ABSTRACT

Can we be sure that children’s talk at the computer will be educationally useful? Can we say what ‘educationally useful’ is — what sort of talk we expect ? And if so, can we encourage the most useful types of talk?