"We know that mothers who follow baby-led weaning tend to be more educated, less likely to return to work before 12 months, and less likely to use feeding practices that have been associated with overweight in children."
Professor Taylor said those practices included parental pressure to eat and rewarding infants with treat foods if they ate their vegetables.
But there were potential problems with baby-led weaning, such as iron deficiency, reduction in growth rates and choking.
Professor Taylor and her colleagues have started recruiting 300 pregnant women, half of whom will be trained to use a form of baby-led weaning that has been modified to address the perceived risks.
The rest will use the standard weaning methods.
The ability of the children in each group to regulate their own energy intake will be compared at two years. They will be offered a buffet of food on two separate occasions, preceded on one day by a high-energy first course and on the other by a low-energy first course.
It is expected this will show if the experimental method makes a difference.Martin Johnston