In the matter of B (a Child) (FC)

This challenge to the making of a care order, made with a view to the child’s adoption, requires the court to consider

The mother, “M”, supported by the father, “F”, with whom she lives, appeals against an order of the Court of Appeal (Rix, Black and Lewison LJJ) dated 14 November 2012. By its order, the Court of Appeal dismissed M’s appeal against an order made by HHJ Cryan in the Principal Registry of the Family Division dated 14 June 2012. By his order, the judge made a care order in relation to the daughter of M and F upon the basis of a care plan that she should be placed for adoption. Amelia (being the name which Lady Hale proposes that we should attribute to the daughter) was born in April 2010 so is now aged three. At birth she was placed with a foster mother on what was intended to be a short- term basis; she remains living with her pending the determination of this appeal.

M is aged 42 and F is aged 45. The applicants for the care order are the London Borough of Barnet, where the parents were living when Amelia was born. Since then they have moved to Islington. M and F each have other children. M has another daughter, aged 14, whom, in accordance with Lady Hale’s proposal, I will call Teresa and with whom M has no contact. F has four other daughters aged between 17 and six. They live with their mother; F has an amiable relationship with them but, for reasons which I will explain in para 15 below, his has not been a stable or responsible presence in their lives and, on a practical level, his involvement with them has been marginal.

Teresa’s father is M’s step-father, “Mr E”, with whom M lived for many years. In 2010, following her separation from him, M applied for a residence order in relation to Teresa. It was in that application that Judge Cryan first became acquainted with the family. He conducted a fact-finding hearing over 20 days which led in April 2011 to his handing down a judgment of 180 paragraphs. The relevant local authority, West Sussex, thereupon issued care proceedings in relation to Teresa, in which the judge conducted three further substantial hearings. With M’s support, he ordered the removal of Teresa from the home of Mr E into foster care, where she remains.

The hearing which led to the making of the care order in relation to Amelia also proceeded for 20 days and led to the judgment under challenge, which extends to 203 paragraphs. It follows that, in arriving at his conclusion that it was not safe for Amelia to be placed with the parents and that it was necessary in her interests that she should be adopted, Judge Cryan drew on extensive exposure to the problems of the family. Indeed the extent of it is beyond my own experience of service for 12 years in the Family Division. M scarcely challenged the judge’s findings of fact in the Court of Appeal; and her challenge to them in this court is, inevitably, so faint that there is no need to add to the reasons which Black LJ gave for rejecting them in her judgment in the Court of Appeal, [2012] EWCA Civ 1475, at paras 133 to 136.

What follows represents as brief a summary as possible of the facts found by Judge Cryan. Greater detail is to be found in the judgment of Black LJ.

M is the victim of grave misfortune. Her life has been hugely dysfunctional. In 1975, when she was aged four, the marriage of her parents broke down and, with her sisters, she moved to live with her grandparents. Several years after her mother’s marriage to Mr E in 1977, M and her sisters moved to live with them. The family was enlarged by the birth of two sons born to M’s mother by Mr E in 1985 and 1986.

Mr E’s influence on the family in general, and on M in particular, was malign in almost every sense. He is egocentric; aggressive; domineering and dishonest. By 1986, although married to her mother, Mr E was having sexual relations with M, then aged 15; in that year she became pregnant by him and had an abortion. Prior to 1999, when she gave birth to Teresa, M was to have six further abortions consequent upon her relationship with him.

Mr E also inflicted grave and protracted physical abuse on one of M’s sisters. When in 1989 the family went to live in Greece, of which Mr E was a citizen, they left the sister behind. In due course the sister was taken permanently into care. Meanwhile the family had returned from Greece.

In 1990 M’s mother left Mr E. For the following 19 years the family in effect comprised Mr E, M, her two half-brothers and, once born, Teresa. On numerous occasions until 2002, when they settled in West Sussex, the family moved home. In the early years, when the half-brothers were still minors and prior to the birth of Teresa, local authorities and police forces became concerned about their safety at the hands of Mr E; and for a month in 1997 they were taken into care.

In 2003 M was found guilty of a series of frauds, which had yielded £30,000 and for which she was sentenced to imprisonment for two years. Her defence had been that the prosecuting officer had conspired with a man who had allegedly raped one of her half-brothers to present a false case against her. In this regard she was later found guilty of attempting to pervert the course of justice and sentenced to a further term of 27 months. Judge Cryan found that, in perpetrating the frauds and concocting the false defence, she had been heavily influenced by Mr E. Nevertheless the convictions, the gravity of which was reflected in the sentences passed upon her in respect of them, represent the first example of conduct on the part of M which, taken together, was to lead the judge to describe her as an habitual and purposeful liar.

For the purposes of her second criminal trial M’s solicitors invited a consultant psychiatrist, Dr Taylor, to assess her fitness to plead. His conclusion was that she was fit to do so; but, following examination of her extensive medical records, he concluded that she suffered from a somatisation disorder. This is a chronic psychiatric disorder, of which the main features are multiple complaints about physical symptoms and requests for medical investigations in circumstances in which, if organic disorder is present at all, it fails to account for the symptoms or for the extent of the sufferer’s pre-occupations. In short it is a condition which drives the sufferer to misuse physical symptoms in order to elicit care from others or for other purposes. For use in the proceedings before Judge Cryan, another consultant psychiatrist, Dr Bass, was instructed to appraise M’s psychological condition in the light of her more recent medical records. Dr Bass, who has considerable expertise in this area, confirmed Dr Taylor’s diagnosis that M suffers a somatisation disorder; and in effect it became an agreed fact. But, as I will explain in para 17 below, Dr Bass went further.

While investigating the allegations of fraud against M, the police discovered paedophilic images in a computer in the family home. Mr E contended that the police had planted them there. In 2004 Mr E complained to the General Medical Council that the family GP had sexually assaulted one of the half-brothers, then aged 18. The complaint was dismissed. Following her ultimate release from prison in 2004, M made various complaints to probation officers, hospitals and social workers that Mr E was abusing her physically and sexually. By June 2009 her life in the home had become intolerable and she left. She did not take Teresa with her; no doubt Mr E would not have allowed her to do so. In effect it marked the end of her relationship with Teresa who, under Mr E’s influence, has refused to have any further dealings with her. Judge Cryan’s conclusion was that, within her long relationship with Mr E, M could fairly be regarded as his victim but that her role had not been entirely inert and that she had actively conspired with him in the many lies, deceptions and false accusations which had been generated in the household.

In the summer of 2009 M met F. They began a relationship which continued following M’s move to Barnet late in 2009; but they did not fully cohabit until late in 2011.

F has a long criminal history and has spent about 15 years of his adult life in prison. He was convicted of 52 offences between 1980 and 2008. Some related to drugs; some included violence but most were offences of dishonesty. In the 1990s he sustained three convictions for robbery, for each of which he received sentences of imprisonment of between two and three and a half years. In 2001 he was sentenced to four years for burglary. In 2007 he was sentenced to three years for further burglary; and the pre-sentence report recorded his admission of a crack cocaine habit and of the occasional use of heroin. But, following his release on licence in May 2009, F has sustained no further convictions. For the first year following his release he was subjected to regular drug tests, of which two proved positive. At a hearing in June 2010 into whether an interim care order in relation to Amelia should be continued, F refused a judge’s invitation to submit strands of his hair to drug-testing; and he later announced that Barnet could “kiss [his] arse” when they next wanted to test him. Ultimately, in July 2011, he underwent a test which showed use of cannabis but not of Class A drugs; and there was no evidence before Judge Cryan that, although admitting to the continued use of cannabis, F was then also using Class A drugs.

I turn to an important part of the history which, it is clear, made Judge Cryan’s decision particularly difficult. It relates to the amount and quality of the contact which M and F have had with Amelia following her removal, at birth, into care. From May 2010 until the judge’s decision in June 2012, contact took place under supervision for one and a half hours on five days a week; since the decision it has taken place for the same duration on three days a week. M and F have been assiduous in attending, in effect, all the periods of contact; and the supervisors’ reports of its quality have been uniformly positive. The judge found that:

It was the diagnosis of Dr Bass, which Judge Cryan accepted, that, beyond abnormal personality traits and in addition to, and more significantly than, her somatisation disorder, M suffers a factitious disorder of mild to moderate intensity. This is a related psychiatric disorder in which the sufferer is driven repeatedly to exaggerate symptoms or altogether to fabricate them and to offer false histories. There is therefore a deceptive dimension to the disorder which was replicated in a mass of other evidence before the judge, unrelated to M’s medical condition, which raised questions about her ability, and for that matter also the ability of F, to behave honestly with professionals. Dr Bass stressed that M’s psychiatric disorders required psychotherapy which might last for a year and which could be undertaken only if she were to acknowledge the problems and to engage honestly with the therapist.

There was debate before Judge Cryan about the effect on Amelia of being placed in the care of M in the event that she was to continue to suffer somatisation and factitious disorders. In the event the judge found that there were risks that she would suffer harm in that regard. They were that M might present Amelia for medical treatment, and, worse, that she might receive medical treatment which was unnecessary; and that in any event Amelia might grow up to believe that the way in which M presented herself for treatment was appropriate and might model herself on it.

But Judge Cryan’s concerns went far wider than that. It was his wider concerns which drove him to the key conclusion that it was not safe for Amelia to be placed with the parents. The wider concerns related to the mass of evidence that each of the parents was fundamentally dishonest, manipulative and antagonistic towards professionals. The expert evidence before the judge as to whether Amelia could safely be placed with the parents was in conflict: Ms Summer, of Marlborough Family Service, and Dr Bass both agreed with Barnet that she could not safely be so placed. Dr Dale, who has a background in social work and is not a medical doctor, suggested otherwise; and the Children’s Guardian, whose contribution, constrained by lack of resources at CAFCASS, was described by the judge as superficial and who sadly died within weeks of the hand-down of his judgment, agreed with Dr Dale. But, importantly if inevitably, all the experts agreed that Amelia could not conceivably be placed in the care of the parents other than pursuant to a programme of multi-disciplinary monitoring and support, which could be implemented only in the event of honest cooperation on the part of the parents. It was, in particular, their cooperation with the local authority which was described by Dr Taylor as very important and by Dr Bass as critical. The judge’s key conclusion was that their honest co-operation with professionals would not be forthcoming; and his subsidiary conclusion was that it might in any event be damaging for a child to grow up in a household permeated by dishonesty and animosity towards professionals in that she would find such attitudes confusing even assuming that she did not find them attractive.

It would not be usual, at this level, to descend into the mass of evidence by reference to which Judge Cryan sought to justify his key conclusion. But I consider that I need to cite 16 examples of it:

In relation to her habitual making of false complaints of a highly unpleasant character about professionals, Judge Cryan described M as an accomplished pupil of Mr E. He accepted the following evidence of Dr Bass:

“I have major concerns about the capacity of [M] to protect any child in her care because of ongoing concerns about her capacity for deception because it is such a dominant feature of her personality, allied to this lack of insight and this lack of acknowledgment and evasiveness and inconsistency.”

The judge observed:

“Provided whoever [M] is dealing with appears to be going along with her without challenge, she will cooperate to achieve her ends. If she is questioned, challenged or thwarted, cooperation is abandoned and entirely unacceptable hostility begins.”

The judge concluded:

“Ultimately, I find that I am persuaded... that what the evidence clearly demonstrates is that these parents do not have the capacity to engage with professionals in such a way that their behaviour will be either controlled or amended to bring about an environment where [Amelia] would be safe... In short I cannot see that there is any sufficiently reliable way that I can fulfil my duty to [Amelia] to protect her from harm and still place her with her parents. I appreciate that in so saying I am depriving her of a relationship which, young though she is, is important to her and depriving her and her parents of that family life which this court strives to promote.”

The judge ended by stating that in those circumstances adoption was “the only viable option” for Amelia’s future care.

Judge Cryan was well aware that, before he could even consider whether to make a care order, section 31(2) of the 1989 Act required him to be satisfied (a) that, when she was first taken into care, namely at birth, Amelia had been “likely” to suffer “significant harm” and (b) that the likelihood was “attributable” to the care likely to be given to her if the order were not made not being what it would be reasonable to expect a parent to give to her. But little separate attention was paid to these threshold requirements in submissions to the judge. Mr Feehan QC, on behalf of the mother, submitted to him that “the evidence barely crosses the threshold”. The guardian presumably considered that the threshold was crossed because at one stage she was advocating a supervision order. But counsel for the father disputed that it was crossed. In the event the judge expressed himself satisfied “that the threshold has been crossed, not perhaps in the most extreme way that is seen in some cases, but crossed it has been”. And, following brief explanation, he turned to the welfare stage of the inquiry.

But in the Court of Appeal, and in particular in this court, much greater attention has been paid to what the threshold requires. It is common ground that, as recently reaffirmed by this court in In re J (Children) (Care Proceedings: Threshold Criteria) [2013] UKSC 9, [2013] 2 WLR 649, a likelihood of significant harm means no more than a real possibility that it will occur but a conclusion to that effect must be based upon a fact or facts established on a balance of probabilities. In the context of the present case it is also noteworthy that, by section 31(9), “harm” means “ill-treatment or the impairment of health or development...” and “development” includes “emotional...development”. Beyond this, however, the debate surrounds two matters.

The first matter is the meaning of the word “significant”. In this regard Parliament chose to help the court to a limited extent by providing in section 31(10) as follows:

“Where the question of whether harm suffered by a child is significant turns on the child’s health or development, his health or development shall be compared with that which could reasonably be expected of a similar child.”

When we read this subsection together with the definition of “harm” in the preceding subsection, we conclude that, whereas the concept of “ill-treatment” is absolute, the concept of “impairment of health or development” is relative to the health or development which could reasonably be expected of a similar child. This is helpful but little more than common sense.

In my view this court should avoid attempting to explain the word “significant”. It would be a gloss; attention might then turn to the meaning of the gloss and, albeit with the best of intentions, the courts might find in due course that they had travelled far from the word itself. Nevertheless it might be worthwhile to note that in the White Paper which preceded the 1989 Act, namely The Law on Child Care and Family Services, Cm 62, January 1987, the government stated, at para 60:

“It is intended that “likely harm” should cover all cases of unacceptable risk in which it may be necessary to balance the chance of the harm occurring against the magnitude of that harm if it does occur.”

It follows that when, in Re C and B (Care Order: Future Harm) [2001] 1 FLR 611, Hale LJ (as my Lady then was) said, at para 28, that “a comparatively small risk of really serious harm can justify action, while even the virtual certainty of slight harm might not”, she was faithfully expressing the intention behind the subsection. But the other interesting feature of the sentence in the White Paper is the word “unacceptable”. I suggest that it was later realised that whether the risk was “unacceptable” was a judgement which fell to be made at the welfare stage of the inquiry; and so a different adjective was chosen.

In Re L (Children) (Care Proceedings: Significant Harm) [2006] EWCA Civ 1282, [2007] 1 FLR 1068, the Court of Appeal allowed an appeal by parents against a judge’s conclusion that their children had suffered and were likely to suffer significant harm and it remitted the issue for re-hearing. The professional evidence had been that the parents’ deficiencies had had “subtle and ambiguous consequences” for the children; and it was not difficult for me, at para 31(a) of my judgment in that court, to conclude that such consequences could not amount to significant harm. The rehearing was conducted by Hedley J and, by his judgment reported as Re L (Care: Threshold Criteria) [2007] 1 FLR 2050, he declined to hold that the threshold was crossed. He observed, at para 50, that “society must be willing to tolerate very diverse standards of parenting, including the eccentric, the barely adequate and the inconsistent”; and, at para 51, that “significant harm is fact-specific and must retain the breadth of meaning that human fallibility may require of it” but that “it is clear that it must be something unusual; at least something more than the commonplace human failure or inadequacy”.

In the present case Mr Feehan seeks to develop Hedley J’s point. He submits that:

“many parents are hypochondriacs, many parents are criminals or benefit cheats, many parents discriminate against ethnic or sexual minorities, many parents support vile political parties or belong to unusual or militant religions. All of these follies are visited upon their children, who may well adopt or ‘model’ them in their own lives but those children could not be removed for those reasons.”

I agree with Mr Feehan’s submission; but the question arises whether, in the light of the judge’s key conclusion, it misses the point.

Mr Feehan proceeds to submit that the proportionality requirement under article 8 of the Convention applies, albeit perhaps only very obliquely, to whether harm is significant for the purposes of the subsection; and he cites observations by Ward LJ in Re MA (Care Threshold) [2009] EWCA Civ 853, [2010] 1 FLR 431, at para 54, that, although it has more relevance to the welfare stage of the inquiry, article 8 “does, none the less, inform the meaning of ‘significant’ and serves to emphasise that there must be a ‘relevant and sufficient’ reason for crossing the threshold”. I consider that, with respect to him, Ward LJ there introduced an inappropriate layer of complexity to the inquiry whether harm is significant. It is the “interference” with the exercise of the right to respect for family life which article 8 addresses. No interference occurs when a judge concludes that the threshold is crossed. The interference occurs only if, at the welfare stage, the judge proceeds to make a care or supervision order; and it is that order which must therefore not fall foul of article 8. I regard section 31(2) as an admirable domestic provision which, by setting a threshold, may make it more likely, although far from inevitable, that any care or supervision order will not fall foul of article 8. But I conclude that the crossing of the threshold does not, of itself, engage the article; and I am fortified in my conclusion by the fact that Lord Neuberger, at para 62, Lady Hale, at para 186, and Lord Clarke at para 134, agree with it and that Lord Kerr, at para 129, tends to agree with it.

The second matter relates to Mr Feehan’s submission that the threshold set by section 31(2) is not crossed if the deficits relate only to the character of the parents rather than to the quality of their parenting. His alternative submission is that harm suffered or likely to be suffered by a child as a result of parental action or inaction may cross the threshold only if, in so acting or failing to act, the parent or parents were deliberately or intentionally to have caused or to be likely to cause such harm. M is, of course, not responsible for her personality traits nor for her psychiatric disorders; and in effect the submission is that the dishonesty, animosities and obstructionism of the parents represent deficits only of character and that, if and insofar as they might cause harm to Amelia, whom they love, the harm is neither deliberate nor intentional.

The first of these alternative submissions represents a false dichotomy: for the character of the parents is relevant to each stage of the inquiry whether to make a care order only to the extent that it affects the quality of their parenting. The second of them is misconceived: for there is no requisite mental element to accompany the actions or inactions which have caused, or are likely to cause, significant harm to the child. Section 31(2)(b)(i) requires only that the harm or likelihood of harm should be “attributable” to the care given or likely to be given to the child not being what it would be reasonable to expect a parent to give to him. Such is a requirement only of causation as between the care and the harm. The provision was prefigured in the White Paper, Cm 62, cited above, also at para 60:

“The court will also have to make a decision as to whether the harm was caused or will in future be caused by the child not receiving a reasonable standard of care or by the absence of adequate parental control. This is not intended to imply a judgment on the parent who may be doing his best but is still unable to provide a reasonable standard of care.”

Judge Cryan’s care order in relation to Amelia with a view to her adoption represented an interference with the exercise by Amelia, by M and by F of their rights to respect for their family life. It was therefore lawful only if, within the meaning of article 8(2) of the Convention, it was not only in accordance with the law but also “necessary” in a democratic society for the protection of the right of A to grow up free from harm. In Johansen v Norway (1996) 23 EHRR 33 the European Commission of Human Rights observed, at para 83, that “the notion of necessity implies that the interference corresponds to a pressing social need and, in particular, that it is proportional to the legitimate aim pursued”.

In a number of its judgments the European Court of Human Rights, “the ECtHR”, has spelt out the stark effects of the proportionality requirement in its application to a determination that a child should be adopted. Only a year ago, in YC v United Kingdom (2012) 55 EHRR 967, it said:

“134.

The Court reiterates that in cases concerning the placing of a child for adoption, which entails the permanent severance of family ties, the best interests of the child are paramount. In identifying the child’s best interests in a particular case, two considerations must be borne in mind: first, it is in the child’s best interests that his ties with his family be maintained except in cases where the family has proved particularly unfit; and secondly, it is in the child’s best interests to ensure his development in a safe and secure environment. It is clear from the foregoing that family ties may only be severed in very exceptional circumstances and that everything must be done to preserve personal relations and, where appropriate, to ‘rebuild’ the family. It is not enough to show that a child could be placed in a more beneficial environment for his upbringing. However, where the maintenance of family ties would harm the child’s health and development, a parent is not entitled under article 8 to insist that such ties be maintained.”

Although in that paragraph it did not in terms refer to proportionality, the court had prefaced it with a reference to the need to examine whether the reasons adduced to justify the measures were relevant and sufficient, in other words whether they were proportionate to them.

In my view it is important not to take any one particular sentence out of its context in the whole of para 134 of the YC case: for each of its propositions is interwoven with the others. But the paragraph well demonstrates the high degree of justification which article 8 demands of a determination that a child should be adopted or placed in care with a view to adoption. Yet, while in every such case the trial judge should, as Judge Cryan expressly did, consider the proportionality of adoption to the identified risks, he is likely to find that domestic law runs broadly in parallel with the demands of article 8. Thus domestic law makes clear that:

The same thread therefore runs through both domestic law and Convention law, namely that the interests of the child must render it necessary to make an adoption order. The word “requires” in section 52(1)(b) “was plainly chosen as best conveying...the essence of the Strasbourg jurisprudence” (Re P (Placement Orders: Parental Consent) [2008] EWCA Civ 535, [2008] 2 FLR 625, para 125).

What is the proper approach of an appellate court to a challenge to the proportionality of a care order made with a view to adoption? Section 6(1) of the Human Rights Act 1998 makes it unlawful for public authorities, which include appellate courts, to act in a way which is incompatible with a Convention right. So the question becomes whether, as the mother submits, section 6(1) dictates that an appellate court should depart from its normal function of secondary review and instead should make a fresh determination of its own when the allegation is that the order made below has violated a Convention right. An analogous submission was made in MT (Algeria) v Secretary of State for the Home Department [2009] UKHL 10, [2010] 2 AC 110, in which two Algerian nationals and a Jordanian national appealed to the Court of Appeal, and from there to the House of Lords, against the dismissal by the Special Immigration Appeals Commission of their appeals against orders for deportation. Their case was that deportation would infringe their rights under article 3 of the Convention; and, by reference to section 6 of the 1998 Act, they claimed to be entitled to a fresh determination of their case both in the Court of Appeal and in the House of Lords. The appellate committee rejected their claim. Lord Phillips pointed out, at paras 67 and 69, that, by section 7 of the Special Immigration Appeals Commission Act 1997, appeal lay from the Commission to the Court of Appeal only on a point of law and that, for the purposes of section 6(2)(a) of the 1998 Act, Parliament had therefore precluded the Court of Appeal from conducting any such fresh determination as the appellants sought. More broadly, however, Lord Hoffmann observed:

“190.

There is nothing in the Convention which prevents the United Kingdom from according only a limited right of appeal, even if the issue involves a Convention right. There is no Convention obligation to have a right of appeal at all. If there is a right of appeal, then of course it must offer a fair hearing before an independent and impartial tribunal in accordance with article 6. But there is no obligation to provide an appeal against the determination of a Convention right. The only concern of the European court with the court structure of the member state is that it should provide a remedy for breach of a Convention right in accordance with article 13. If a SIAC hearing does so, that is an end of the matter and the extent of the right of appeal, if any, is irrelevant.”

It is therefore clear that the Convention itself does not require appellate courts to address issues arising under it with any particular degree of intensity. Appellate courts must discharge their domestic duty under section 6(1); but the manner in which they seek to do so is a matter for Parliament or for rules made under its authority. No one suggests, for example, that the appellate court should itself rehear all the evidence relevant to a Convention issue. On any view it will adopt much of the relevant material from the survey conducted by the trial judge. Civil appellate courts other than the Supreme Court operate in accordance with rule 52.11 of the Civil Procedure Rules 1998, made pursuant to the Civil Procedure Act 1997. Paragraph 1 of the rule provides that “every appeal will be limited to a review of the decision of the lower court unless...(b) the court considers that in the circumstances of an individual appeal it would be in the interests of justice to hold a re-hearing”. Such courts should in my view seek to discharge their duty under section 6 of the 1998 Act by determining a Convention issue in accordance with this paragraph.

In paras 83 to 90 of his judgment Lord Neuberger provides fuller reasons for concluding that section 6 of the 1998 Act does not mandate fresh appellate determination of a Convention-related issue. Like Lord Clarke, at para 136 of his judgment, I agree with Lord Neuberger’s reasons as well as with his conclusion; it follows that, with respect, I do not agree with the contrary opinions expressed by Lady Hale at para 205 and by Lord Kerr at paras 116 to 127 of their judgments. Although the view of the majority is therefore that the requisite appellate exercise is that of conventional review, a question still remains about the criterion for review apt to appeals against determinations made in care proceedings.

G v G [1985] 1 WLR 647 was a dispute between parents as to which of them should have residence of the children. Lord Fraser gave the classic exposition of the role of the appellate court in reviewing a trial judge’s order in a dispute between members of a family about arrangements for a child. He described the order, at p 649, as having been made in the exercise of the judge’s discretion. This classification, which was not controversial, is hard-wired into the mind-set of family lawyers in England and Wales; and, although in Kacem v Bashir, [2011] 2 NZLR 1, the Supreme Court of New Zealand made an interesting suggestion, at para 32, that the decision in such a case was evaluative as opposed to discretionary, this is not the moment to consider whether – subject to para 45 below - to depart from the conventional classification or the consequences, if any, of doing so. Lord Fraser said at p 651:

“The Jurisdiction in such cases is one of great difficulty, as every judge who has had to exercise it must be aware. The main reason is that in most of these cases there is no right answer. All practicable answers are to some extent unsatisfactory and therefore to some extent wrong, and the best that can be done is to find an answer that is reasonably satisfactory. It is comparatively seldom that the Court of Appeal, even if it would itself have preferred a different answer, can say that the judge’s decision was wrong, and unless it can say so, it will leave his decision undisturbed.”

He added, at p 652:

“Certainly it would not be useful to inquire whether different shades of meaning are intended to be conveyed by words such as ‘blatant error’ used by the President in the present case, and words such as ‘clearly wrong’, ‘plainly wrong’, or simply ‘wrong’ used by other judges in other cases. All these various expressions were used in order to emphasise the point that the appellate court should only interfere when they consider that the judge of first instance has not merely preferred an imperfect solution which is different from an alternative imperfect solution which the Court of Appeal might or would have adopted, but has exceeded the generous ambit within which a reasonable disagreement is possible.”

The concept of the generous ambit of reasonable disagreement was derived from the judgment of Asquith LJ in Bellenden (formerly Satterthwaite) v Satterthwaite [1948] 1 All ER 343, at p 345, which Lord Fraser had already quoted.

Lord Fraser proceeded, at p 653, to quote with approval from the judgment of Bridge LJ in In re F (A Minor)(Wardship: Appeal) [1976] Fam 238. This was a dispute between a father and a grandmother about the residence of a child. Bridge LJ, at p 266, reminded himself that, in granting residence to the grandmother, the trial judge had been exercising a discretion. He observed that none of the factors which often vitiate the exercise of a discretion and so require it to be re-exercised – namely that the judge had considered an irrelevant matter, failed to consider a relevant matter, erred in law or applied a wrong principle – was present. The judge’s error, according to Bridge LJ, was in the balancing exercise, in other words that he had given too little weight to the factors favourable to the father’s case or too much weight to those adverse to them. Bridge LJ went on to hold that, where a judge’s conclusion was not justified by his advantage in seeing and hearing the witnesses and was vitiated by an error in the balancing exercise, an appellate court could set it aside.

It is clear, however, that, in quoting with approval the proposition of Bridge LJ that even only an error in the balancing exercise might justify appellate intervention, Lord Fraser was not intending to redraw any part of his earlier delineation of the boundaries of intervention. Thus an error in the balancing exercise justifies intervention only if it gives rise to a conclusion that the judge’s determination was outside the generous ambit of reasonable disagreement or wrong within the meaning of the various expressions to which he had referred.

Into its review of a trial judge’s determination of a child case an appellate court needs to factor the advantages which the judge had over it in appraising the case. In Piglowska v Piglowski [1999] 1 WLR 1360 Lord Hoffmann said, at p 1372:

“The appellate court must bear in mind the advantage which the first instance judge had in seeing the parties and the other witnesses. This is well understood on questions of credibility and findings of primary fact. But it goes further than that. It applies also to the judge’s evaluation of those facts. If I may quote what I said in Biogen Inc v Medeva plc [1997] RPC 1, 45:

The need for appellate caution in reversing the trial judge’s evaluation of the facts is based upon much more solid grounds than professional courtesy. It is because specific findings of fact, even by the most meticulous judge, are inherently an incomplete statement of the impression which was made upon him by the primary evidence. His expressed findings are always surrounded by a penumbra of imprecision as to emphasis, relative weight, minor qualification and nuance...of which time and language do not permit exact expression, but which may play an important part in the judge’s overall evaluation.”

Lord Hoffmann’s remarks apply all the more strongly to an appeal against a decision about the future of a child. In the Biogen case the issue was whether the subject of a claim to a patent was obvious and so did not amount to a patentable invention. Resolution of the issue required no regard to the future. The Piglowska case concerned financial remedies following divorce and the issue related to the weight which the district judge had given to the respective needs of the parties for accommodation. In his assessment of such needs there was no doubt an element of regard to the future. But it would have been as nothing in comparison with the need for a judge in a child case to look to the future. The function of the family judge in a child case transcends the need to decide issues of fact; and so his (or her) advantage over the appellate court transcends the conventional advantage of the fact-finder who has seen and heard the witnesses of fact. In a child case the judge develops a face-to-face, bench-to-witness-box, acquaintanceship with each of the candidates for the care of the child. Throughout their evidence his function is to ask himself not just “is this true?” or “is this sincere?” but “what does this evidence tell me about any future parenting of the child by this witness?” and, in a public law case, when always hoping to be able to answer his question negatively, to ask “are the local authority’s concerns about the future parenting of the child by this witness justified?” The function demands a high degree of wisdom on the part of the family judge; focussed training; and the allowance to him by the justice system of time to reflect and to choose the optimum expression of the reasons for his decision. But the corollary is the difficulty of mounting a successful appeal against a judge’s decision about the future arrangements for a child. In In re B (A Minor) (Adoption: Natural Parent) [2001] UKHL 70, [2002] 1 WLR 258, Lord Nicholls said:

“16.

... There is no objectively certain answer on which of two or more possible courses is in the best interests of a child. In all save the most straightforward cases, there are competing factors, some pointing one way and some another. There is no means of demonstrating that one answer is clearly right and another clearly wrong. There are too many uncertainties involved in what, after all, is an attempt to peer into the future and assess the advantages and disadvantages which this or that course will or may have for the child.

...

19.

... Cases relating to the welfare of children tend to be towards the edge of the spectrum where an appellate court is particularly reluctant to interfere with the judge’s decision.”

In this appeal M challenges both Judge Cryan’s determination that the threshold set by section 31(2) was crossed and, alternatively, his ultimate determination that Amelia’s welfare demanded that he should make the care order. The time has come for me to address the criterion for appellate review of each of these determinations; but, whatever the criterion, the appellate court will factor into its review the trial judge’s enjoyment of the advantages to which I have referred.

On any view there is nothing discretionary about a determination of whether the threshold is crossed. I consider that in the Court of Appeal Black LJ was correct, at para 9, to categorise it as, instead, a value judgement, particularly, but not only, when the court is surveying likelihood. Black LJ proceeded to adopt the approach of Ward LJ in the Court of Appeal in Re MA (Care Threshold), cited above, at para 56, that the question on an appeal against the refusal of a judge to hold that the threshold had been crossed was whether it exceeded the generous ambit of reasonable disagreement. In my judgment in that case, from the outcome of which I dissented, I asked, at para 34, whether it had been “open” to the judge to refuse to do so. In her judgment Hallett LJ asked, at para 44, whether the judge had been “plainly wrong” to refuse to do so. Although these are matters of little more than nuance, I consider in retrospect that in that case none of the three of us afforded sufficient weight to the evaluative, as opposed to the discretionary, nature of a determination whether the threshold is crossed. Ward LJ’s reference to the generous ambit of reasonable disagreement seems apt only to the review of an exercise of discretion, as in G v G. My own reference to whether the judge’s determination had been “open” to him now seems to me to have been singularly uninformative. Perhaps Hallett LJ came closest to the appropriate test in her reference to whether the determination had been “plainly wrong”. But it is generally better to allow adjectives to speak for themselves without adverbial support. What does “plainly” add to “wrong”? Either the word adds nothing or it serves to treat the determination under challenge with some slight extra level of generosity apt to one which is discretionary but not to one which is evaluative. Like all other members of the court, I consider that appellate review of a determination whether the threshold is crossed should be conducted by reference simply to whether it was wrong.

I turn to the criterion for appellate review of the ultimate determination of whether to make a care order. This is an order which the court “may” make (section 31(1) of the 1989 Act), albeit that the determination is governed by the paramountcy of the child’s welfare (section 1(1)) and can be made only following regard to specified factors (section 1(3) and (4) (b)). The court’s apparent discretion whether to make a care order has led family practitioners readily to assume that the criterion for appellate review is identical to that applicable to review of what are taken to be discretionary determinations relating to children in private law, namely that explained by Lord Fraser in G v G cited above; and it was not controversial when Sir Mark Potter P formally so ruled in Re C (Adoption: Best Interests of Child) [2009] EWHC 499 (Fam), [2009] 2 FLR 1293, at para 33. But, by contrast with the issue between the members of this court as to the impact of section 6(1) of the 1998 Act upon the role of the appellate judge, there is no issue that, since that Act came into force, the task of the trial judge in applications for care (or supervision) orders – and indeed in such applications for private law orders as can sensibly be said to represent a suggested interference with a person’s right to respect for his or her family life – is more than to exercise a discretion. His task is to comply with an obligation under the subsection not to determine the application in a way which is incompatible with that right. It follows therefore that the review which, according to the majority, falls to be conducted by the appellate court must focus not just on the judge’s exercise of a discretion but on his compliance or otherwise with an obligation. The criterion enunciated in G v G, in particular the concept of the generous ambit of reasonable disagreement, is in my view inapt to that review (as opposed, for example, to a review of a case management decision made within care proceedings: see Re TG (A Child) [2013] EWCA Civ 5, at para 38).

Lord Neuberger, at paras 90 and 91, and Lord Clarke, at para 139, suggest that the criterion for appellate review of an ultimate determination to make (or to refuse to make) a care order should, as in respect of the threshold, be whether it was wrong (or vitiated by serious irregularity). Just as in my view rule 52.11(1) of the Civil Procedure Rules helps to identify the role of an appellate court in a challenge to the determination of a Convention-related issue, so, as Lord Clarke there suggests, rule 52.11(3) helps to identify the criterion which it should adopt in that it provides: “The appeal court will allow an appeal where the decision of the lower court was – (a) wrong; or (b) unjust because of a serious procedural or other irregularity...” I agree. To be driven to jettison the principles in G v G in this context is not to say that the factors which often vitiate the exercise of a discretion – namely that the judge considered an irrelevant matter, failed to consider a relevant matter, erred in law or applied a wrong principle – become irrelevant. On the contrary they may well generate a conclusion that the determination was wrong and should be set aside and either that it should be reversed or that the application should be remitted for consideration afresh. By contrast a judge’s failure to give adequate reasons for his determination is likely to lead to its being set aside as unjust within the meaning of rule 52.11(3) (b).

There is therefore an attractive symmetry between the criterion for review of a determination of whether the threshold is crossed and that for review of a determination of whether a care order should be made. In each case it is no more and no less than whether the determination is wrong. But the simplicity of the criterion should not disguise the difficulty, in some cases, of its application.

As Mr Verdan QC on behalf of the new Children’s Guardian submits (in the course of a series of submissions of a quality which partly compensates for the deficiencies of the previous Guardian), Judge Cryan was not wrong to determine that the threshold to the making of a care order in relation to Amelia, set by section 31 of the 1989 Act, was crossed. Nor in my view was he wrong to proceed to make a care order in relation to her with a view to her adoption. In the present case the reasons for each of the two determinations can be summarised together. There were a number of features relative to the personalities of the parents and to the psychiatric conditions of the mother which raised a real possibility that, in their care, Amelia would suffer impairment of her emotional development. Nor is there anything to suggest that, in principle, emotional harm is less serious than physical harm. But Barnet realistically concede that, in the absence of one particular feature, the court might not have concluded that the likely harm was significant nor, alternatively, have felt driven to make the care order with a view to adoption. That one feature relates to the judge’s key conclusion that the characters of the parents disabled them from offering the elementary cooperation with professionals which Amelia’s safety in their home would require. Family courts regularly make allowance for the negative attitude of parents towards the social workers who personify their employers’ applications for care orders. But the level of the dishonest, manipulative, antagonistic obstructionism of the parents in this case was of a different order. Such attributes of course betokened a lack of insight into the needs of Amelia which raised wider concerns; but more immediately, they precluded the success of any rehabilitative programme, whatever its precise composition. It would have been risking enough in terms of Amelia’s welfare for the judge to have countenanced any further delay in her long-term placement following what by then had already been two years in foster care; but, had there been evidence that a way might be found of lowering the barriers erected by the parents, he might have directed an adjournment, to be measured surely in no more than a few months, in order to explore it. It might have been the proportionate response to the positive features of the parents’ case and loyal to the decision of the ECtHR in Kutzner v Germany (2002) 35 EHRR 653, at para 75, that it was “questionable whether the domestic administrative and judicial authorities [had] given sufficient consideration to additional measures of support as an alternative to what is by far the most extreme measure, namely separating the children from their parents”. In the present case, however, that avenue was not open to the judge. In a concluding sentence which correctly reflected both domestic law and the Convention’s proportionality requirement, he described adoption as “the only viable option” for Amelia’s future care. “There was no halfway house”, said Lewison LJ in his reluctant concurrence in the Court of Appeal’s dismissal of the appeal. Its dismissal was in my view the disposal which accorded with principle. This court should uphold it.

The point which His Honour Judge Cryan ultimately had to decide in this case was whether to make a care order in respect of a child, Amelia, with a view to her being adopted, against the wishes of her natural parents. To determine this point, the Judge had to resolve two main issues. The first issue was whether, in the light of the evidence, the threshold in section 31(2) of the Children Act 1989 (“the 1989 Act”) was satisfied. If he decided (as he did) that that threshold was crossed, the second main issue for the Judge was whether it was appropriate to make a care order.

Section 31(2) of the 1989 Act (“section 31(2)”) is set out in para 177 of Lady Hale’s judgment. In order to determine whether it was crossed in this case, the task the Judge faced can be analysed as involving three steps. He was required (i) to determine the factual issues, which involved resolving a substantial amount of disputed evidence, (ii) to identify the nature of the threshold, which involved construing section 31(2), and (iii) to decide whether on the primary facts he had found and the assessments he had made, that threshold was crossed. Having resolved the disputed primary facts, he decided that, in the light of those facts and the assessments he had made of the various witnesses (including Amelia’s mother and father, and a number of other factual and expert witnesses), the threshold had been crossed, without expressly discussing its nature.

As to the first step, Lady Hale, in paras 146-175, and Lord Wilson, in paras 2-22, have set out the unusual and troubling facts as agreed or as found by the Judge, as well as the procedural history.

There is no question of this court interfering with, or indeed being asked to interfere with, the findings of primary fact made by the Judge. Bearing in mind that it is a second appeal tribunal, the Supreme Court is virtually never even asked to reconsider findings of primary fact made by the trial judge. The Court of Appeal, as a first appeal tribunal, will only rarely even contemplate reversing a trial judge’s findings of primary fact.

As Lady Hale and Lord Kerr explain in para 200 and para 108 respectively, this is traditionally and rightly explained by reference to good sense, namely that the trial judge has the benefit of assessing the witnesses and actually hearing and considering their evidence as it emerges. Consequently, where a trial judge has reached a conclusion on the primary facts, it is only in a rare case, such as where that conclusion was one (i) which there was no evidence to support, (ii) which was based on a misunderstanding of the evidence, or (iii) which no reasonable judge could have reached, that an appellate tribunal will interfere with it. This can also be justified on grounds of policy (parties should put forward their best case on the facts at trial and not regard the potential to appeal as a second chance), cost (appeals on fact can be expensive), delay (appeals on fact often take a long time to get on), and practicality (in many cases, it is very hard to ascertain the facts with confidence, so a second, different, opinion is no more likely to be right than the first).

The second and third steps involved in the threshold issue can be combined into the single question of whether the primary facts found and assessments made by the Judge were capable of justifying the conclusion he reached that the threshold contained in section 31(2) was satisfied.

The second step is to determine the meaning of section 31(2), which is an issue of pure law. In relation to such an issue, the function of this Court (like that of the Court of Appeal) is uninhibited by the fact that it is an appellate tribunal. That is because there is a single “right or wrong” answer, which an appellate court has to determine for itself, although it often derives assistance from the reasoning of the court or courts below.

As to the meaning of section 31(2), and in particular the meaning of “likely to suffer significant harm” and “care … likely to be given”, Lord Wilson is rightly anxious not to encumber the comparatively simple wording of section 31(2), as expanded somewhat by section 31(10), with too much judicial encrustation. However, it seems to me that some authoritative guidance for judges and lawyers in this very important and difficult area is appropriate, in order to ensure as much predictability as possible and to minimise the likelihood of appeals. In my view, such guidance may be found in the analyses of Lady Hale at paras 179–193, Lord Wilson at paras 23-31, and Lord Kerr at para 108. I would not think it helpful to expand on what constitutes “significant harm” save to emphasise that it is interrelated with the likelihood of it being suffered, so that, as Lady Hale explains in para 188 and as she said in In re S-B (Children) (Care Proceedings: Standard of Proof) [2010] 1 AC 678, para 9, the more significant the harm, the less the required level of likelihood, and vice versa.

The final step in relation to the section 31(2) threshold issue required the Judge to address the question whether, on the primary facts he had found and assessments he had made, the threshold was crossed in this case. The decision on that question is certainly not one of law, but it is not one of primary fact either. It is a type of decision which is often described as involving the exercise of judgement, but it may fairly be said that this is not a very illuminating characterisation, because the determination of an issue of law or of an issue of fact also involves the exercise of judgement. As Lady Hale at para 199 and Lord Wilson at para 44 each say, it can be categorised as a value judgment (as Ward LJ said in In re MA (Care: Threshold) [2010] 1 FLR 431, para 56, and Black LJ said below, [2012] EWCA Civ 1475, para 9). It can also be said to be an appraisal, as Lord Kerr describes it in para 109, or an evaluation, to use Clarke LJ’s characterisation in Assicurazioni Generali SpA v Arab Insurance Group [2003] 1 WLR 577, paras 16 and 17, cited with approval by the House of Lords in Datec Electronics Holdings Ltd v United Parcels Service Ltd [2007] 1 WLR 1325, para 46.

In many cases, reversing the trial judge’s evaluation on an issue such as whether the section 31(2) threshold has been crossed, would involve an appellate court effectively disagreeing with (i) primary findings of fact made by the judge, or (ii) the impressions he obtained from seeing the witnesses (eg in terms of trustworthiness as to future conduct). In such cases, depending on the precise basis on which the appeal is mounted, the reasons for giving primacy to the trial judge’s conclusion (good sense, policy, cost, delay, and practicality) will either apply in the same way as, or will apply with somewhat less force than, they do in relation to findings of primary fact. This point is tellingly made by Lord Hoffmann in Piglowska v Piglowski [1999] 1 WLR 1360, 1372 (citing his earlier remarks in Biogen Inc v Medeva plc [1997] RPC 1, 45), in a passage quoted by Lord Wilson at para 41. It is perhaps worth adding that, immediately after that passage, Lord Hoffmann observed: “The exigencies of daily court room life are such that reasons for judgment will always be capable of having been better expressed. This is particularly true of an unreserved judgment … but also of a reserved judgment based upon notes”.

In the following paragraph of his judgment, para 42, Lord Wilson suggests that Lord Hoffmann’s remarks apply “all the more strongly” to an appeal against a decision involving the future of a child, and that is supported by an observation of Lord Nicholls quoted at the end of the paragraph. I agree: in a case such as this, the court is concentrating its focus on future multi-factorial possibilities, as opposed to present or past questions, such as the present needs of divorcing spouses (as in Piglowska) or past likely opinions which would have been formed by skilled people as in (Biogen).

When it comes to an evaluation, the extent to which the benefit of hearing the witnesses and watching the evidence unfold will result in the trial judge having a particular advantage over an appellate tribunal will vary from case to case. Accordingly, it is not possible to lay down any single clear general rule as to the proper approach for an appeal court to take where the appeal is against an evaluation (see also in this connection Robert Walker LJ in Bessant v South Cone Inc [2002] EWCA Civ 763, para 26, May LJ in EI du Pont de Nemours & Co v ST Dupont [2003] EWCA Civ 1368, para 94, and Laws LJ in Subesh v Secretary of State for the Home Department [2004] EWCA Civ 56, para 44). Accordingly, as already explained, even where the issue raised is not one of law, the reasons which justify a very high hurdle for an appeal on an issue of primary fact apply, often with somewhat less force, in relation to an appeal on an issue of evaluation.

I therefore agree with what Lord Wilson and Lord Kerr say about the right approach of an appellate court in relation to a question of evaluation in para 44 and in paras 110 and 113 respectively.

Whether article 8 of the European Convention on Human Rights (“the Convention”) has a part to play in relation to the threshold seems to me to be rather an arid issue: the important point is that the court acknowledges that no substantive order is made without all Convention rights being taken into account. Having said that, I consider that article 8 of the Convention (“article 8”) has no part to play in deciding whether the threshold is crossed, although it obviously comes very much into play when considering the issue of whether to make a care order. The threshold merely represents a hurdle which has to be crossed before the court can go on to consider whether to make a substantive order – i.e. an order which actually has an effect on a child and her parents (and sometimes on others). It is, of course, common ground that the court must consider any Convention rights before deciding whether to make a substantive order.

In this case, Lady Hale in paras 206-214 considers the evidence and findings in relation to this third step involved in this issue and concludes, albeit, with “some hesitation”, that the Judge was entitled to reach the conclusion that, on the primary factual findings and assessments of the parties which he had made, the threshold had been crossed. Lord Wilson in para 48 has less hesitation in reaching the same conclusion, as does Lord Kerr for the reasons he gives at paras 131-132.

In agreement with Lord Wilson and Lord Kerr, I consider that the Judge was fully justified in coming to the conclusion that the threshold was crossed on the primary facts as he had found them, and in the light of his assessment of the witnesses and of the risks facing Amelia if she remained with her parents. I have in mind in particular the ultimate views he formed (based on the primary facts he had found and the opportunity he had had to assess the witnesses) which are identified by Lady Hale at paras 169- 170 and by Lord Wilson at paras 20-22.

The nature of the harm which concerned Judge Cryan was (i) “the emotional harm to [Amelia] likely to be caused by” (a) “the Mother’s somatisation disorder and factitious illness disorder”, (b) “concerns … about the parents’ personality traits”, (c) “her mother’s lying”, (d) her father’s “active, but less chronic, tendency to dishonesty and vulnerability to the misuse of drugs”, and (ii) “physical harm to [Amelia]” which “cannot be discounted, for example, by over treatment or inappropriate treatment by doctors”. As to the possibility of such harm being prevented or acceptably mitigated, the Judge concluded that Amelia’s parents did not have “the capacity to engage with professionals in such a way that their behaviour will either be controlled or amended to bring about an environment where [Amelia] would be safe”. He explained that the result of this was that he could think of no “sufficiently reliable way” in which he could “fulfil [his] duty” to Amelia “to protect her from harm and still place her with her parents”.

Those conclusions are concerned with what may be characterised as risks, prospects or possible outcomes, and they are not, therefore, findings of primary fact, let alone conclusions of law. As explained above, they are evaluations based on the findings of primary fact, and on assessments of character and likely behaviour and attitudes, made by the Judge as a result of many days of considering oral and written evidence and also as a result of hearing argument. They are evaluations which are also plainly dependant on the Judge’s overall assessment of the witnesses, and in particular on his opinion as to the character and dependability of Amelia’s mother and father, and as to the reliability of the assessments of the expert witnesses. His conclusions appear to me to be ones to which, to put it at its lowest, he was fully entitled to come on the evidence he had heard and assessed. In other words, they were justified in terms of logic and common sense in the light of his findings of primary fact and his assessment of the witnesses, and they were coherently formulated. There is no basis in my view, for saying that they were wrong.

I understand the concern which Lady Hale expresses in her judgment at paras 208-222, which in many respects reflect the very wise remarks made by Hedley J in Re L (Care: Threshold Criteria) [2007] 1 FLR 2050, 2063. Although they have been referred to by Lady Hale at paras 181-182 and Lord Wilson at para 27 and were set out in full by Black LJ in the Court of Appeal, [2012] EWCA Civ 1475, para 116, those remarks merit repetition, not least because they have resonance in relation to both main issues in this case:

“50.

What about the court's approach, in the light of all that, to the issue of significant harm? In order to understand this concept and the range of harm that it's intended to encompass, it is right to begin with issues of policy. Basically it is the tradition of the UK, recognised in law, that children are best brought up within natural families. Lord Templeman, in In re KD (A Minor: Ward) (Termination of Access) [1988] 1 AC 806, 812, said this:

‘The best person to bring up a child is the natural parent. It matters not whether the parent is wise or foolish, rich or poor, educated or illiterate, provided the child's moral and physical health are not in danger. Public authorities cannot improve on nature.’

… It follows inexorably from that, that society must be willing to tolerate very diverse standards of parenting, including the eccentric, the barely adequate and the inconsistent. It follows too that children will inevitably have both very different experiences of parenting and very unequal consequences flowing from it. It means that some children will experience disadvantage and harm, while others flourish in atmospheres of loving security and emotional stability. These are the consequences of our fallible humanity and it is not the provenance of the state to spare children all the consequences of defective parenting. In any event, it simply could not be done.

51.

That is not, however, to say that the state has no role, as the 1989 Act fully demonstrates. Nevertheless, the 1989 Act, wide ranging though the court’s and social services’ powers may be, is to be operated in the context of the policy I have sought to describe. Its essence, in Part III of the 1989 Act, is the concept of working in partnership with families who have children in need. Only exceptionally should the state intervene with compulsive powers and then only when a court is satisfied that the significant harm criteria in section 31(2) is made out. …. It would be unwise to a degree to attempt an all embracing definition of significant harm. One never ceases to be surprised at the extent of complication and difficulty that human beings manage to introduce into family life. Significant harm is fact specific and must retain the breadth of meaning that human fallibility may require of it. Moreover, the court recognises, as Lord Nicholls of Birkenhead pointed out in In re H [1996] AC 563 that the threshold may be comparatively low. However, it is clear that it must be something unusual; at least something more than the commonplace human failure or inadequacy.”

It is true that much of the harm which the Judge considered could befall Amelia in this case might be said to flow from “eccentric, … barely adequate [or] inconsistent” parenting, and that it can also be said that the harm would result, to an extent at any rate, from her mother’s and father’s characteristics rather than from their parenting. There are, however, two answers to that.

The first is that it is a question of fact and degree whether the defective parenting which Amelia would undergo if she remained with her parents fell outside the wide spectrum of the acceptable “very diverse standards” (to quote Hedley J) such as would justify the state stepping in. I agree with Black LJ when she said at para 128 in her judgment in the Court of Appeal, that that was an issue which the Judge was particularly well placed to assess, and, while he could have discussed the issue more fully than he did, it seems to me that, particularly bearing in mind the very unusual facts, the expert evidence, the combination of risks, and the value of seeing the witnesses over a long hearing, it is impossible to fault the Judge’s conclusion.

The importance of a trial judge giving clear and coherent reasons for his decision carries particular force where the judgment is very likely to result in a child being adopted against her parents’ wishes. However, even in such a case, an appellate court must be careful of placing an unrealistically high burden on the trial judge. As Lord Hoffmann acknowledged in the passage quoted by Lord Wilson from Piglowska (quoting from an earlier judgment he gave), “specific findings of fact, even by the most meticulous judge are inherently an incomplete statement of the impression which was made upon him by the primary evidence”. In her recent judgment in In re L and B (Preliminary Findings: Power to Reverse) [2013] 1 WLR 634, para 46, Lady Hale, while emphasising the importance of “a fully and properly reasoned judgment”, as a means of achieving finality, recognised the “vicissitudes” which can beset any judge.

Secondly, in so far as it is said that the threatened harm was attributable to the character of Amelia’s parents rather than to their parenting activities, the parents’ characteristics which concerned the Judge would inevitably be reflected in the way they looked after, or “parented”, Amelia. In particular, it was not merely the potential emotional (and even physical) harm to Amelia owing to her parents’ deficiencies which worried the Judge. It is of central importance to understanding his conclusion that he was also concerned by the fact that the parents (especially the mother) would, or at least appeared very likely to, impede the professional people who would need access to Amelia in order to mitigate the risk or effect of any harm she might suffer. That not only served to reinforce the degree of risk (or at least to remove a way of mitigating the risk); it also amounted to a finding that, by their activities the parents would actively impede an important and beneficial source of mitigating and monitoring the harm which Amelia would face. That, on any view, must amount to defective parenting, and, in the circumstances of this case in the light of the risks which the Judge thought that Amelia would face, I consider that it amounted to defective parenting which a judge could reasonably conclude satisfied the section 31(2) threshold.

I turn, then, to the second main issue which the Judge had to decide, namely, given his finding that the threshold had been crossed, whether he should make a care order in respect of Amelia. He decided that he should do so, and, in that connection, it seems to me that three potential questions arise on an appeal against such a decision. Those questions are: (i) whether the judge applied the right legal test when resolving to make the care order; (ii) if he applied the right test, the correct approach of an appellate court on an appeal against the decision to make a care order; and (iii) whether the judge’s decision can stand, if the appellate tribunal subjects it to that test.

I turn to consider the first question, which involves first identifying the correct test. The effect of section 1(1) of the 1989 Act is that, when considering whether to make a care order, the court must treat the welfare of the child as the paramount consideration, and this involves taking into account in particular the factors identified in section 1(3), which includes, in para (g), the range of powers available to the court As Lady Hale (who knows more about this than anybody) says in para 194, the 1989 Act was drafted with the Convention in mind; in any event, with the coming into force of the Human Rights Act 1998 (“the 1998 Act”), the 1989 Act must now, if possible, be construed and applied so as to comply with the Convention. So too the Adoption and Children Act 2002 (“the 2002 Act”) must, if possible, be construed and applied so as to comply with the Convention. It also appears to me that the 2002 Act must be construed and applied bearing in mind the provisions of the UN Convention on the Rights of the Child 1989 (“UNCRC”).

A care order in a case such as this is a very extreme thing, a last resort, as it would be very likely to result in Amelia being adopted against the wishes of both her parents.

As already mentioned, it is clear that a judge cannot properly decide that a care order should be made in such circumstances, unless the order is proportionate bearing in mind the requirements of article 8.

It appears to me that, given that the Judge concluded that the section 31(2) threshold was crossed, he should only have made a care order if he had been satisfied that it was necessary to do so in order to protect the interests of the child. By “necessary”, I mean, to use Lady Hale’s phrase in para 198, “where nothing else will do”. I consider that this conclusion is clear under the 1989 Act, interpreted in the absence of the Convention, but it is put beyond doubt by article 8. The conclusion is also consistent with UNCRC.

It seems to me to be inherent in section 1(1) that a care order should be a last resort, because the interests of a child would self-evidently require her relationship with her natural parents to be maintained unless no other course was possible in her interests. That is reinforced by the requirement in section 1(3)(g) that the court must consider all options, which carries with it the clear implication that the most extreme option should only be adopted if others would not be in her interests. As to article 8, the Strasbourg court decisions cited by Lady Hale in paras 195-198 make it clear that such an order can only be made in “exceptional circumstances”, and that it could only be justified by “overriding requirements pertaining to the child’s welfare”, or, putting the same point in slightly different words, “by the overriding necessity of the interests of the child”. I consider that this is the same as the domestic test (as is evidenced by the remarks of Hale LJ in Re C and B [2001] 1 FLR 611, para 34 quoted by Lady Hale in para 198 above), but it is unnecessary to explore that point further.

The high threshold to be crossed before a court should make an adoption order against the natural parents’ wishes is also clear from UNCRC. Thus, Hodgkin and Newell, Implementation Handbook for the Convention on the Rights of the Child, Unicef, 3rd ed (2007), p 296, state that “there is a presumption within the Convention that children’s best interests are served by being with their parents wherever possible”. This is reflected in UNCRC, which provides in article 7 that a child has “as far as possible, the right to know and be cared for by his or her parents”, and in article 9, which requires states to ensure that

“a child shall not be separated from his or her parents against their will, except when competent authorities subject to judicial review determine, in accordance with applicable law and procedures, that such separation is necessary for the best interests of the child”.

Having identified the test, the other aspect of the first question is whether the Judge purported to apply that test in this case. In my view, he did, or, to put it at its lowest, his conclusions were expressed in a way which makes it clear that he considered that the test was satisfied. In the passage to which I have already referred, quoted by Lord Wilson in para 22, the Judge said that he could not see “any sufficiently reliable way that [he could] fulfil [his] duty to [Amelia] to protect her from harm and still place her with her parents”, and he immediately went on to explain that this was despite the fact that “this court strives to promote” her relationship with her parents and their family life together. He also described adoption as “the only viable option” for Amelia’s future care. As a matter of ordinary language, it seems to me clear that the Judge was there applying the test laid down by the Strasbourg court, and concluding that it was satisfied.

It is next necessary to address the question of the approach to be adopted by an appellate court when an appeal is brought against a judge’s decision to make a care order in a case such as this, and where the judge has applied, or at least has purported to apply, the correct test. As a matter of pure domestic law, this is an evaluative exercise and therefore it might appear that the approach discussed in paras 57-61 above applies. However, the issue is potentially complicated by the fact that article 8 is engaged.

There is no doubt but that Judge Cryan, as the trial judge, the first instance tribunal, was required to decide for himself whether the care order which he was proposing to make satisfied the test, which the Strasbourg jurisprudence establishes is required by article 8. The issue to be addressed concerns the correct approach of an appellate court when confronted by an appeal against the making of such an order.

What the Strasbourg jurisprudence requires (and, I would have thought, what the rule of law in a modern, democratic society would require) is that no child should be adopted, particularly when it is against her parents’ wishes, without a judge deciding after a proper hearing, with the interests of the parents (where appropriate) and of the child being appropriately advanced, that it is necessary in the interests of the child that she is adopted.

So far as any appeal against such a decision is concerned, as Lord Hoffmann said in a passage quoted by Lord Wilson in para 35, “[t]here is no Convention obligation to have a right of appeal at all”. However, to an extent at any rate, that begs the question as to the correct approach for an appellate court to adopt where there is a right of appeal. In that connection, I respectfully disagree with Lady Hale’s view as expressed in para 204 that an appellate court is under a positive obligation on every such appeal to assess the question of proportionality for itself, if that means that the Court of Appeal in this case was required to decide for itself, effectively de novo, whether the requirements of article 8, as explained in the cases mentioned in paras 195-198 of Lady Hale’s judgment, were satisfied so far as the making of a care order in respect of Amelia was concerned.

It is well established that a court entertaining a challenge to an administrative decision, ie a decision of the executive rather than a decision of a judge, must decide the issue of proportionality for itself – see the statements of principle in R (SB) v Governors of Denbigh High School [2007] 1 AC 100, paras 29-30 and 63, and in Belfast City Council v Miss Behavin’ Ltd [2007] 1 WLR 1420, paras 12-14, 24-27, 31, 42-46 and 89-91. However, this does not mean that an appellate court entertaining a challenge to a judicial decision, as opposed to an executive decision, must similarly decide the issue of proportionality for itself. If it did, it would mean that (subject to obtaining permission to appeal) litigants would be entitled to (or forced to undergo) two separate sequential judicial assessments of proportionality. I do not consider that there is anything in the Strasbourg jurisprudence or in the 1998 Act which suggests that such an entitlement should exist, even where there is a right of appeal.

That is not to say that the fact that Convention rights are involved is irrelevant if there is a right of appeal. The appeal process must “offer a fair hearing before an independent and impartial tribunal in accordance with article 6” (to quote again from Lord Hoffmann in MT (Algeria)), and, if the appeal process involves a challenge to the trial judge’s assessment of proportionality, that challenge would have to be properly and fairly addressed. But in my view, the fact that a Convention right is involved does not require an appellate domestic court to consider again the issue of proportionality for itself. What it requires is that a court considers the question of proportionality and that, if there is an appeal, any appeal process involves a proper consideration of the question of proportionality. In other words, the court system as a whole must fairly determine for itself whether the requirement of proportionality is met, but that does not mean that each court up the appeal chain does so.

I agree with Lord Wilson at para 36 that, subject to the requirements of article 6 of the Convention, it must be a question of domestic law as to how the challenge to proportionality is to be addressed on an appeal. There is, in my view, no reason why the Court of Appeal in a case such as this should not have followed the normal, almost invariable, approach of an appellate court in the United Kingdom on a first appeal, namely that of reviewing the trial judge’s conclusion on the issue, rather than that of reconsidering the issue afresh for itself.

That this is the normal function of the Court of Appeal is made clear by CPR 52.11, which states that, save in exceptional cases, every appeal is limited to a review rather than a rehearing and the appeal will be allowed only where the decision of the lower court was “wrong” or “unjust because of a serious procedural or other irregularity in the proceedings in the lower court”. The “exceptional cases” are, as a matter of principle and experience, almost always limited to those where the Court of Appeal (i) decides that the judge has gone wrong in some way so that his decision cannot stand, and (ii) feels able to reconsider, or “rehear”, the issue for itself rather than incurring the parties in the cost and delay of a fresh hearing at first instance.

As I see it, this limitation on the function of an appellate court is based on similar grounds as are set out in paras 53 and 57-61 above - see per Lord Diplock in Hadmor Productions Ltd v Hamilton [1983] AC 191, 220 and per May LJ in EI du Pont para 94. If, after reviewing the judge’s judgment and any relevant evidence, the appellate court considers that the judge approached the question of proportionality correctly as a matter of law and reached a decision which he was entitled to reach, then the appellate court will not interfere. If, on the other hand, after such a review, the appellate court considers that the judge made a significant error of principle in reaching his conclusion or reached a conclusion he should not have reached, then, and only then, will the appellate court reconsider the issue for itself if it can properly do so (as remitting the issue results in expense and delay, and is often pointless).

Not only is this consistent with the normal practice of an appeal court in this jurisdiction but it is also consistent with good sense. In many cases, and this is one, the trial judge will have seen the witnesses and had a full opportunity to assess the primary facts and to make relevant assessments (I refer again to what Lord Wilson says at paras 41-42). Once one accepts that this means that the appellate court should defer to the trial judge at least to some extent (as Lady Hale rightly does in para 205), then, unless the appellate court is confined to a primarily reviewing function, it will have some sort of half-way house role between review and reconsideration. This would seem to me to be unprincipled and to be liable to cause confusion to actual and potential litigants as well as to the judiciary. Additionally, the introduction of a second layer of judicial assessment of proportionality is likely to lead to increased cost and delay in many cases. Of course, where the trial judge has not heard oral evidence or where his findings have not depended on his assessment of the witnesses’ reliability or likely future conduct, then the appellate court will normally be in as good a position as the trial judge to form a view on proportionality.

The argument that the Convention or the 1998 Act requires the Court of Appeal to form its own view in every case where a trial judge’s decision on proportionality is challenged, appears to me to be wrong in principle and potentially unfair or inconvenient. The argument is wrong in principle because, if the function of the Court of Appeal is as I have described, then, in my view, there can be no breach of the Convention or the 1998 Act, if it conducts a review of the trial judge’s decision and only reverses it if satisfied that it was wrong. The only basis for challenging that view is, on analysis, circular, as it involves assuming that the Court of Appeal’s primary function is to reconsider not to review. The argument is potentially unfair or inconvenient, because in cases where the appeal court could not be sure whether the trial judge was right or wrong without hearing the evidence and seeing the witnesses, it would either to have to reach a decision knowing that it was less satisfactorily based than that of the judge, or it would have to hear the evidence and see the witnesses for itself.

That conclusion leaves open the standard which an appellate court should apply when determining whether the trial judge was entitled to reach his conclusion on proportionality, once the appellate court is satisfied that the conclusion was based on justifiable primary facts and assessments. In my view, an appellate court should not interfere with the trial judge’s conclusion on proportionality in such a case, unless it decides that that conclusion was wrong. I do not agree with the view that the appellate court has to consider that judge’s conclusion was “plainly” wrong on the issue of proportionality before it can be varied or reversed. As Lord Wilson says in para 44, either “plainly” adds nothing, in which case it should be abandoned as it will cause confusion, or it means that an appellate court cannot vary or reverse a judge’s conclusion on proportionality of it considers it to have been “merely” wrong. Whatever view the Strasbourg court may take of such a notion, I cannot accept it, as it appears to me to undermine the role of judges in the field of human rights.

I appreciate that the attachment of adverbs to “wrong” was impliedly approved by Lord Fraser in the passage cited from G v G (Minors: Custody Appeal) [1985] 1 WLR 647, 652, by Lord Wilson at para 38, and has something of a pedigree – see eg per Ward LJ in Assicurazioni [2003] 1 WLR 577, para 195 (although aspects of his approach have been disapproved – see Datec [2007] 1 WLR 1325, para 46). However, at least where Convention questions such as proportionality are being considered on an appeal, I consider that, if after reviewing the trial judge’s decision, an appeal court considers that he was wrong, then the appeal should be allowed. Thus, a finding that he was wrong is a sufficient condition for allowing an appeal against the trial judge’s conclusion on proportionality, and, indeed, it is a necessary condition (save, conceivably, in very rare cases).

There is a danger in over-analysis, but I would add this. An appellate judge may conclude that the trial judge’s conclusion on proportionality was (i) the only possible view, (ii) a view which she considers was right, (iii) a view on which she has doubts, but on balance considers was right, (iv) a view which she cannot say was right or wrong, (v) a view on which she has doubts, but on balance considers was wrong, (vi) a view which she considers was wrong, or (vii) a view which is unsupportable. The appeal must be dismissed if the appellate judge’s view is in category (i) to (iv) and allowed if it is in category (vi) or (vii).

As to category (iv), there will be a number of cases where an appellate court may think that there is no right answer, in the sense that reasonable judges could differ in their conclusions. As with many evaluative assessments, cases raising an issue on proportionality will include those where the answer is in a grey area, as well as those where the answer is in a black or a white area. An appellate court is much less likely to conclude that category (iv) applies in cases where the trial judge’s decision was not based on his assessment of the witnesses’ reliability or likely future conduct. So far as category (v) is concerned, the appellate judge should think very carefully about the benefit the trial judge had in seeing the witnesses and hearing the evidence, which are factors whose significance depends on the particular case. However, if, after such anxious consideration, an appellate judge adheres to her view that the trial judge’s decision was wrong, then I think that she should allow the appeal.

I am conscious that the analysis in paras 80-90 appears to differ somewhat from that of Lady Hale in paras 204-205 and of Lord Kerr in paras 116-127. However, at least in my opinion, it would, essentially for two reasons, be a very rare case where their approach would produce a different outcome from mine. First, it is only my category (iv) which gives rise to disagreement, in that they would not, as I understand it, accept that such types of case exist. However, many, probably most, cases that on my approach would fall into that category would, on their approach (especially in the light of what they say about the weight to be given to the trial judge’s assessment) be in category (iii), which would yield the same outcome. Secondly, the advantage which the trial judge has in hearing the evidence and seeing the witnesses will mainly apply to his findings of primary fact, inferences of fact, and assessment of probable outcomes, which then feed into his assessment of proportionality (and, in this case, necessity). When those factors come to be weighed on the question of proportionality (or necessity), the advantage the trial judge has will normally be of less significance, and sometimes even of very little, if any, significance.

It is unnecessary to decide whether the approach described in paras 85-90 is appropriate to any appeal concerning an evaluation even where no Convention right is involved, including the sort of issue considered in G v G, in Bellenden (formerly Satterthwaite) v Satterthwaite [1948] 1 All ER 343 and in Assicurazioni. I am not convinced that the approach is necessarily different from that advocated in those cases, although the precise formulation is undoubtedly different. However, it was not a topic on which we had any argument, as the submissions were limited to the proper approach on an appeal on proportionality, or necessity, under the Convention, and I say no more about it.

So I reach the final question on this second issue, and on this appeal: was the Judge’s decision to make a care order in this case proportionate, in the light of the conclusions he reached on the primary facts and on the assessments he made, and bearing in mind the strong general desirability of maintaining the family unit, and the possibility of other remedies? In other words, in the light of these factors, was it necessary to make such order to protect Amelia, bearing in mind the overriding necessity of the interests of the child?

In the light of the Judge’s findings, quoted by Lord Wilson in para 21, when considered against the detailed findings of primary fact and assessment which he made, it seems to me that he directed himself appropriately and came to a decision which I cannot possibly characterise as wrong. I agree with Lord Wilson’s analysis in para 48, as well as with what Lord Kerr says at paras 131-133.

While I understand, and have real sympathy with, the points made by Lady Hale in paras 216-223, I do not view the enquiries she wishes to be made on any remission as realistically open to an appellate court in the light of the various conclusions reached by the Judge. I take this view above all because of his finding, which is admirably explained by Lord Kerr in para 132 as “the inescapable difficulty … that the parents have been found to be incapable of co-operating to the necessary extent with professionals whose intervention is considered to be indispensable to the safeguarding of Amelia’s happy and fulfilled future”. That was a finding plainly open to the Judge, not least in the light of all his unchallenged findings of primary fact and his assessment of Amelia’s parents.

To put the point another way, if we were to remit the case on the basis proposed by Lady Hale, it would be to enable another judge to draw conclusions as to the future behaviour or attitude of Amelia’s mother and/or father, which were inconsistent with the findings made by Judge Cryan. He concluded that the parents would not co-operate with professionals whose access to Amelia, for whose well-being such access would be essential. But, as I have already said, those findings were ones with which an appellate court had no grounds, in my opinion, for interfering. It follows that I cannot accept that this case is one which could be properly remitted.

In deference to Lady Hale’s conclusions, I see how it could be argued that (i) the question of co-operation was given too much weight by the Judge or (ii) the possibility of setting up a system whereby co-operation was assured could have been more fully explored at the hearing. However, to allow the appeal on either ground would, I think, be wrong, in principle and in practice. In principle, because the Judge’s view was justified in the light of his findings as to the past behaviour of the parents, the impression which he formed of the parents and other important witnesses, and his assessment of the future likelihoods and risk, all of which were open to him. In practice, it is almost always possible for parties who have lost a case because of the judge’s assessment of their likely behaviour, to contend that they should be given a second chance to explore matters more fully with a view to achieving a different result. To allow this appeal on that ground would justify a remittal for fuller consideration in any case where a party was dissatisfied with a trial judge’s decision based on the assessment of the future. Of course, the issue in the present case is particularly important and sensitive, but finality is important, not just in the public interest, but for the good of Amelia (and her parents).

Having reached this conclusion, it is only right to refer to the very brief, but important judgments of Rix and Lewison LJJ in the Court of Appeal, which proleptically echo the concerns expressed by Lady Hale. They agreed with Black LJ’s full and careful reasons for dismissing the appeal, but Lewison LJ, at para 147, was worried that the Judge might have fallen foul of Hedley J’s wise remarks quoted in para 67 above, and Rix LJ wondered at para 150:

“whether this case illustrates a powerful but also troubling example of the state exercising its precautionary responsibilities for a much loved child in the face of parenting whose unsatisfactory nature lies not so much in the area of physical abuse but in the more subjective area of moral and emotional risk.”

These observations are also reflected by concerns expressed more broadly by Sloan, “Conflicting rights: English adoption law and the implementation of the UN Convention on the Rights of the Child” [2013] CFLQ 40. That Article at pp 49-50, suggests that, whereas UNCRC is “neutral about the desirability of adoption” (quoting Hodgkin and Newell, op cit p 294), the 2002 Act “unashamedly aimed to bring about ‘more adoptions more quickly’ for children in care” (quoting Harris-Short, “New Legislation: The Adoption and Children Bill – A Fast Track to Failure?” [2001] CFLQ 405). More specifically, the Article identifies a suggested inconsistency between the approach of the Court of Appeal in Re C (A Child) (Adoption: Duty of Local Authority), reported as C v XYZ County Council [2008] Fam 54, at para 15, and that of the High Court in Re A (A child) (Disclosure of Child’s Existence to Paternal Grandparents, reported as Birmingham City Council v S [2007] 1 FLR 1223, at paras 73 and 76. In Re C, it was said that “the 2002 Act does not privilege the birth family over the adoptive parents simply because they are the birth family”. In the Birmingham case, which Sloan suggests is more in line with the policy of UNCRC, Sumner J described adoption as “a last resort for any child” to be invoked only “when neither of the parents nor the wider family and friends can reasonably be considered as potential carers for the child”, and he went on to recognise a child’s “right to be brought up by her own family”.

We were not addressed on this Article or on those two cases. However, they all give added weight to the importance of emphasising the principle that adoption of a child against her parents’ wishes should only be contemplated as a last resort – when all else fails. Although the child’s interests in an adoption case are “paramount” (in the UK legislation and under article 21 of UNCRC), a court must never lose sight of the fact that those interests include being brought up by her natural family, ideally her natural parents, or at least one of them.

Hodgkin and Newell, op cit, suggest that, under UNCRC, an “adoption can only occur if parents are unwilling or are deemed by judicial process to be unable to discharge” their responsibilities towards the child. The assessment of that ability to discharge their responsibilities must, of course, take into account the assistance and support which the authorities would offer. That approach is the same as that suggested by Hedley J in the passage quoted in para 67 above, and I agree with it. It means that, before making an adoption order in such a case, the court must be satisfied that there is no practical way of the authorities (or others) providing the requisite assistance and support.

In this case, I revert to the melancholy fact that the Judge was satisfied that (i) without such assistance and support, the parents would not discharge their responsibilities to Amelia, (ii) that, as a result, there would be a grave risk of harm to her, and, crucially, (iii) that the parents would seriously impede the authorities in providing such essential assistance and support. There was ample evidence to support that conclusion, and therefore the appeal must be dismissed.

Three different types of judicial decision in care proceedings have been authoritatively identified by Lady Hale in para 199 of her judgment. The first concerns factual decisions on the evidence; the second involves consideration of whether the statutory threshold has been crossed; and the third deals with decisions as to the type of order that should be made. For the reasons that she has given, with which I agree, it is important to recognise the different intellectual exercise which is in play in each of these contexts because that will dictate the proper approach of the appellate court to a challenge about the correctness of a judge’s decision.

A conclusion by a judge at first instance on which facts have been proved, and which have not been, involves the judge sifting the evidence that has been led, assessing it and then deciding whether it has brought him or her to the necessary point of conviction of its truth and accuracy. Although an appellate court is competent to hear appeals against the findings of fact that the judge has made, of necessity, its review of those findings is constrained by the circumstance that, usually, the initial fact-finder will have been exposed to a wider range of impressions that influence a decision on factual matters than will be available to a court of appeal. This is not simply a question of assessing the demeanour of the witnesses who gave evidence on factual matters, although that can be important. It also involves considering the initial impact of the testimony as it unfolds – did it appear frank, candid, spontaneous and persuasive or did it seem to be contrived, lacking in conviction or implausible. These reactions and experiences cannot be confidently replicated by an analysis of a transcript of the evidence. For this reason a measure of deference to the conclusions reached by the initial fact finder is appropriate. Unless the finding is insupportable on any objective analysis it will be immune from review.

The second species of decision in care proceedings (whether the threshold has been crossed) is of a plainly different order from that of fact-finding. In deciding whether a child is suffering or is likely to suffer significant harm, a judge must exercise judgment. There may be factual elements to the decision such as, for instance, whether a particular type of harm occurred. Ultimately, however, the dominant character of the decision is that of an appraisal by the judge as to whether the harm is significant. With due respect to Hallett LJ’s contrary view in Re MA (Care Threshold) [2009] EWCA Civ 853, [2010] 1 FLR 431, para 42, this cannot be characterised as a finding of fact. It is a judgment made on the import of the facts found, rather than a factual finding.

Given that the determination as to whether the threshold has been crossed is one involving the exercise of judgment, what should the approach of the appellate court be to a review of that decision? Leaving aside for the moment the question of proportionality, there is much to be said for the proposition that the measure of deference that an appellate court should show to this decision approximates to that which is appropriate to a review of factual findings. Like Lord Neuberger, however, I believe that to cast the test of reviewability in this sphere as an examination of whether the judge was “plainly wrong” is potentially misleading. A finding on whether the threshold has been crossed will, in many cases, be a matter for fine judgment, however. The conclusion on this issue will be informed, at least to some extent, on the judge’s impression of the evidence. While the weight to be given to his or her conclusion as to whether the threshold has been crossed operates in a different way from that where the judge reaches a conclusion on disputed facts, since the assessment of the evidence is influential in the threshold decision, a degree of reticence on the part of an appellate court on whether to interfere with the decision is warranted. If the appellate court considers that the judge was wrong, however, it should not shrink from reversing his or her decision.

The third species of decision in care proceedings (the selection of the appropriate disposal) is, as Lady Hale has said (in para 199), on the face of it, a matter of discretion. One proceeds on the basis that there is a range of options available to the judge, all of which are defensible, but that one is chosen, in the exercise of the judge’s discretion, as that considered best to meet the judge’s perception of what is needed to fulfil the requirements of the particular case. In truth, of course, this decision partakes of the exercise of judgment as well as discretion. The distinction (although it cannot be a sharp-edged one) between a decision on disposal and the threshold decision is that in the latter case, the judge must arrive at a firm conclusion as to whether the appropriate standard has been met whereas, in the case of disposal, the judge can acknowledge that there is an array of possible disposals from which he or she selects what is considered to be the best.

Where what is under review by an appellate court is a decision based on the exercise of discretion, provided the decision-maker has not failed to take into account relevant matters and has not had regard to irrelevant factors and has not reached a decision that is plainly irrational, the review by an appellate court is at its most benign. Truly, in that instance, an appellate court which disagrees with the challenged decision of the judge will be constrained to say, even though we would have reached a different conclusion, we cannot interfere.

Absent the proportionality dimension, that is probably also true of a review of the decision on whether the threshold has been crossed. The judgment as to whether there has been or is likely to be significant harm to the child must be influenced to a large extent by what the judge finds to have been proved as a matter of fact. To reverse a decision on this will almost invariably involve a review of the correctness of the judge’s conclusion on some of the facts. For the reasons given earlier, this is potentially perilous territory for the appellate court. So, even if it feels that it would have reached a different conclusion, it should refrain unless it concludes that the decision of the trial judge is simply insupportable.

Likewise – and obviously – where what is under challenge is the selection of the order deemed by the judge to be required to meet the particular circumstances of the individual case, the view of the appellate court that it would have reached a different conclusion should not, apart from proportionality, prevail. Although this decision consists of the exercise of judgment as well as discretion, it is essentially one in which the judge is selecting one of a possible range of options about what is best required to meet the requirements of a particular case. Because that decision is inevitably influenced by, among other things, the judge’s impression of the evidence, the appellate court should be slow to substitute its view of what is best required.

Into all of this discussion, however, must come the question of proportionality. Significantly different considerations are in play when the proportionality of the decision is in issue. A decision as to whether a particular outcome is proportionate involves asking oneself, is it really necessary. That question cannot be answered by saying that someone else with whose judgment I am reluctant to interfere, or whose judgment can be defended, has decided that it is necessary. It requires the decision-maker, at whatever level the decision is made, to starkly confront the question, “is this necessary”. If an appellate court decides that it would not have concluded that it was necessary, even though it can understand the reasons that the first instance court believed it to be so, or if it considered that the decision of the lower court was perfectly tenable, it cannot say that the decision was proportionate.

Lord Wilson has said in para 36 that the European Convention on Human Rights does not require appellate courts to address issues as to Convention rights with any particular degree of intensity. He also says that it is not incumbent on appellate courts to re-hear all the evidence relevant to a Convention issue. I agree with both propositions. But an appellate court which is required to review the proportionality of a decision may not discharge its duty under section 6(1) of the Human Rights Act 1998 by merely saying that the lower court has reached a decision which is not wrong. The observations of Lord Phillips and Lord Hoffmann in MT (Algeria) v Secretary of State for the Home Department [2009] UKHL 10, [2010] 2 AC 110, which Lord Wilson quotes in para 35 of his judgment, address the argument that an appellate court was required to conduct a full-blown investigation of the facts surrounding the question whether a Convention right had been violated. What was in issue was how an appellate court should inquire into a claim that a Convention right had been breached. The passages quoted were not concerned with the quite different question as to what the appellate court had to decide. In the present case both appellate courts, the Court of Appeal and this court, have to decide whether the making of a care order was proportionate. Neither court is required to conduct a complete re-hearing of the evidence. But both courts must address directly the question whether they have been satisfied that the making of a care order is proportionate. They may not do so by reference only to the defensibility of Judge Cryan’s decision. What the appellate courts are required to decide, therefore, is not whether Judge Cryan’s conclusion is wrong. What they must decide is whether the decision was proportionate and that is a matter for their judgment, not one on which they may defer to the judgment of others.

Deferring to the judgment of others is, of course, quite a different thing from taking into account the judgment of others. An appellate court, tasked with the function of deciding whether a decision is proportionate, may – indeed, should – take into consideration any properly reasoned conclusion by a judge at first instance as to proportionality. Each member of the appellate court must ask whether he or she is satisfied that the decision is proportionate but that does not mean that the first instance judge’s reasoning should be disregarded. The distinction between examining a first instance judge’s decision to see whether it falls within the “generous ambit of reasonable disagreement” and considering it in order to decide what influence it should have on one’s own decision may seem somewhat narrow but the two assessments are importantly different from each other. In the first instance, one is not concerned (or at least not principally concerned) with whether one would have reached a different conclusion. In the second, that question is of critical importance but it can properly be influenced by an earlier process of reasoning with which one can agree.

Another, perhaps more simple, way of expressing that concept is this: where an appellate court is in the realm of review of a lower court's decision without the dimension of proportionality, if the decision is not one which the appellate court would have reached, it is obliged to consider whether the lower court’s conclusion nevertheless falls within the generous ambit of permissible decisions. If it does, it should not be reversed. If, on the other hand, the review must comprehend proportionality, that is not the approach. Generous ambit considerations do not arise. But that does not mean that the appellate court may completely disregard the reasons given by the first instance judge for his or her conclusion. These must be taken into account and given such weight as they deserve, bearing in mind that the judge has had the advantage of seeing the witnesses, hearing the evidence given in real time etc. Ultimately, however, the appellate court must frankly address the question "is the challenged measure proportionate", "is it really necessary". If the court of appeal concludes that it is not, then, notwithstanding its consideration of the first instance judge's view, the decision must be reversed.

The decision by an appellate court on whether the making of an adoption order is proportionate cannot be determined by an approach which is geared solely to testing the adequacy of the trial judge’s assessment of the proportionality issue. In my view this is impermissible because it removes the appellate court from the area of responsibility which it has to ensure that a Convention right is not infringed. Moreover, an approach that contemplates the endorsement by an appellate court of a decision on proportionality which it does not affirmatively find to be correct involves an abdication of the court’s statutory duty as a public authority. Section 6 of the Human Rights Act makes it unlawful for a public authority, including a court, to act in a way which is incompatible with a Convention right.

The inaptness of review of the trial judge’s decision on proportionality rather than the appellate court reaching its own conclusion on the issue can, perhaps, be best illustrated by a series of sequential propositions:

In my view, an appellate court cannot avoid the imperative of section 6 of HRA by viewing the matter of proportionality through the prism of the defensibility of the trial judge’s decision. An appeal in an adoption case requires the appellate court to confront the possibility that its decision could involve the infringement of a Convention right. The duty not to act in a way which is incompatible with such a right gives rise to an inevitable, concomitant duty to inquire whether the order that the court makes would have that consequence. That is an inquiry which cannot be satisfactorily answered by the conclusion that another agency has so decided. The inquiry must require the appellate court to decide for itself if the freeing order is proportionate/necessary. It is simply not an option for the appellate court to adopt a stance of agnosticism and say, well we have not reached a firm conclusion ourselves but we cannot be sure that the trial judge was wrong and therefore his decision on proportionality holds the field.

That is not to say that the trial judge’s decision is irrelevant. It is entirely consistent with the proper discharge of the duty under section 6 of HRA for an appellate court to give considerable weight to the decision of the first instance judge, endowed as he is with the conspicuous advantage of having heard the evidence unfold for the first and most pertinent time. But giving the trial judge’s conclusion on the question of proportionality appropriate weight is a crucially different exercise from saying that it withstands review because it is defendable. Giving it weight because of the benefits it enjoys involves the appellate court reaching its own decision, influenced by the conclusion of the trial judge. Deferring to the judge’s decision because it is immune from review does not require (or, indeed, permit) the appellate court to reach its own view except as a test of the viability of the judge’s view.

Lord Neuberger considers that neither the HRA nor the ECHR requires in terms that an appellate court must “decide proportionality for itself”. I agree. I also agree that ECHR does not require that there be a system of appeals in every case. What ECHR requires is that, where an appeal is available, it must be conducted in a way which is “Convention-compliant”. What Convention-compliant means in the context of this case is important and I will say something about that presently.

The critical point, it seems to me, is to identify what an appellate court is about in deciding an appeal in a care proceedings case. Ultimately, its decision comes down to the question whether to allow the parents’ wishes to be overridden. It may only do so if that course is proportionate. The fact that the decision is taken on an appeal from an earlier judicial finding does not detract from that central element of the appellate court’s function. The difficult question of how, in the context of a decision on proportionality, an appeal court should exercise its appellate role must begin at this fundamental starting point. In the final analysis, the decision to allow the parents’ wishes to be overborne, by whatever court that decision is taken, must be shown to be necessary.

The statement that an appeal must be conducted in a Convention compliant way is normally made in relation to such issues as equality of arms, access to relevant material, the ability to know the case against one etc. But it must also mean that the appeal is conducted in a way that will mean that the Convention right is vindicated in a way that is practical and effective. To have these attributes the right must not be interfered with unless the interference is proportionate and the court (any court) which deals with that question must be satisfied of that. And satisfied of that “for itself”. So, although neither HRA nor ECHR requires in terms that an appellate court has to decide proportionality for itself, the outworking of the court’s duty under section 6 of HRA in particular makes a decision on the proportionality of a freeing order unavoidable for any court deciding that question.

This does not mean, however, that the appellate court has to conclude that it is “in as good a position” as the trial judge or that it has to order that the evidence be heard again. The appellate court decides whether it is satisfied that the decision is proportionate on the basis of the material that is put before it in accordance with the normal rules that attend the hearing of an appeal. That material includes the judge’s rehearsal of the evidence, the factual findings that he has made and the reasoning which underpins his decision on the question of proportionality. All of these go into the appellate mix. Ultimately, however, the appellate court has to say to itself, are we persuaded that the decision was proportionate. I cannot see how that question can be avoided or elided if faith is to be kept with the duty under section 6 of HRA.

A suggested formulation for the test of reviewability in this area has been, “was the first instance judge wrong”. As I understand the suggestion, this is something which goes to result as well as process so that the appellate court is not merely inquiring whether the lower court went about it the right way but also whether it arrived at the right outcome. If that is so, the question arises, how does the appellate court address the question, did the lower court get it right or wrong (as opposed to did it follow the correct route to its conclusion)? It seems to me that it can only do so by asking what is the right result; in other words, by deciding the issue for itself.

Proportionality does not arise in relation to fact finding by the trial judge. It plainly is relevant to the question of disposal in care proceedings. Does it arise in relation to the crossing of the threshold? Lord Wilson thinks not. He considers that proportionality becomes material only when interference with the right to respect for family life under article 8 is in prospect and this only occurs when the threshold has been crossed and the making of a care order is in contemplation. Ward LJ in Re MA (Care Threshold) [2010] 1 FLR 431, at para 54 thought that, given the underlying philosophy of the Human Rights Act, the requirement in article 8 to have respect for family life informed the meaning of significant harm and emphasised the need for a sufficient reason for crossing the threshold.

I tend to agree with Lord Wilson that this may introduce an unnecessary layer of complexity to the inquiry whether the harm was significant. The backdrop to the decision whether sufficiently serious harm has occurred or is apprehended in order to hold that the threshold has been crossed is that this opens the gateway to a possible care order. Recognition that this is a draconian step provides sufficient emphasis on the need for the harm to be significant without adding further colour by recourse to article 8.

Whether or not article 8 has any part to play in the threshold decision, it certainly comes into full flower at the disposal stage. Lady Hale and Lord Wilson have both referred to emphatic statements by ECtHR in such cases as Johansen v Norway (1996) 23 EHRR 33, K and T v Finland (2001) 36 EHRR 255, R and H v United Kingdom (2011) 54 EHRR 28, [2011] 2 FLR 1236 and YC v United Kingdom (2012) 55 EHRR 967 concerning the stringent requirements of the proportionality doctrine where family ties must be broken in order to allow adoption to take place. I agree with Lady Hale’s statement (in para 198 of her judgment) that the test for severing the relationship between parent and child is very strict and that the test will be found to be satisfied only in exceptional circumstances and “where nothing else will do”. I also agree with what Lord Wilson has said in para 34 of his judgment, that “a high degree of justification” is required before an order can properly be made.

Both Lady Hale (with some reluctance) and Lord Wilson (more readily) have accepted that the threshold in this case was crossed. I am happy to acknowledge the great strength of their combined experience and expertise in this area of the law. Quite apart from this, however, I am personally satisfied that the threshold was indeed crossed.

The psychiatric conditions from which the mother has suffered and the way in which the parents have reacted in the past must be set against the apparently exemplary care and concern that they have exhibited towards Amelia (to use Lady Hale’s pseudonym). But the latter does not, in my estimation, counteract the former factors. Without rehearsing the facts which have been so extensively reviewed in the judgments of Judge Cryan, Black LJ, Lady Hale and Lord Wilson, it seems to me clear that there is a sufficient likelihood that the way in which, however well-intentioned they may be, the parents’ care for Amelia would be blighted by their well-established difficulties and that her emotional well-being and development would suffer significantly in consequence. The inescapable difficulty in this case is that the parents have been found to be incapable of co-operating to the necessary extent with professionals whose intervention is considered to be indispensable to the safeguarding of Amelia’s happy and fulfilled future. Of course this was not a question to be judged solely by reference to experience in previous relations with social services. As Lady Hale has pointed out, the evidence in relation to this was not universally adverse. But the established inability to co-operate, combined with the dishonesty and antagonism displayed by the parents, unmistakably presaged the impossibility of ensuring that this child would not suffer significant harm.

My conclusion on the threshold issue leads me inexorably to the same view on the question of disposal. If the difficulties that the parents presented could not be overcome – and, on my analysis of the evidence, there was no prospect of this – there really was no alternative to the care order. While I do not entirely agree with Lord Wilson on what I understand to be his view as to how an appellate court should approach the question of proportionality, I do agree with him as to the outcome of the appeal. I consider that it should be dismissed.

I agree that this appeal should be dismissed for the reasons given by Lord Wilson, Lord Neuberger and Lord Kerr. I do not detect any difference between them save as to the correct test to be adopted by an appellate court in a case of this kind. Which test is adopted does not, as I see it, affect the correct answer to each of the questions for determination in the particular circumstances of this case, namely the correct analysis of the facts, whether the section 31(2) threshold was crossed and whether a care order with a view to adoption should have been made.

However, there is a difference in principle between the approaches of an appellate court to the making of a care order adopted by Lord Wilson and Lord Neuberger on the one hand and Lord Kerr and Lady Hale on the other. I suspect that in the vast majority of cases that difference would not affect the ultimate disposal of a case of this kind, in which it is agreed on all sides that a care order cannot be made unless it is necessary in the best interests of the child. Nothing less than necessity will do, either under our domestic law or under the European Convention on Human Rights. Only in a case of necessity will an adoption order removing a child from his or her parents be proportionate.

The importance of this court addressing the difference is that one of its roles is to give guidance to the courts below and it is, to my mind, critical that there should, at the very least, be a clear majority for one approach. I agree with Lord Neuberger and Lord Wilson that the correct approach of an appellate court to the making of a care order is to treat the exercise as an appellate exercise and not as a fresh determination of necessity or proportionality. On that basis the question arises how the exercise should be approached by the appellate court. In the course of argument there was some debate whether, absent some error of principle, the Court of Appeal could only interfere with the decision of the judge if satisfied that the judge was plainly wrong.

In England and Wales the jurisdiction of the Court of Appeal is set out in CPR 52.11(3), which provides that “the appeal court will allow an appeal where the decision of the lower court was (a) wrong; or (b) unjust because of a serous procedural or other irregularity in the proceedings in the lower court”. The rule does not require that the decision be “plainly wrong”. However, the courts have traditionally required that the appeal court must hold that the judge was plainly wrong before it can interfere with his or her decision in a number of different classes of case. I referred to some of them in Assicurazioni Generali SpA v Arab Insurance Group [2003] 1 WLR 577, to which Lord Neuberger refers at para 57, at my paras 9 to 23. It seemed to me then and it seems to me now that the correct approach of an appellate court in a particular case may depend upon all the circumstances of that case. So, for example, it has traditionally been held that, absent an error of principle, the Court of Appeal will not interfere with the exercise of a discretion unless the judge was plainly wrong. On the other hand, where the process involves a consideration of a number of different factors, all will depend on the circumstances. As Hoffmann LJ put it in In re Grayan Building Services Ltd [1995] Ch 241 at 254,

“generally speaking, the vaguer the standard and the greater the number of factors which the court has to weigh up in deciding whether or not the standards have been met, the more reluctant an appellate court will be to interfere with the trial judge's decision.”

In the present context, it seems to me, in agreement with Lord Neuberger at para 58, that the court should have particular regard to the principles stated by Lord Hoffmann in Piglowska v Piglowski [1999] 1 WLR 1360 at 1372, which are quoted by Lord Wilson at para 41.

As I read their judgments, Lord Neuberger, Lord Kerr and Lord Wilson all conclude that on the question whether the section 31 threshold was crossed the test is whether the judge was wrong, not whether he was plainly wrong. Lord Neuberger and Lord Wilson have reached the same conclusion on the ultimate question, namely whether a care order should be made.

I agree with them. CPR 52.11(3) provides that the appeal court will allow an appeal where the decision of the lower court was “wrong”. As already indicated, I appreciate that the courts have given the expression “wrong” a different meaning in different contexts. However, in the context of care orders, where the court must be satisfied that it is necessary make the order, the better course is to ask whether the judge was wrong to make the order and not to ask whether he was plainly wrong. In ordinary language there is a difference between wrong and plainly wrong. If a plainly wrong test is adopted, it will be possible for an appellate court to hold that the judge was wrong to make an adoption order but was not plainly wrong to do so. How it might then be asked can it be said that it was necessary to make the order? If it was a wrong order how can it have been a necessary order? This consideration seems to me to argue strongly for the approach adopted by Lord Neuberger and Lord Wilson. For simplicity, I would apply the same test to decisions as to whether or not the threshold is crossed.

For the avoidance of doubt, as I see it, this does not mean that the judge will only be held to be wrong if he or she has made a decision which no reasonable judge could have come to. It means that the judge’s decision is wrong if the case is in one of the three categories identified by Lord Neuberger in para 93 as (v), (vi) or (vii). That is where the view expressed by the judge is one which the appellate court is doubtful about but on balance concludes was wrong, or one which the appellate court concludes was wrong or insupportable. These categories are to be contrasted with Lord Neuberger’s categories (i), (ii), (iii) and (iv). They include category (iv), where the appellate court cannot say whether the judge’s view was right or wrong. In short, I agree with the approach proposed by Lord Neuberger in paras 93 and 94.

I would only add that, as I read Lord Kerr’s judgment, he is of the opinion that, if (contrary to his view) the exercise is that of an appellate court and not that of a court determining the issue of necessity or proportionality for itself, the correct test (absent an error of principle) is whether the decision of the judge was wrong, not whether it was plainly wrong. If that is correct, there is a majority in favour of the appellate approach (Lord Neuberger, Lord Wilson and myself) and, on such an approach, a majority in favour of the test being whether the judge was wrong (Lord Neuberger, Lord Kerr, Lord Wilson and myself).

This approach will simply mean that a care order can only be made where a judge has held that such an order was necessary and the Court of Appeal (or this court on appeal from the Court of Appeal) has declined to hold that the judge was wrong. I would expect appeals to this court in adoption cases to be very rare indeed, since on this approach there will very rarely be any basis for a further appeal to this court, with all the expense and delay such an appeal entails.

This case raises some profound questions about the scope of courts’ powers to take away children from their birth families when what is feared is, not physical abuse or neglect, but emotional or psychological harm. We are all frail human beings, with our fair share of unattractive character traits, which sometimes manifest themselves in bad behaviours which may be copied by our children. But the State does not and cannot take away the children of all the people who commit crimes, who abuse alcohol or drugs, who suffer from physical or mental illnesses or disabilities, or who espouse anti-social political or religious beliefs. Indeed, in Dickson v United Kingdom (2007) 46 EHRR 937, the Strasbourg court held that the refusal of artificial insemination facilities to a convicted murderer and the wife whom he had met while they were both in prison was a breach of their rights under article 8 of the European Convention. How is the law to distinguish between emotional or psychological harm, which warrants the compulsory intervention of the State, and the normal and natural tendency of children to grow up to be and behave like their parents?

Added to this is the problem that the harm which is feared may take many years to materialise, if indeed it ever does. Every child is an individual, with her own character and personality. Many children are remarkably resilient. They do not all inherit their parents’ less attractive characters or copy their less attractive behaviours. Indeed some will consciously reject them. They have many other positive influences in their lives which can help them to resist the negative, whether it is their schools, their friends, or other people around them. How confident do we have to be that a child will indeed suffer harm because of her parents’ character and behaviour before we separate them for good?

Perhaps above all, however, this case raises the issue of when it is proper for an appellate court to interfere in the decisions of the trial judge who has heard and read all the evidence and reached his conclusions after careful cogitation following many days of hearing in court and face-to-face contact with the people involved. We all agree that an appellate court can interfere if satisfied that the judge was wrong. We also all agree that a court can only separate a child from her parents if satisfied that it is necessary to do so, that “nothing else will do”. I have come to the conclusion that the judge was indeed wrong to be so satisfied in this case. As my fellow Justices have reached a different conclusion, I must explain myself.

We are concerned with a little girl, whom I shall call Amelia, who was born on 22 April 2010 and is now nearly three years old. Her mother, Ms M, was born in 1970 and is now 42 years old. M’s parents separated when she was around five years old and when she was around seven years old her mother remarried Mr E, who thus became M’s step-father. They had two sons, born in February 1985 and 1986, who are M’s half-brothers. In her early teens M, who had been living with her grandparents, moved in with her mother and Mr E. In 1986, when M was 15 years old, Mr E began a sexual relationship with her, she became pregnant by him and had an abortion. In 1987, M’s younger sister J, then aged 12, also moved in with them. Within a remarkably short time J was the subject of care proceedings arising from a major assault upon her in November 1987. The family left her behind when they moved to Greece for several months in 1989. While there, M again became pregnant by Mr E and had another abortion (she later had five more). After the family returned to this country, M’s mother left Mr E and their two sons, then aged five and three and a half, and she began divorce proceedings in 1990. M (then 20) and the two boys remained living with Mr E.

On 17 March 1999, when M was aged 29, she and Mr E had a daughter, whom I shall call Teresa. She is now nearly 14 years old. In 2003, M was convicted of a series of frauds against financial institutions. She was sentenced to two years’ imprisonment. In her defence, she alleged that the proceedings were a conspiracy between a BT engineer and a police officer and that the police officer had indecently assaulted her. She was prosecuted for and convicted of perverting the course of justice and perjury in respect of the latter allegation and sentenced to a further three years’ imprisonment, which was reduced to 27 months on appeal because of her “serious psychological problems”. M now says that Mr E was the prime mover behind all these offences, telling her how to accomplish the frauds, taking her to the premises and waiting for her outside, and also concocting her defence. Judge Cryan heard and accepted evidence from the barrister who defended her at the time, to whom it was apparent that M was completely dominated by Mr E. Judge Cryan held that Mr E “knew all along the nature of M’s defence and, at least, was in league with her or, perhaps even, was the driving force behind it” (Judgment of 27 April 2011, para 101). But he also held that M was a “habitual and purposeful liar and an accomplished fraudster” (para 98).

In 2003, in the first criminal proceedings, Dr Spoto, presumably a consultant forensic psychiatrist, prepared a report on M, suggesting a diagnosis of Munchausen’s syndrome. We have not seen that report, but we do know from the later reports that M has a long and complicated history of frequent complaints to a large number of hospitals and medical practitioners for which a physical explanation could not always be found. In 2004, there was a report in the second criminal proceedings from Dr Taylor, who is a consultant forensic psychiatrist. M complained to him of sexual abuse by her own father but not by Mr E. His diagnosis was somatisation disorder: that is, the repeated presentation of physical symptoms, together with persistent requests for medical investigations, but where any physical symptoms present do not explain the nature and extent of the symptoms or the distress and preoccupation of the patient. M had had some real disorders, but these were not such as to explain her repeated presentations with complaints primarily of vaginal bleeding and abdominal pain. But he disagreed with Dr Spoto: there was no definitive evidence that she had feigned or fabricated symptoms, so he did not then diagnose Munchausen’s syndrome (or factitious disorder as it is now known) still less Munchausen’s syndrome by proxy (which is the reporting of feigned or fabricated symptoms in a child in order to secure medical attention for the child).

In June 2009, M was admitted again to hospital for medical investigations and this time she did not return to the family home with Mr E, her half-brothers and her daughter, Teresa. She began to make serious allegations of abuse to the Surrey police against Mr E. Mr E countered these with allegations against her. He made an ex parte application to the court and obtained an interim residence order in relation to Teresa. While M was accommodated in a refuge in Eastbourne, Mr E traced her and harassed her by telephone call and text, necessitating her move to a refuge in Hastings. While there, her younger half brother sought her out and harassed her to come home, which Judge Cryan found highly likely to be a joint exercise involving Mr E.

It was during this period that M met Amelia’s father, F, in Hastings and soon began a relationship with him. He has a great many convictions for serious criminal offences and has also used class A drugs. He has been a somewhat distant father to his four older children. Amelia must have been conceived in August 2009. In November 2009, the couple moved with the support of the police to live in the London Borough of Barnet. The first social work statement to the court reports that M “accessed ante-natal care appropriately throughout her pregnancy”. In January 2010, a midwife asked her about scarring on her body and M told hospital staff that she was fleeing domestic violence, that she had a ten year old daughter who was still living with her partner, and described 23 years of serious abuse by Mr E.

In March 2010, Barnet midwifery services made a referral to Barnet social services, because of their concerns about the wellbeing of M and the safety of Teresa. Barnet began an initial assessment but although M gave information over the phone she was reluctant to attend appointments or receive home visits. On 17 March 2010, West Sussex County Council held an initial child protection case conference about Teresa. M’s solicitors later sent the Minutes of this case conference to Barnet social services. From these they learned, not only of the serious allegations which M was making against Mr E, but also of the (less serious) allegations which he was making against her, of her criminal record and reported Munchausen’s syndrome.

Amelia was born on 22 April 2010 at only 32 weeks’ gestation and was placed in the special baby care unit. On 6 May 2010, Barnet began care proceedings in respect of her and asked for an interim care order. They did not send the parents the usual pre-proceedings letter setting out their concerns and asking for the parents’ response. But both the parents had been evasive and obstructive with them. An interim care order was granted on 10 May 2010 and Amelia was placed in foster care immediately on discharge from the hospital. Greatly to her benefit, to the credit of the local authority and her foster carer, and despite the barrage of complaints from the parents which they had to endure in the early days, she has remained with the same foster carer ever since.

Given the complexity of the situation at that point, and the difficulties they were having with the parents, it is understandable that the local authority applied for an interim care order and that the court granted it. Their concerns will have become even greater in September 2010, when Dr Bass, a consultant liaison psychiatrist and renowned authority on somatisation and factitious illness disorders, made his first report. He examined M’s medical history in great detail and agreed with Dr Taylor that she satisfied the diagnostic criteria for somatisation order. He also concluded from the “evidence that she exaggerates and possibly fabricates her biographical history as well as her medical history” that she had factitious disorder as well. So he wished to re-interview her after the fact-finding hearing.

The fact-finding hearing began in March 2011 before Judge Cryan and lasted 20 days. Technically, it was part of the private law proceedings concerning Teresa, M having made a cross application for a residence order in October 2010. West Sussex County Council took an active part in the proceedings but Barnet did not. But it was contemplated that the care proceedings about Amelia would heard together with the residence proceedings about Teresa. On 27 April 2011, Judge Cryan delivered a long and detailed judgment considering the allegations made by M against Mr E, by M’s mother against Mr E, by J against Mr E, and by Mr E against M. He prefaced this judgment with a comment on “the forensic challenge” with which we can all sympathise: “The history is chaotic and complex, and it is a forensic challenge made all the more difficult by the unreliability of almost all of the main witnesses, particularly [M and Mr E]” (para 7). In general, he concluded that Mr E was a violent, controlling and bullying personality, who used violence from time to time (para 172), and that M finally left home, not for fear of any particular danger or increase in risk, but “rather the cumulative effect of highly intolerable conduct” (para 175).

His conclusions are summarised in a “Scott schedule” prepared by the advocates, which the judge confirms is a reasonable reflection of his findings. Items 1 – 14 are M’s allegations against Mr E. Items 1 – 3 are allegations of rape. The judge found that there was a lengthy sexual relationship which began when M was 15 years old, when she was not able to give her consent, she became pregnant and had an abortion at his instigation to cover up that fact. After she reached 16, their relationship was not to be characterised as rape “but was part of a dysfunctional relationship in which he was the dominant partner”. Items 4 – 6 are allegations of repeated violence towards her. No specific findings were made, save where these were corroborated by other evidence (for example, of an assault outside court during the criminal proceedings against her), but “Mr E was a domineering, bullying and occasionally violent man who controlled the household by these means”. He did not find that the extensive scarring to M’s body was caused by the offensive actions of Mr E “but see above”. He did not find item 8, that Mr E forced M to have sex with other people so that he could watch, proved. But he did find proved item 7, that Mr E placed undue influence on M in respect of the criminal proceedings; item 9, that Mr E made threats to kill M; item 10, that Mr E would continue his abuse of M regardless of the presence of their child; item 11, that Mr E placed offensive material about M on her Facebook account; item 12, that Mr E constructed or used a website to post offensive photographs of M, and make defamatory claims about her; item 13, that he harassed her after she left, intending to “intimidate her into returning to the home of her child [Teresa] in order that he could continue his abuse of her”; and item 14, that his actions “have caused her physical and emotional injury, whilst living in his home she was in fear for her own safety, and in fleeing from his home she has remained in fear for her own safety as a result of his continuing threatening behaviour”.

Items 15 – 16 concerned his behaviour towards his wife, M’s mother; the judge found that there was domination of and violence towards her, together with some sexually aberrant behaviour. Items 17 – 26 concerned his very serious physical and psychological abuse of J, all of which the judge found proved. In addition the judge found that paedophile pornographic material had twice been found on the family’s computer; on the second occasion M could not have been involved as she had already left; on the first, he did not find that she had been involved.

Items 27 to 34 consisted of Mr E’s allegations against M. Save for one (that M had alleged that her brother had assaulted her in Hastings – which allegation was true), none of these allegations were found to be proved. Mr E had engaged in seriously dishonest conduct and lied to this and other courts, he had coached [Teresa] into making false allegations against M, and persuaded his sons to give false evidence on his behalf.

The immediate result was that West Sussex County Council issued care proceedings in relation to Teresa, initially seeking an interim supervision order, but changing this to an interim care order in the light of Mr E’s attitude towards them. In a second judgment, dated 24 May 2011, Judge Cryan granted that application. The care proceedings relating to Teresa had not been completed when the case came before us, but she has maintained the negative view of her mother into which the judge found that she was coached by Mr E, and steadfastly refuses to have anything to do with her. Dr Bentovim, a well-known child psychiatrist, has produced two reports confirming the “brainwashing” of Teresa and the effect upon her emotional development. Also in those proceedings there has been a report from the Lucy Faithfull Foundation, dated 2 August 2011, pointing to the possible links between the sexual and other abuse suffered by M and her somatisation disorder, and taking the view that the risks of sexual abuse and of exposure to the effects of the mother’s mental ill-health were currently low and could be managed.

Meanwhile, the care proceedings relating to Amelia continued. Dr Bass produced his second report, in the light of the fact finding hearing, on 4 July 2011. If M had ceased attending doctors with complaints of pelvic pain and vaginal bleeding since September 2009, it did demonstrate an improvement. He confirmed the diagnosis of somatisation disorder, but it was “difficult to be clear about” evidence that she habitually exaggerates and lies about symptoms (para 5.2). His main concern was that she had remained so long with an abusive partner and reared her daughter in this abusive environment, so that he could not be confident that she would keep a child in her care safe. There was also a risk that she might expose Amelia to unnecessary medical attention and she could present a moral risk to her child as a result of her habitual lying and deception (para 5.3).

Dr Taylor produced a second report on 12 July 2011. He accepted that there were incidents providing some evidence of co-existing factitious disorder, but he thought that her repeated presentations to doctors with unexplained symptoms were “predominantly as a result of somatisation disorder rather than factitious disorder” (para 7). He also had “some concerns extrapolating the presence of factitious disorder to the risk of fabricated or induced illness in children” (para 8).

Dr Bass had said that a parenting assessment might help to identity any abnormalities in M’s parenting style and attachment to her child. In September 2011, Barnet instructed the Marlborough Family Service to conduct a viability assessment. Their parenting assessment co-ordinator, Ms Summer (a psychologist), produced a report on 4 November 2011 concluding that further assessment of either the mother or the father “would not assist the Court, because neither of them can be relied upon to be honest in their reporting of events and to work cooperatively with child protection workers and agencies”. The parents, however, with the support of the child’s guardian, sought a further parenting assessment from Dr Dale, an experienced social work consultant and researcher. He did not form such a negative view of the parents’ ability to co-operate. His two reports, dated 20 January 2012 and 22 February 2012, concluded that Amelia “should be reunified into the care of her parents in the context of a risk management and family support programme without delay” (para 9.1). He commented that “this case raises important social policy questions worthy of public debate, about the nature and level of perceived risks of future emotional harm required for a local authority to recommend to a court that an infant be subject to compulsory adoption” (para 17.4).

The final hearing of these proceedings began on 5 March 2012. The guardian, in her report dated 7 March 2012, was unable to make a recommendation to the court. She understood the local authority’s concerns based upon the reports of Dr Bass and Dr Taylor but she was unsure whether these were enough to warrant permanent removal from the parents. She felt that Amelia had been lost in the ongoing battle between the local authority and the parents. In an addendum report after hearing the evidence of the experts and M, but not the father, she concluded that adoption was too draconian and should only be utilised if there was no alternative. She firmly believed that there was an alternative and recommended a supervision order.

She acknowledged that M and F “may not be the very best parents”. However, they had shown 100% commitment to contact, attending assiduously for one and a half hours, five days a week. The quality of the contact was appropriate and there were no major concerns from the supervisors. Amelia was clearly attached to her parents and knows them as her mum and dad. The mother of the father’s four daughters confirmed that he had always been a good, if often absent, father. When he was with them he was caring, loving and attentive. There was also nothing to suggest that M had taken Teresa to the doctor excessively.

The hearing before Judge Cryan lasted some 15 days. Once again, his judgment, dated 14 June 2012, is long (203 paragraphs), careful and detailed. He reviews in detail the evidence about the father, about the relationship between the father and the mother, about the conduct of the mother since Amelia’s birth, and about the mother’s dishonesty. The most striking feature of the parents’ relationship was the strength of and consistency of their united wish to have their daughter placed in their care. They had attended all the court hearings, all the statutory meetings with the local authority, and every contact session. Throughout the contact they had behaved unimpeachably towards their daughter. They each have a warm and loving relationship with her. But M had not lost the tendency, developed when living with Mr E, of seeking to “control by complaint” and make false allegations as a way of diverting attention. There were numerous examples of the mother’s dishonesty throughout the evidence.

He then reviewed the expert evidence, beginning with that of Dr Bass and Dr Taylor. Both maintained their diagnosis of somatisation disorder. Both described the resulting risk to Amelia as “the intergenerational transmission of abnormal health behaviour”. Dr Bass also considered that there was an unquantifiable risk of Amelia being subject to excessive medicalisation (para 111). However, there were no ongoing presentations at hospital or medicalisation of stresses or emotional problems (para 114). Further, in cross examination, Dr Bass had clearly modified his view of the mother’s factitious disorder and reached the conclusion that it was “less severe than one normally sees”, putting it at mild to moderate (para 117). But both experts agreed that her somatisation disorder brought future risk and necessitated “a plan and strategy for the future to ensure that all health care professionals are aware of [the mother’s] past and are able to intervene to protect [Amelia] should the symptoms resurface” (para 114).

He next reviewed at length the evidence of risk management from Ms Nabi of the Lucy Faithfull Foundation, Ms Summer of the Marlborough Family Service, Dr Dale, and the guardian. He shared Ms Summer’s concerns about how any child would cope with the high levels of dishonesty exercised by her mother (para 155). He noted that “even [Dr Dale] considers that here there is a risk which would require management by a risk management and family support programme” (para 177). The guardian, although firmly of the view that adoption was the wrong order here, had modified her recommendation from a supervision order to placement with the parents under a care order (paras 180, 182). Overall, he found the guardian “an unimpressive witness whose input to this complex case was little short of superficial” (para 188).

Turning to his conclusions, he found that the threshold required by section 31(2) of the Children Act 1989 “has been crossed, not perhaps in the most extreme way that is seen in some cases, but crossed it has been” (para 189). He did not there spell out the nature and degree of the future harm which Amelia would be likely to suffer if an order were not made, or the degree of likelihood that such harm would materialise, but referred back to the risk identified by Dr Bass and Dr Taylor and to Ms Summer’s concerns about the mother’s willingness to leave Teresa exposed to the risks of living with Mr E and her “quite exceptional proneness to lie”.

Having found the threshold crossed, he went on to consider Amelia’s welfare in terms of the checklist of factors set out in section 1(3) of the 1989 Act. Under her “physical, emotional and educational needs” he set out the perceived risks in clearer terms than he had done when finding the threshold crossed:

“The concerns of the local authority focus primarily on the emotional harm to [Amelia] likely to be caused by the Mother’s somatisation disorder and the factitious illness disorder. Those emotional risks are coupled with the concerns expressed by Drs Bass and Taylor and by Ms Summer, which I have accepted, about the parents’ personality traits, and her mother’s vulnerability to accommodation disorder, her mother’s lying and her father’s active, but less chronic tendency to dishonesty and vulnerability to the misuse of drugs. Whilst primarily these are engines for emotional harm, it is submitted, and I accept, that physical harm to [Amelia] cannot be discounted, for example by over treatment or inappropriate treatment by doctors” (para 192).

As to what to do about it, he rejected the views of Dr Dale and the guardian. Dr Dale’s criticisms of the local authority and Ms Summer had been unfair. He considered that the parents were “controlling and wilful”. Their unacceptable behaviour was not merely reactive to the mishandling of events by others. He concluded (para 197):

“Ultimately, I find that I am persuaded by the other group of witnesses that what the evidence clearly demonstrates is that these parents do not have the capacity to engage with professionals in such a way that their behaviour will be either controlled or amended to bring about an environment where [Amelia] would be safe and protected from emotional and/or physical harm identified by Drs Bass and Taylor”.

The father would not be able to protect Amelia from the risks because he simply did not accept them. Amelia could not be placed with her father alone because there was a high probability that the parents would not separate in any meaningful sense. There would be no way in which the situation could be effectively monitored (para 199). In any event, he had very serious reservations about the father, who “has lived a turbulent life with a very serious history of criminality, imprisonment and drug abuse” (para 200). Adjournment to make further enquiries of the extended paternal family would simply delay the inevitable (para 201). So he was left with the local authority’s care plan as the only viable option (para 203).

On appeal, it was argued that the risks identified were not sufficient to constitute significant harm, that they were not imminent, and that it was disproportionate to respond to them by permanent removal of the child. Various criticisms were made of the judge’s treatment of the evidence, in particular of his failure to deal with the detailed criticisms of Dr Bass’s diagnosis of factitious illness, with the numerous factual errors and invalid assumptions in Ms Summer’s evidence, with the detailed rebuttal of and explanation of the allegations in relation to M’s dishonesty and lack of cooperation, and on the other hand to refer to the Lucy Faithfull Foundation’s view that M did not present a risk or to mention the social worker’s evidence that there was no physical, sexual or educational risk to the child, and no suggestion that the parents would not offer her adequate physical care and emotional warmth, and that the parents’ attitude had never prevented her from undertaking her work properly with Amelia. The father also had a number of criticisms of the judge’s decision not to allow him to be assessed with a view to caring for Amelia on his own.

It is no doubt an indication of the complex and troubling nature of this case that Black LJ, who delivered the principal judgment, took enormous care in reviewing the evidence and considering the detailed criticisms made by the parents of the judge’s approach to that evidence: [2012] EWCA Civ 1475. She reminded herself that the Court of Appeal must avoid approaching the case as if it were making the determination at first instance (para 112). She took the view that the judge was aware of the need to separate the issue of what harm there actually was from the question of whether the parents would cooperate sufficiently with social services. He had intervened to explain to the social worker that it did not matter how unco-operative parents were with social services if there was no risk against which social services needed to guard (para 121).

The harm was of two kinds: that stemming from the mother’s illness-related behaviour (para 122) and that stemming from her chronic lying and the father’s dishonesty (para 123). Black LJ attached particular importance to the mother’s position in Mr E’s household, where she could not argue that her role had been entirely inert – she was a habitual and purposeful liar and an accomplished fraudster and her use of complaining tactics since she left Mr E had shown her to be his “accomplished pupil”. Her vindictive behaviour when a relative of the father had withdrawn her offer of help was “redolent of the E household” (para 125). Counsel had argued that these non-medical risks were “not what the Children Act was driving at”, but she agreed with the local authority that it was a question of degree which the judge was best placed to assess and make the necessary value judgment (para 128). She concluded “In short, the catalogue of problems identified by the judge went beyond the routine; the problems were undoubtedly more than commonplace human failure or inadequacy” (an echo of the words of Hedley J quoted at para 182 below). The judge was entitled to conclude that any strategy to manage the risks would have to go beyond the watchful eye of the GP and involve social services and that the parents would not be able to engage with professionals to ensure that Amelia was safe from harm (para 130). She went on to reject the detailed criticisms made of the judge’s approach to some of the evidence.

Lewison and Rix LJJ were clearly deeply troubled by the case. Lewison LJ was concerned about proportionality: here was a child who had not suffered any harm, who had a warm and loving relationship with her parents; the threshold had not been crossed in the most extreme way, but the order made was the most extreme that could have been made (para 142). But their task was not to make the decision but to examine whether it fell “outside the generous ambit within which reasonable disagreement is possible” so he would not push his doubts to dissent (para 148). Rix LJ also acknowledged the difficulties in the case, but agreed that one should trust the judge of trial.

Lewison LJ associated himself with Rix LJ’s concluding sentence: “I also wonder whether this case illustrates a powerful but also troubling example of the state exercising its precautionary responsibilities for a much loved child in the face of parenting whose unsatisfactory nature lies not so much in the area of physical abuse but in the more subjective area of moral and emotional risk” (para 150).

It is not the task of this court to review the factual findings of the judge in order to decide whether he was entitled to make them in the light of the evidence before him. The Court of Appeal has already performed that task with conspicuous care. This Court gave permission to appeal because of the general public importance of, and concern about, the point made by Rix LJ. On giving permission, the Court identified four specific, though inter-related, questions of law:

The threshold set by section 31(2) of the Children Act 1989 requires that the court be satisfied:

“(a)

that the child concerned is suffering, or is likely to suffer, significant harm; and (b) that the harm, or likelihood of harm, is attributable to - (i) the care given to the child, or likely to be given to him if an order were not made, not being what it would be reasonable to expect a parent to give to him; or (ii) the child’s being beyond parental control.”

By section 31(9) “‘harm’ means ill-treatment or the impairment of health or development, including, for example, impairment suffered from seeing or hearing the ill-treatment of another”; “‘ill-treatment’ includes sexual abuse and forms of ill-treatment which are not physical’”; “‘health’ means physical or mental health”; and “‘development’ means physical, intellectual, emotional, social or behavioural development”. There is no definition of “significant”, but section 31(10) provides that “Where the question of whether harm suffered by a child is significant turns on the child’s health or development, his health or development shall be compared with that which could reasonably be expected of a similar child”. Thus, while the standard of parenting expected by section 31(2)(b) is the objective standard of a reasonable parent, the level of development expected of the child is the subjective level to be expected of a child like him. Furthermore, as Munby J said in In re K, A Local Authority v N and Others [2005] EWHC 2956 (Fam), [2007] 1 FLR 399, at para 26, “the court must always be sensitive to the cultural, social and religious circumstances of the particular child and family”.

Since well before the Children Act came into force, the courts have recognised that there is a line to be drawn between parents whose personal characteristics mean that they may be less than perfect parents and parents who may cause harm to their children. Lord Templeman put the point this way in his well-known words in In re KD (A Minor)(Ward: Termination of Access) [1988] AC 806, 812:

“The best person to bring up a child is the natural parent. It matters not whether the parent is wise or foolish, rich or poor, educated or illiterate, provided the child’s moral and physical health are not endangered. Public authorities cannot improve on nature.”

If, by that last sentence, Lord Templeman was making a factual statement, then some might disagree: if local authorities remove children from unsatisfactory parents at birth and swiftly place them with highly satisfactory adoptive parents they can undoubtedly improve on nature. But in my view Lord Templeman was making a normative statement: public authorities have no right to improve on nature.

That thought has been followed through in numerous cases since. As Wall LJ pointed out in Re L (Children) (Care Proceedings: Significant Harm) [2006] EWCA Civ 1282, [2007] 1 FLR 1068, at 1084, “There are, of course, many statements in the law reports warning of the dangers of social engineering”, citing in particular Butler-Sloss LJ in Re O (A Minor) (Custody: Adoption) [1992] 1 FLR 77, 79:

“If it were a choice of balancing the known defects of every parent with some added problems that this father has, against idealised perfect adopters, in a very large number of cases, children would immediately move out of the family circle and towards adopters. That would be social engineering …”

Re L is an important case because it concerned parents with learning difficulties, very considerable in the case of the mother. The judge had found significant harm on the basis of the report of a psychologist who had not been asked to assess this. She had acknowledged that there was no obvious harm, no explicit malicious abuse or extreme abuse: “On the contrary my concern in this family relates to the more subtle and ambiguous consequences on the children flowing from parental deficiencies”. Wilson LJ commented: “So which was it? ‘Significant harm’ or ‘subtle and ambiguous consequences’? Speaking for myself, I regard the two concepts as mutually exclusive” (para 31). For these and many other concerns about the report, the case was sent back to be re-heard in the High Court. In Re L (Care: Threshold Criteria) [2007] 1 FLR 2050, at 2063, Hedley J, having quoted Lord Templeman, continued (para 50):

“It follows inexorably from that, that society must be willing to tolerate very diverse standards of parenting, including the eccentric, the barely adequate and the inconsistent. It follows too that children will inevitably have both very different experiences of parenting and very unequal consequences flowing from it. It means that some children will experience disadvantage and harm, while others flourish in atmospheres of loving security and emotional stability. These are the consequences of our fallible humanity and it is not the provenance [semble: province] of the state to spare children all the consequences of defective parenting.”

But clearly we do remove some of those children. The difficulty is to identify what it is that tips the case over the threshold. Although every parent, every child, every family is different, and, as Hedley J put it, “significant harm is fact specific and must retain the breadth of meaning that human fallibility may require of it”, there must be some consistency in the approach of both local authorities and the courts. Hedley J went on say that “it must be something unusual; at least something more than commonplace human failure or inadequacy”. It does appear that he considered that the children were suffering, and likely to suffer, some harm to their intellectual development as a result of their parents’ inadequacies, but that it was not of a character or significance to justify the compulsory intervention of the state.

The Court of Appeal wrestled with the problem of separating “harm” from “significant harm” in Re MA (Care Threshold) [2009] EWCA Civ 853, [2010] 1 FLR 431. The trial judge had found that although the parents had ill-treated another child, whose presence in their household remained something of a mystery, their own children were well cared-for, healthy, well-nourished and had strong bonds with their parents. The eldest child had been slapped, kicked and hit on the head by one or other of her parents. But that was not significant harm. Wilson LJ, at para 29, quoted Booth J in Humberside County Council v B [1993] 1 FLR 257, at p 263:

“Significant harm was defined by Miss Black, in accordance with the dictionary definition, first as being harm that the court should consider was either considerable or noteworthy or important. Then she expressed it as harm which the court should take into account in considering a child’s future.”

Wilson LJ went on to comment that while “I might not have expressed myself in quite such broad terms, they certainly foreshadow the view of Lord Nicholls, expressed three years later, that, in relation to the likelihood of harm, the threshold is set at a comparatively low level”.

At paragraph 51, Ward LJ emphasised, correctly in my view, that Lord Nicholls’ remark, in In re H [1996] AC 563, was directed, not at the threshold as a whole, nor at the threshold of significant harm, but at the threshold of likelihood of harm in the future. Lord Nicholls said nothing in that case, or in any later case, to suggest that the threshold of significance was comparatively low. Ward LJ went on, at para 54, to express the difference between “harm” and “significant harm” thus:

“Given the underlying philosophy of the Act, the harm must, in my judgment, be significant enough to justify the intervention of the state and disturb the autonomy of the parents to bring up their children by themselves in the way they choose.”

The point can fairly be made, both of this definition and of the second of the two definitions suggested by Miss Black to Booth J (para 183 above), that they are somewhat circular: the state is justified in intervening if the harm is sufficient to justify the state’s intervening. But it serves to make the point that not all harm which children may suffer as a result of their parents’ care falling short of what it is reasonable to expect is significant for this purpose. The dictionary definition, “considerable, noteworthy or important”, is to my mind more helpful. It chimes with the Guidance given by the Department of Health and Social Security when the Act first came into force: “It is additionally necessary to show that the ill-treatment is significant, which given its dictionary definition means considerable, noteworthy or important” (para 3.19). There would be no point in the threshold if it could be crossed by trivial or unimportant harm.

As to the suggestion made by Ward LJ (at para 54), that article 8 of the European Convention on Human Rights does “inform” the meaning of “significant”, I agree that it is only the court’s order, and not its finding that the threshold has been crossed, which constitutes an interference with the article 8 right. However, the reason why the threshold is crossed forms part of the court’s reasons for making the order, and these must be “relevant and sufficient”. It is not sufficient that the child would be better off in another family. That is the reason for the existence of the threshold (which was substituted for the more precise criteria laid down in the Children and Young Persons Act 1969 and the Child Care Act 1980). Furthermore, there is a relationship between this debate and the approach taken to proportionality, discussed in paragraph 197 below, which I believe to be common ground between us. If permanent removal is proportionate if it is the only way of avoiding the identified risk of harm, then it is also important that the threshold of harm is not set at too low a level, for otherwise the reasons for removal will not be sufficient: say, for example, that it is highly likely that a child will turn into an unhealthy couch potato like her parents, and only permanent removal could reliably prevent this, it would nevertheless not be a justifiable interference with family life to permit this.

Added to the difficult question of identifying significant harm is the question of identifying the degree of likelihood that such harm will be suffered in the future which is necessary to take the case over the threshold. It was held, albeit strictly obiter, in In re H (Minors) (Sexual Abuse: Standard of Proof) [1996] AC 563 that “likely” does not mean “probable” or “more likely than not”. It means, in Lord Nicholls’ well-known words, “a real possibility, a possibility that cannot sensibly be ignored having regard to the nature and gravity of the feared harm in the particular case” (at 585F). That standard has been adopted or approved in numerous later cases, including recently in this court in In re J (Children) (Care Proceedings: Threshold Criteria) [2013] UKSC 9; [2013] 2 WLR 649. It is in this respect, and this respect alone, that Lord Nicholls observed that the threshold is “comparatively low”.

The reason for adopting a comparatively low threshold of likelihood is clear: some harm is so catastrophic that even a relatively small degree of likelihood should be sufficient to justify the state in intervening to protect the child before it happens, for example from death or serious injury or sexual abuse. But it is clear that Lord Nicholls did not contemplate that a relatively small degree of likelihood would be sufficient in all cases. The corollary of “the more serious the harm, the less likely it has to be” is that “the less serious the harm, the more likely it has to be”.

Of course, another reason for adopting a test of “real possibility”, rather than “more likely than not”, is that it is extremely difficult to predict the future and to do so with the sort of accuracy which would enable a court to say that it was more likely than not that a parent would harm a child in the future. Once again, this is a particular problem with emotional or psychological harm, which may take many years to manifest itself. The Act does not set limits upon when the harm may be likely to occur and clearly the court is entitled to look to the medium and longer term as well as to the child’s immediate future.

However, the longer term the prospect of harm, the greater the degree of uncertainty about whether it will actually happen. The child’s resilience or resistance, and the many protective influences at work in the community, whether from the wider family, their friends, their neighbourhoods, the health and social services and, perhaps above all, their schools, mean that it may never happen. The degree of likelihood must be such as to justify compulsory intervention now, for there is always the possibility of compulsory intervention later, should the “real possibility” solidify.

The second element in the threshold sheds some light upon these questions. The harm, or the likelihood of harm, must be “attributable to the care given to the child, or likely to be given to him if an order were not made, not being what it would be reasonable to expect a parent to give to him” (s 31(2)(b)). This reinforces the view that it is a deficiency in parental care, rather than in parental character, which must cause the harm. It also means that the court should be able to identify what that deficiency in care might be and how likely it is to happen.

Allied to this is the definition of “harm” itself (see para 178 above). It is wide, but it is not infinite. The focus is upon the child suffering that harm, so upon the child suffering ill-treatment or suffering the impairment of her health or development. Ill-treatment will generally involve some active conduct, whether physical or sexual abuse, bullying or other forms of active emotional abuse. Impairment may also be the result of active conduct towards the child, but it could also be the result of neglecting the child’s needs, for food, for warmth, for shelter, for love, for education, for health care. Generally speaking, however, the harm is likely to be the result of some abusive or neglectful behaviour towards the child. But this is not invariably the case, as is shown by the inclusion, by way of example, “impairment suffered from seeing or hearing the ill-treatment of another”. We now know that serious harm may be done to the development of children who see or hear domestic violence between their parents.

I agree entirely that it is the statute and the statute alone that the courts have to apply, and that judicial explanation or expansion is at best an imperfect guide. I agree also that parents, children and families are so infinitely various that the law must be flexible enough to cater for frailties as yet unimagined even by the most experienced family judge. Nevertheless, where the threshold is in dispute, courts might find it helpful to bear the following in mind:

Once the threshold is crossed, section 1(1) of the Children Act requires that the welfare of the child be the court’s paramount consideration. In deciding what will best promote that welfare, the court is required to have regard to the “checklist” of factors in section 1(3). These include, at (g), the range of powers available to the court in the proceedings in question. By section 1(5), the court must not make any order unless it considers that doing so would be better for the child than making no order at all. The Act itself makes no mention of proportionality, but it was framed with the developing jurisprudence under article 8 of the European Convention on Human Rights very much in mind. Once the Human Rights Act 1998 came into force, not only the local authority, but also the courts as public authorities, came under a duty to act compatibly with the Convention rights.

It is well-established in the case law of the European Court of Human Rights that “the mutual enjoyment by parent and child of each other’s company constitutes a fundamental element of family life, and domestic measures hindering such enjoyment amount to an interference with the right protected by article 8 of the Convention” (Johansen v Norway (1996) 23 EHRR 33, among many others). However, such measures may be justified if aimed at protecting the “health or morals” and “the rights and freedoms” of children. But they must also be “necessary in a democratic society”. The court has recently summed up the principles in the context of an order freeing a child for adoption, in R and H v United Kingdom (2011) 54 EHRR 28, [2011] 2 FLR 1236, at para 81:

“In assessing whether the freeing order was a disproportionate interference with the applicants’ article 8 rights, the court must consider whether, in the light of the case as a whole, the reasons adduced to justify that measure were relevant and sufficient for the purposes of paragraph 2 of article 8 of the Convention (see, among other authorities, K and T v Finland (2001) 36 EHRR 255, para 154). . . . The court would also recall that, while national authorities enjoy a wide margin of appreciation in deciding whether a child should be taken into care, stricter scrutiny is called for as regards any further limitations, such as restrictions placed by those authorities on parental rights of access, and as regards any legal safeguards designed to secure the effective protection of the right of parents and children to respect for their family life. Such further limitations entail the danger that the family relations between a young child and one or both parents would be effectively curtailed (see Elsholz v Germany (2000) 34 EHRR 1412, para 49, and Kutzner v Germany (2002) 35 EHRR 653, para 67). For these reasons, measures which deprive biological parents of the parental responsibilities and authorise adoption should only be applied in exceptional circumstances and can only be justified if they are motivated by an overriding requirement pertaining to the child’s best interests (see Aune v Norway (Application No 52502/07) 28 October 2010, para 66; Johansen v Norway (1996) 23 EHRR 33, para 78; and, mutatis mutandis, P, C and S v United Kingdom (2002) 35 EHRR 31, para 118).”

The Strasbourg court itself has consistently applied a stricter standard of scrutiny to the national courts’ decisions to restrict or curtail contact between parent and child than it has to the decision to take a child into care in the first place. This is because, as stated, for example, by the Grand Chamber in K and T v Finland (2001) 36 EHRR 255, at para 178, there is:

“… the guiding principle whereby a care order should in principle be regarded as a temporary measure, to be discontinued as soon as circumstances permit, and that any measures implementing temporary care should be consistent with the ultimate aim of reuniting the natural parents and the child. The positive duty to take measures to facilitate family reunification as soon as reasonably feasible will begin to weigh on the responsible authorities with progressively increasing force as from the commencement of the period of care, subject always to its being balanced against the duty to consider the best interests of the child.”

Thus it is not surprising that Lewison LJ was troubled by the proportionality of planning the most drastic interference possible, which is a closed adoption, in a case where the threshold had not been crossed in the most extreme way (see para 174 above). However, I would not see proportionality in such a linear fashion, as if the level of interference should be in direct proportion to the level of harm to the child. There are cases where the harm suffered or feared is very severe, but it would be disproportionate to sever or curtail the family ties because the authorities can protect the child in other ways. I recall, for example, a case where the mother was slowly starving her baby to death because she could not cope with the colostomy tube through which the baby had to be fed, but solutions were found which enabled the child to stay at home. Conversely, there may be cases where the level of harm is not so great, but there is no other way in which the child can be properly protected from it.

Nevertheless, it is quite clear that the test for severing the relationship between parent and child is very strict: only in exceptional circumstances and where motivated by overriding requirements pertaining to the child’s welfare, in short, where nothing else will do. In many cases, and particularly where the feared harm has not yet materialised and may never do so, it will be necessary to explore and attempt alternative solutions. As was said in Re C and B [2001] 1 FLR 611, at para 34,

“Intervention in the family may be appropriate, but the aim should be to reunite the family when the circumstances enable that, and the effort should be devoted towards that end. Cutting off all contact and the relationship between the child or children and their family is only justified by the overriding necessity of the interests of the child.”

The judgments involved in care proceedings are of (at least) three different types. First are the decisions on the facts: for example, who did what to whom and in what circumstances. Second is the decision as to whether the threshold is crossed, which involves the various questions set out in para 193 above. In In re MA (Care: Threshold) [2010] 1 FLR 431, at para 56, Ward LJ was inclined to think that this was a value judgment rather than a finding of fact; and in the Court of Appeal in this case, Black LJ was also inclined to categorise it “as a value judgment rather than as a finding of fact or an exercise of discretion” (para 9). I agree and so, I think, do we all. It is certainly not a discretion and it will entail prior findings of fact but in the end it is a judgment as to whether those facts meet the criteria laid down in the statute. Third is the decision what order, if any, should be made. That is, on the face of it, a discretion. But it is a discretion in which the requirements, not only of the Children Act 1989, but also of proportionality under the Human Rights Act 1998, must be observed. What is the role of an appellate court in relation to each of these three decisions?

As to the first, the position is clear. The Court of Appeal has jurisdiction to hear appeals on questions of fact as well as law. It can and sometimes does test the judge’s factual findings against the contemporaneous documentation and inherent probabilities. But where findings depend upon the reliability and credibility of the witnesses, it will generally defer to the trial judge who has had the great advantage of seeing and hearing the witnesses give their evidence. The question is whether the findings made were open to him on the evidence. As Lord Hoffmann explained in Biogen Inc v Medeva plc [1997] RPC 1, the need for appellate caution is “based upon much more solid grounds than professional courtesy”. Specific findings of fact are “inherently an incomplete statement of the impression which was made upon him by the primary evidence. His expressed findings are always surrounded by a penumbra of imprecision as to emphasis, relative weight, minor qualifications and nuance . . .” In child cases, as Lord Wilson points out, there is the additional very important factor that the court’s role is as much to make predictions about the future as it is to make findings about the past.

As to the second, in Piglowska v Piglowski [1999] 1 WLR 1360, 1371, Lord Hoffmann cautioned the same appellate restraint in relation to the trial judge’s evaluation of the facts as to his factual findings themselves. In In re MA, Wilson LJ would have allowed the appeal “on the stark basis that, on the evidence before him, it was not open to Roderic Wood J, of all people, to reach the conclusion which he did” (para 34). Hallett LJ considered the question to be one of fact and was “not persuaded that the judge was plainly wrong” to decline to find that the threshold has been crossed (para 44). Ward LJ, having inclined to the view that it was a value judgment rather than a finding of fact, held that “it does not matter for the test this court has to apply is essentially similar, namely whether he has exceeded the generous ambit within which there is room for reasonable disagreement” (para 56). In this case, Black LJ adopted the approach of Ward LJ in In re MA (para 9).

In fact, the “generous ambit” or “plainly wrong” tests were developed, not in the context of value judgments such as this but in the context of a true discretion. In G v G (Minors: Custody Appeals) [1985] 1 WLR 647, Lord Fraser of Tullybelton approved the statement of Asquith LJ in Bellenden (formerly Satterthwaite) v Sattherthwaite [1948] 1 All ER 343, at 345:

“It is, of course, not enough for the wife to establish that this court might, or would, have made a different order. We are here concerned with a judicial discretion, and it is of the essence of such a discretion that on the same evidence two different minds might reach widely different decisions without either being appealable. It is only where the decision exceeds the generous ambit within which reasonable disagreement is possible, and is, in fact, plainly wrong, that an appellate body is entitled to interfere.”

In relation to evaluating whether the threshold has been crossed, we are all agreed that the proper appellate test is whether the trial judge was “wrong” to reach the conclusion he did. This is the test laid down in CPR 52.11(3) and there is no reason why it should not apply in this context. “Plainly” adds nothing helpful, unless it is simply to explain that the appellate court must be in one of the three states of mind described by Lord Neuberger at paragraph 93 considering the trial judge’s decision (v) on balance wrong, (vi) wrong or (vii) insupportable.

Lord Neuberger, Lord Clarke and Lord Wilson would adopt the same approach to the question of proportionality. The question here is what section 6(1) of the Human Rights Act requires of appellate courts. This is not a case such as R (Daly) v Secretary of State for the Home Department [2001] 2 AC 532, in which the courts were engaged in the careful scrutiny of the proportionality of a decision which Parliament had given to the executive to make. The courts are the primary decision makers in care cases. They are not conducting a judicial review of the local authority’s decisions. Local authorities have a range of statutory duties to help children in need and their families, to investigate and to take steps to protect children from harm. It is right, it seems to me, that they should generally follow a precautionary principle. But they do not have the power to intervene compulsorily between parent and child without the sanction of a court order. The courts are the guardians of the rights of both the children and their parents. Hence the courts, as public authorities, have the statutory duty under section 6(1) not to act incompatibly with the Convention rights. This means that the courts have the duty to assess the proportionality of the proposed interference for themselves.

Does this mean that an appellate court has the same duty to assess the proportionality of the proposed interference as does the court at first instance? This is a difficult question, but it seems to me that if the court has the duty to assess the proportionality of the decisions of a board of school governors, or of the Secretary of State, or of the immigration appellate authorities, it must a fortiori have the duty to assess the proportionality of the decisions of the trial judge in a care case. It must of course give due weight to the enormous benefit which he has had of reading and hearing all the evidence, of assessing not only the credibility and reliability of the witnesses but also their characters and personalities and the professionalism of the professional witnesses, of living and breathing the case over so many days and weeks. And it must be alive to the risks of being over-critical of the way in which a judge has expressed himself, bearing in mind the wise words of Lord Hoffmann in Biogen quoted earlier. But the court which makes the final decision is the public authority which is responsible for the invasion of Convention rights. I agree with Lord Kerr that it must decide for itself whether the order will be compatible with those rights. But I also agree that this will only make a difference in cases within Lord Neuberger’s category (iv), where the appellate judge cannot say whether the trial judge was right or wrong.

The judge collected all his self-directions on the law at the beginning of his judgment, including references to In re H, Re MA, Re L, Re B, and Re C and B. In themselves these directions can scarcely be faulted. But when he gave his reasons for concluding that threshold was crossed he did not clearly spell out (i) what the feared harm was, (ii) whether it was significant, and (iii) how likely it was to happen. Both in the Court of Appeal and in this court it was necessary for us to try and do so.

The nature of the harm feared

Black LJ was alive to the need to separate the nature of the harm feared from the question of whether the parents would co-operate with social services (para 121). The parents have no legal duty to co-operate with social services (as opposed to the health and educational services) unless the threshold is crossed. She identified two kinds of harm:

Is it significant?

Black LJ considered that this was a matter of degree, which the trial judge was best placed to assess (para 128). For my part, I would draw a distinction between the harm stemming from over-medicalisation and the other harms identified. A child whose mother actively feigns or fabricates symptoms undoubtedly suffers significant harm, but that was not the harm which the judge found here. A child whose mother exaggerates and sees the worst and thereby exposes her to unnecessary medical investigations and even treatment may well suffer significant harm. But it will be a question of degree, depending upon its frequency and severity. Many of us are anxious mothers and take our children to the doctor far more often than we should. Some of us, of course, are not anxious enough and do not take our children to the doctor when we should. There was evidence that the mother was over-anxious during the early days when Amelia was in foster care and that she over-dramatised an occasion when Amelia was taken to hospital with breathing difficulties. On the other hand, there was no evidence at all that her older daughter had been subject to excessive medicalisation, despite the fact that the mother was then much more actively engaged in her own over-medicalisation. It is clear that the judge did not place a great deal of weight upon this feature, simply commenting that “it cannot be discounted” (para 192).

The other harms, it seems to me, present much more of a problem. They are indeed the harms which stem from parental character defects. In relation to the mother’s somatisation disorder, the harm would be the emotional (and possibly also physical) damage which Amelia would suffer from copying her mother’s behaviour. In relation to the parents’ dishonesty and use of complaining tactics, the harm would be the emotional damage which Amelia would suffer, either from copying this behaviour, or from the confusion and divided loyalties resulting from her growing realisation that her mother’s version of the truth differed from her own. I accept entirely that the significance of such harms must be a question of degree. I also accept that the mother’s problems are indeed out of the ordinary. But then so were the learning difficulties of the parents in Re L. Put at their highest, these do not seem to me to come very high in the hierarchy of bad behaviours which children may learn from their parents. The father’s serious criminality, including violence, and drug abuse would come much higher.

How likely are they to be suffered?

The issue of significance and the issue of likelihood are inter-related. It is very difficult, reading the judgment of the judge, to get any impression of how likely he thought it was that these harms would materialise. I have the impression that he did not think it very likely, though he could not discount, that Amelia herself would be subject to over-medicalisation. It is difficult to gauge the likelihood of the other risks materialising, bearing in mind that the mother’s behaviour had been different since extricating herself from the household of Mr E, and the other protective factors there might be in play. It is one thing to say that the father does not take the local authority’s concerns seriously enough. It is another thing to say that this father, who is a very different person from Mr E, would not be able to counteract some of the prospective harm. In any event it is clear that these are subtle and slowly developing harms which will only materialise, if they materialise at all, in the longer term.

It is possible to get too close to the trees. I have the gravest doubts as to whether, properly analysed, the harm which is feared here is of sufficient significance or sufficient likelihood to justify a finding that the threshold has been crossed. It is difficult to discern whether the judge ever asked himself the question of degree, particularly in relation to the degree of likelihood. It is one thing to find that there are risks which the professionals have identified. It is another thing to find that those risks amount to a sufficient likelihood of sufficiently significant harm to meet the statutory threshold for compulsory intervention in the family.

However, I have to bear in mind that this extremely careful and experienced judge spent many weeks with this case. He will undoubtedly have acquired a “feel” for those questions of degree which no appellate judge could possibly acquire however close her reading of the appellate papers. Provided that we can be satisfied that he asked himself the right questions, it would be difficult indeed to interfere with his assessment. The questions which the judge must ask himself are different from the questions which the professionals must ask themselves. I would have preferred him to spell out his conclusions more clearly and confronted head on the question posed by Hedley J in Re L. But it has to be accepted that the behaviours which caused concern were both extremely unusual and unusually persistent.

No-one wants to compound the abuse which a parent, often a mother, has suffered in her own childhood by finding that that abuse renders her unable to parent her own children safely. It would be possible to see this mother wholly as a victim - the victim of the abuse which she suffered in childhood, from which her own mother was not able to protect her, and the victim of the relationship with her wicked stepfather which was established while she was still a child and from which she was unable to extricate herself for more than 20 years. But that is not the whole picture. As Black LJ identified, although she deserves our sympathy for what she has endured, “the judge’s findings disentitle her from arguing that she was solely a passive victim and that her problematic behaviour will not recur” (para 125). The judge addressed the issue when he commented of the mother: “Clearly, she was for many years under the thrall of Mr E, and might fairly be thought of as a victim within that long-standing relationship. However, having seen and heard her over a long period and having regard to her subsequent conduct, I find it difficult to see her role as being entirely inert” (para 22). We are all these days very well aware of how difficult the victims of domestic violence and abuse can find it to escape, because of the variety of subtle and not-so-subtle ways in which they can be dominated by their oppressors. But the picture which the judge gained of this mother was more complicated than that and involved a degree of collusion in the abusive environment in which her half brothers and her older daughter Teresa were being brought up.

Not without some hesitation, therefore, I am driven to the conclusion that this court is not in a position to interfere with the judge’s finding that the threshold was crossed in this case.

Was the order proportionate?

But that is not the end of the story. We all agree that an order compulsorily severing the ties between a child and her parents can only be made if “justified by an overriding requirement pertaining to the child’s best interests”. In other words, the test is one of necessity. Nothing else will do.

The judge referred to proportionality when directing himself as to the law at the beginning of his judgment, but he did not remind himself of the test when it came to making his decision. The basis of his decision was the inability of the parents to work with professionals. But it must first be asked what work with professionals would be necessary, before asking whether the parents would co-operate. Mr Feehan complains that neither the judge nor the Court of Appeal responded adequately to the detailed criticisms which he had made of the local authority’s case about this.

Dr Bass and Dr Taylor had said that the mother’s diagnosis necessitated a “plan and strategy for the future to ensure that all health care professionals are aware of M’s past and are able to intervene to protect Amelia should the symptoms resurface” (para 114). What reason was there to suppose that the parents would not co-operate with health care professionals? There was no evidence that Teresa had been exposed to inappropriate attention from the medical professions. The mother had accessed ante-natal care appropriately during her pregnancy. The mother had co-operated with the investigations conducted by Dr Bass and Dr Taylor and the father had co-operated with Dr Bass. The mother had been, of course, a frequent user of medical services, and this was the main reason for concern. But none of this evidence suggests that it would not be possible to devise a “plan and strategy” to enable the health care professionals to be aware of the situation and take appropriate action should it become necessary.

But what about the need for co-operation with the social services? Even Dr Dale, the professional who was most supportive of the parents’ case, accepted that there would need to be a “risk management and family support programme” (First report, para 20.1) although details would require clarification. There was little evidence about what this might entail, other than the brief enquiries made by the guardian during the hearing. There was conflicting evidence about the parents’ ability to co-operate with such a programme, whatever it might be. On the one hand, West Sussex social services confirmed that the mother had tried to get them to intervene to protect Teresa after she had left and her solicitors had sent a copy of the case conference minutes about Teresa to Barnet social services while she was pregnant with Amelia. She had also cooperated with the enquiries by the Lucy Faithfull Foundation and by Dr Dale (as was to be expected). The parents had been able to co-operate with a succession of workers who were supervising their contact with Amelia over the whole of her life. Their initial relationship with the foster carer was not a happy one, but it had much improved. And they had been able to co-operate with the child’s guardian.

On the other hand, they had not been so co-operative with Barnet social services. As the judge found, “from the start she has failed to co-operate reasonably with the local authority and at times has behaved in a singularly unconstructive way” (para 140). This was, as some of the examples given by Lord Wilson show, putting it mildly. Perhaps this is not to be wondered at. Their original contact was in the context of concern about Teresa and the household of Mr E. The response was to seek an interim care order separating mother and baby without taking the usual step of a pre-proceedings letter explaining matters to them. Anyone who has had to leave a premature baby in a special baby care unit can empathise with the feelings of a mother who is prevented from taking her baby home when, miracle of miracles, that baby is well enough to be discharged from hospital. Of course, the first social work statement to the court explained why the authority was making the application. But the scene was set for a rocky relationship. And this will not have been improved by the parents’ frequent complaints about Amelia’s progress in foster care.

The other negative relationship was with Ms Summer of the Marlborough Family Service. In her oral evidence, Ms Nabi of the Lucy Faithfull Foundation, who was generally supportive of the mother, was surprised and worried by this. Ms Summer had adopted the method of challenging the parents about the various aspects of their behaviour which were a matter of concern. This had clearly not gone down well with them, they had been at times dishonest, evasive, petulant and immature. In effect, the parents were willing to be helpful when they perceived that a professional was helping them but not when they perceived the professional to be the enemy.

But it was essential to set all this evidence against the evidence of the harm which was feared that Amelia might suffer in the future and the sort of programme which might be needed to protect her. It was not established that the mother was immediately in need of the sort of intensive psychological therapy which would make such challenging demands upon her. The question was what monitoring and support was an “overriding requirement pertaining to the child’s best interests”.

It must not be forgotten that this is a child who as yet has suffered no harm at all (except possibly the harm of being separated from her mother so soon after birth). She has had the advantage of remaining with the same foster carer throughout, where she is doing well. She has also had the enormous advantage of establishing a strong and loving relationship with her parents, who have given her “child centred love and affection in spades”, as the judge put it. Their commitment has been excellent and the fact that in all the circumstances their behaviour during contact has attracted so little criticism and so much praise is extraordinary. She will eventually have to move on from her foster home and the only question is whether she moves to a completely new home with adoptive parents as yet unidentified or whether she moves to live with the parents she knows and loves and who know and love her.

In all the circumstances, I take the view that it has not been sufficiently demonstrated that it is necessary to bring the relationship between Amelia and her parents to an end. In the circumstances of this case, it cannot be said that “nothing else will do” when nothing else has been tried. The harm that is feared is subtle and long term. It may never happen. There are numerous possible protective factors in addition to the work of social services. There is a need for some protective work, but precisely what that might entail, and how the parents might engage with it, has not yet been properly examined.

Accordingly, I would have allowed the appeal and sent the case back for a fresh and in-depth enquiry by the child’s new guardian (her original guardian having sadly died soon after the judge’s judgment), who would be able to examine both the necessity for and the viability of the sort of measures which were only beginning to be explored by the previous guardian. My understanding of the careful submissions made to us on her behalf is that this would have been her preferred solution had we not now been so far down the road. Of course the safest solution for Amelia now is almost certainly adoption. But I take the view that the judge was indeed wrong to hold this a proportionate response to the risks which he had identified and that it is my duty to say so.