In June I argued on this blog that the Supreme Court’s decision in A Reference by the Attorney General for Northern Ireland [2026] UKSC 16 broke faith with a discipline that runs from chapter 29 of the Magna Carta through the writ of habeas corpus to Article 5: that a confinement must rest on a reason which can be stated, recorded and tested before someone other than the confiner. The Deprivation of Liberty Safeguards were that discipline exercised by proxy, for people who cannot demand it for themselves. The Reference withdrew the proxy from anyone who can be read as content, by treating valid consent as independent of capacity. I said then that contentment is not consent, and that those most affected would be those least able to object.
The decision has since been examined on this blog as an exercise in the apex court correcting its own reading of Strasbourg downwards (Hleb Buziuk, 30 June). Four months on, the first instance courts have begun to apply it to actual people. This post is about two of those decisions, delivered three weeks apart, applying the same judgment to a man of 25 and a boy of 15 under comparable restrictions. In Salford City Council v AR [2026] EWCOP 45 HHJ Burrows asked what the Supreme Court’s consent test requires, found that it requires a great deal. In In re Mustafa (A Child) [2026] EWHC 2159 (Fam) Mr Recorder Jack, sitting as a High Court judge and faced with a boy who had said one year that he did not like being physically restrained and the next that it was ‘fine’, did not ask the question at all. Together they show what the Reference has produced: not a test but a set of gaps, which contentment can be used as a veneer to conceal.
The test
Strasbourg has long held that a deprivation of liberty has two elements: an objective one, confinement in a restricted space for a not negligible time, and a subjective one, the absence of valid consent to that confinement. Cheshire West treated them in sequence: the objective element was met if the person was under continuous supervision and control and not free to leave, and the subjective element followed automatically if the person lacked capacity, since someone who cannot consent cannot have consented. The Reference changed both. The objective element is now a multifactorial assessment in which the purpose of the restrictions, the relative normality of the placement and the person’s own compliance or lack of objection all count [53(i)], [53(iv)]. The subjective element is now an ‘autonomous concept’: a person without legal capacity may still give valid consent if they are ‘conscious of their environment’, have ‘a basic understanding of their living circumstances’ and ‘manifest their acceptance’ of them [53(ii)], [151]. The two elements overlap, so that acceptance can defeat a claim at either stage. Mere compliance or acquiescence is not enough, and where there is serious doubt no consent is to be inferred [53(v)], [191]. What ‘acceptance’ means, and how a court is to find it, is not said.
The field
At least eight published decisions have applied the Reference since June, and they sort by one principle: objection. Where the person objects, they are deprived of their liberty, as Theis J held of a fifteen-year-old under 3:1 supervision who found it ‘oppressive’ (Essex County Council v MH [2026] EWHC 2328 (Fam)) and of a man who wanted to return to Scotland (Argyll and Bute Council v RF [2026] EWCOP 41). Where the person does not object, they are not, whether or not they could consent: non-verbal twins who ‘appear not to object’ were held not confined (Wigan MBC v PK [2026] EWCOP 52), and so was a woman who told the judge she hated her placement, because ‘many people would’ (Derby City Council v SM [2026] EWCOP 46). Salford and Mustafa are the two in which the court had to confront the harder question: what to make of a person who does not object, cannot consent in the ordinary sense, and whose acceptance is the only thing standing between them and the safeguards.
Salford: the gaps named
AR is 25, with a moderate to severe learning disability and autism, living under 24-hour one-to-one supervision, two-to-one in the community, behind locked doors and gates, with window restrictors and CCTV. The evidence of his wishes was ‘mixed’: attachment to his parents, enjoyment of activities, attempts to open the gate, a suitcase from home he would not unpack.
HHJ Burrows held that he was not deprived of his liberty on the objective element. The decisive matters ([65]) were that AR ‘will always require a significant level of supervision’, ‘will never be “free to leave” wherever he safely resides’, and that the restrictions are ‘for his own good’ and ‘relatively normal’. Normality, in other words, is measured against a person with AR’s needs, which is the comparison Lady Hale refused to make in Cheshire West and the Reference requires. Its logic runs one way: the more profound the disability, the more supervision is normal, and the harder it becomes for the person to enforce their right to liberty. The judge saw this (‘those facts cannot be minimised’, [54]) and was bound to look past it.
What follows is the most serious judicial engagement so far with the subjective element. He made three points. First, the new concept sits awkwardly in the statute: Schedule A1 to the Mental Capacity Act defines the ‘mental capacity requirement’ for the safeguards by reference to capacity, and the Supreme Court ‘has imported into the statute a further interpretative guide’ under which a person who fails that requirement can nonetheless consent, one which ‘appears to focus on compliance and non-capacitous agreement with (otherwise known as acquiescence to) restrictions’ ([43]-[45]). That is a judge saying, politely, that the new concept is acquiescence renamed.
Secondly, the practical questions are unanswered: ‘Is evidence required as to agreement and disagreement, and can it be challenged? What about the situation where P’s position changes from time to time? Who decides?’ ([72]). His answer was that a court ‘must always err on the side of caution’. The phrase is Lady Hale’s, from the judgment the Reference overruled for its ‘misplaced policy concern’.
Thirdly, he applied the Reference’s standard to AR and found it unmet. AR could express ‘simple likes, dislikes and preferences’, and the judge accepted those as real. The difficulty was understanding:
there is little evidence that he understands the nature of the choice between living at the placement and living elsewhere, the restrictions which operate within the placement, or that he has sufficient appreciation of those arrangements to make an autonomous decision either accepting or rejecting them. His ability to communicate contentment or dissatisfaction is real, but that is not necessarily the same as understanding the nature of the arrangements to which that contentment relates ([76]).
Contentment with a home is one thing; consent to the restrictions that make it one is another; and the Reference’s ‘basic understanding of their living circumstances’ is a standard about the second. On that standard AR could not consent, though it made no difference to the result.
Mustafa: the gaps unfilled
Mustafa has Prader-Willi syndrome and developmental delay. He has lived at a specialist residential school for five years, 52 weeks a year, under a care order, behind locked doors, with a locked kitchen at night, one-to-one support rising to two-to-one and constant supervision. There had been 48 recorded incidents since July 2025, and physical restraint, two members of staff each holding an arm, was being used ‘approximately every 9-10 days’ and increasing in duration and intensity. After the Reference the local authority applied to withdraw its application to extend the deprivation of liberty order.
The local authority’s case was that the placement is ‘his home’, that he ‘likes living at [the School]’, and that valid consent is ‘the critical feature of the present case’. The Guardian’s evidence was less tidy. Mustafa, ‘responded almost immediately in a positive manner, raising some concern as to whether he was processing information’. At his most recent review he had said he felt ‘fine’ about being held; at the previous review he had said he did not like restraint. The Guardian’s conclusion was that ‘it remains difficult to know whether this is compliance by Mustafa rather than valid consent’. That is the case the Supreme Court described at [191] as requiring ‘anxious consideration’, and the case of a person whose ‘position changes from time to time’.
The Recorder did not address the doubt. He held that a child who is not Gillick competent may nonetheless consent. He accepted that the Guardian was ‘right to point out’ that Mustafa might withdraw any consent, ‘even (or indeed especially) whilst physical restraint is being used on him’. He then held that this did not matter, because schools may restrain pupils at common law and under section 93 of the Education and Inspections Act 2006 regardless of the pupil’s wishes: ‘where the school uses physical restraint, it either has Mustafa’s consent or it is entitled to use restraint under its common law powers. Either way, there is no deprivation of liberty.’
This conflates the objective and subjective parts of the test. The Recorder made no finding on the objective element. The locked doors, the supervision and the restraint every nine days are recorded without dissent; a deprivation of liberty order had been made, and later extended, on those same facts. The conclusion is reached in a single step. If Mustafa consents, there is no deprivation. Further, if he does not, the school may restrain him anyway, so there is still no deprivation.
That reasoning, if followed, would leave Article 5 with nothing to do. Whether a school may restrain a child goes to the justification of a confinement, not whether they are constrained in fact. The Supreme Court criticised Lord Neuberger in Cheshire West for confusing ‘the concept of justification of deprivation of liberty … with the criteria for identifying a deprivation in the first place’ ([94]). Mustafa makes the same confusion in reverse: the restraint would be lawful, so there is no deprivation to justify. On that logic no instance of a lawful detention would count as a detention at all.
Article 5 is built on the opposite premise. It lists in sub-paragraphs (a) to (f) the detentions that are lawful and subjects every one of them to the guarantees in Article 5(4), because a power to confine is the thing the Article exists to limit, not a reason for the limit to fall away. Many powers authorise restraint in the right conditions: the Mental Capacity Act, the Children Act, the inherent jurisdiction, the common law. The question is never whether some power authorises it, but whether the restraint, under whatever power, deprives the person of liberty, in which case the power must be exercised within the Article’s limits. One power cannot be substituted for another to take the exercise out of the Article’s reach. If it could, every deprivation of liberty order ever made for a child at a residential school was unnecessary, since the school always had section 93.
Further, the judgment records that Mustafa is ‘generally happy in his placement’, and that is the whole of its engagement with the subjective element. The questions the Salford judge asked are all raised on the face of the Guardian’s evidence; none is answered. The court heard inconsistent contentment with restrictions, treated it as consent, and said that either way it did not matter because another power could be used to justify the restraints.
What it means to be content
Set side by side, the two cases show the Reference’s central gap: contentment is ‘usually decisive’ ([151]), and nothing is said about what it is. HHJ Burrows supplied it by asking what the person understands and whether the acceptance attaches to the restrictions; the Recorder supplied it from the care provider’s record. Nothing in the Reference says which of them is right.
In June I argued that the inarticulate would be read as content by default. Mustafa is that prediction on the facts: a boy who answers before he processes, recorded as ‘fine’ in front of his carers. The subjective element is being read against the person in both directions, silence as acceptance and complaint as ordinary life; only in Salford was it read with the ‘anxious consideration’ the Supreme Court prescribed, though with no effect on the outcome.
In the Reference, the Supreme Court suggested that those removed from Article 5 keep Article 8, the Mental Capacity Act and the Care Act ([10], [141], [185]). Mustafa shows what that looks like at the point of use. Once the authority withdrew, it was doubtful whether the Guardian could keep the question before the court, and the Recorder suggested that the remedy was for the Official Solicitor or a parent to sue for false imprisonment. Instead of prospective safeguards against the state, a child lacking legal capacity must then rely on a litigation friend to bring a private law claim on their behalf.
Conclusion
The Reference was decided without facts on which to test the ruling. The facts have now arrived, and they show a test in which the question of what contentment consists of, for a person who cannot consent, is answered or not according to the judge. The first judge to answer it found that it demands an understanding of the arrangements and not merely a liking for the home, and that the court must refuse to infer consent where there is doubt. The next did not ask it, found a power to restrain elsewhere. That is an insufficiently restricted power, which was the whole of my objection in June. What Salford shows is that the Reference, applied with care, leads to the conclusion that the more severe a person’s disability, the less likely they are to be legally deprived of their liberty. What Mustafa shows is that the whole of its test, objective and subjective limbs together, can be reduced to a gloss of ‘contentment’ without further justification. Neither outcome is satisfactory.
Craig Wells is a lecturer in constitutional law, legal theory and social work law at the University of Kent. His articles ‘Habeas Corpus by Proxy: Liberty, the Inarticulate Subject and the Deprivation of Liberty Reference’ and ‘Contentment is not Consent: Capacity, the Subjective Element of Article 5, and the Unequal Liberty of Disabled People’ are forthcoming in Public Law and the Medical Law Review respectively.
(Suggested citation: C. Wells, ‘The Veneer of Contentment: Mustafa, Salford and the Gaps the Deprivation of Liberty Reference Left Behind’, U.K. Const. L. Blog (19th October 2026) (available at https://ukconstitutionallaw.org/))
Leave a Reply