First: Lift the veil of secrecy. There is no area of government, or indeed of professional activity of any kind, which functions better when officials know that what they do will escape public inspection. It is inconceivable that the quality of decision-making too often characteristic of the family courts would be tolerated in the criminal courts. The only explanation for the difference is that the criminal courts are public: incompetence and errors are made public immediately. In the family courts, it is almost impossible even to identify, let alone to rectify, poor or incompetent decision-making. Of course children need to be protected from publicity. But it is not impossible to devise a system which combines publicity for the procedures and the bare facts of the case with anonymity for the children involved. It works in rape trials, where the identity of the victim is kept secret even though the facts of the case are made public.

Second: Produce a clear definition of critical notions such as “the best interests of the child” and “harm to the child”. The ambiguity and vagueness of those notions has led to their being deprived of all concrete meaning, which has in turn led to a situation in which it is impossible for the courts to assess claims that “a child will be harmed” if allowed to remain with his parents. Moreover, social services, and courts, rarely seem to recognise that placing a child in local authority care is usually in itself extremely harmful. That harm is consequently never weighed against whatever harm is alleged to result from allowing the child to remain with his or her parents. The assumption is that local authority care will be better than parental care: in fact, that is almost never true. Parental care has to be abysmally awful for local authority care to be a less harmful alternative.

Third: Induce a sense of scepticism in judges as to the reliability of claims made by social services and the experts they produce. Mr Justice McFarlane manifested an admirable and wholly appropriate scepticism in his review of the EPO in the case discussed above. But as that case reveals, many of those on the bench are unwarrantedly credulous of social services’ claims, rather than healthily sceptical. They are too willing simply to endorse, rather than to scrutinise, them.

None of these initial steps would be difficult to implement. The depressing thing is that all of them have been recommended in the past — and the Government has never acted on them. What will it take to produce the recognition that reform is essential? The possibility that thousands of children are taken wrongly from their parents every year is clearly not enough. I wonder what is.