25. This regrettably lengthy prologue enables me, I hope fairly, to review Mr Owen’s main submissions summarised in para 7 above more briefly than would otherwise have been possible.

26. I do not accept that, because the tribunal lacked the power to secure compliance with its conditions, it lacked the coercive power which is one of the essential attributes of a court. What Article 5(1)(e) and (4) require is that a person of unsound mind compulsorily detained in hospital should have access to a court with power to decide whether the detention is lawful and, if not, to order his release. This power the tribunal had. Nothing in article 5 suggests that discharge subject to conditions is impermissible in principle, and nothing in the Convention jurisprudence suggests that the power to discharge conditionally (whether there are specific conditions or a mere liability to recall), properly used, should be viewed with disfavour. Indeed, the conditional discharge regime, properly used, is of great benefit to patients and the public, and conducive to the Convention object of restricting the curtailment of personal liberty to the maximum, because it enables tribunals to ensure that restricted patients compulsorily detained in hospital represent the hard core of those who suffer from mental illness, are a risk to themselves or others and cannot be effectively treated and supervised otherwise than in hospital. If there is any possibility of treating and supervising a patient in the community, the imposition of conditions permits that possibility to be explored and, it may be, tried.

27. When, following the tribunal’s order of 3 February 2000, it proved impossible to secure compliance with the conditions within a matter of a few months, a violation of the appellant’s Article 5(4) right did occur. It occurred because the tribunal, having made its order, was precluded by the authority of the Oxford case from reconsidering it. The result was to leave the appellant in limbo for a much longer period than was acceptable or compatible with the Convention. I would accordingly endorse the Court of Appeal’s decision to set aside the Oxford ruling and I would adopt the ruling it gave in para 71 of its judgment quoted above. Evidence before the House shows that that ruling is already yielding significant practical benefits. Mr Owen was, I think, right to submit that the tribunal could have achieved the same result, consistently with the Oxford case, by a judicious use of its power to adjourn and by proleptic indication of the conditions it had in mind to impose, but it is undesirable to restrict the procedural freedom of tribunals in a field as important and sensitive as this, where personal liberty and safety and public protection are all at stake: the outcome should not turn on procedural niceties.

28. There was no time between 3 February 2000 and 25 March 2002 when the appellant was, in my opinion, unlawfully detained, and there was thus no breach of Article 5(1)(e). There is a categorical difference, not a difference of degree, between this case and that of Johnson. Mr Johnson was a patient in whose case the Winterwerp criteria were found not to be satisfied from June 1989 onwards. While, therefore, it was reasonable to try and ease the patient’s reintegration into the community by the imposition of conditions, the alternative, if those conditions proved impossible to meet, was not continued detention but discharge, either absolutely or subject only to a condition of liability to recall. His detention became unlawful shortly after June 1989 because there were, as all the doctors agreed, no grounds for continuing to detain him. The present case is quite different. There was never a medical consensus, nor did the tribunal find, that the Winterwerp criteria were not satisfied. The tribunal considered that the appellant could be satisfactorily treated and supervised in the community if its conditions were met, as it expected, but the alternative, if these conditions proved impossible to meet, was not discharge, either absolutely or subject only to a condition of recall, but continued detention. The appellant was never detained when there were no grounds for detaining him. To the extent that Buxton and Sedley LJJ differed from the Master of the Rolls on this point in K, the opinion of the latter is to be preferred.

29. The duty of the health authority, whether under section 117 of the 1983 Act or in response to the tribunal’s order of 3 February 2000, was to use its best endeavours to procure compliance with the conditions laid down by the tribunal. This it did. It was not subject to an absolute obligation to procure compliance and was not at fault in failing to do so. It had no power to require any psychiatrist to act in a way which conflicted with the conscientious professional judgment of that psychiatrist. Thus the appellant can base no claim on the fact that the tribunal’s conditions were not met. This conclusion makes it unnecessary, in my opinion, to address a question on which the House heard argument, but which was not considered below, whether in a context such as this psychiatrists were or could be a hybrid public authority. Determination of that question is best left to a case in which it is necessary to the decision. We are nonetheless grateful to the Royal College of Psychiatrists for its submissions on this point.

30. I do not consider that the violation of article 5(4) which I have found calls for an award of compensation since (a) the violation has been publicly acknowledged and the appellant’s right thereby vindicated, (b) the law has been amended in a way which should prevent similar violations in future, and (c) the appellant has not been the victim of unlawful detention, which Article 5 is intended to avoid.