The On-Ramp

A safeguard is supposed to be a brake — a point where someone independent can look at what the state wants to do to you and say no. Canada builds these brakes in series: to be held against your will, the paperwork must pass a justice of the peace, then a physician, and your detention can later be reviewed by a lawyer and a board. Four hands on the lever. The promise is redundancy. The reality is a relay: each hand finds the prior signature already on the page and adds its own, and the one direction the machine cannot easily move is stop. Then turn it around. When a person arrives at the hospital and asks to be let in — the textbook case the whole apparatus exists to catch — the same machine that will authorize two weeks of detention on a stamped chain has no on-ramp for him at all. That is not a system that measures danger. It is a system that measures who started the file.

§01 — The two doors

Every psychiatric admission enters through one of two doors, and they are not built the same way. Through the first door the state brings you: police apprehend a person, or a justice of the peace orders it, and the machinery of authorization engages — forms, signatures, a 72-hour assessment, a certificate that can hold you for two weeks. Through the second door you bring yourself: you walk into an emergency room and say you are in crisis. The first door has an elaborate apparatus of authorization built around it. The second door has a triage nurse, a waiting room, and a security desk.

The series’ method is to watch where a state puts its procedure, because procedure marks what the state takes seriously enough to formalize. Here the procedure clusters almost entirely around the first door — the one through which the state acts on you. Almost none governs the second — the one through which you ask. Hold that asymmetry in view; the rest of the case is its consequence. A machine this carefully built to authorize detention, and this casually built to receive a plea, is telling you what it is for.

• P1 Prior independent authorization is a genuine safeguard — the Charter requires a neutral arbiter, capable of refusing, between the citizen and the state’s power to detain. The case concedes this in full.
• P2 A safeguard placed in series is a brake only if each gate independently re-opens the question and can really say no.
• P3 In practice the gates run one way — ex parte, on documents the applicant drafts, with no automatic review of the front end — and the same machinery offers no authorized pathway to the person who self-presents.
• P4 So the output is announced as “independent judicial oversight” and “care.” That the gates exist does not establish that they brake — and a machine that auto-admits the pushed-in while turning away the asking is not weighing need. It is laundering the direction of state will into authority. The asymmetry is the tell.

Counter: redundancy that only runs one direction is not redundancy. Four gates that each defer to the last are not four reviews. They are one decision, signed four times.

§02 — Deference in depth

Take the first door — the one through which the state commits you — and walk the chain. In Ontario the sequence is named on the forms. A Form 2 is an order, signed by a justice of the peace, directing police to apprehend a person and bring them for examination; anyone may swear the information that asks for it. A Form 1 is the physician’s application that then authorizes up to 72 hours of assessment. A Form 3, signed by a second physician, is the certificate of involuntary admission that holds the person for up to two weeks. Each is presented as an independent check on the one before. Each, in operation, mostly ratifies it.

The law itself says these must be more than rubber stamps. The Supreme Court has been emphatic that the authorizing officer is a brake, not a clerk. In Hunter v. Southam (1984) the Court held that the person authorizing a search “must…at a minimum, be capable of acting judicially,” assessing the competing interests of state and individual “in an entirely neutral and impartial manner.” In R. v. Araujo (2000) it spelled out the duty: the authorizing judge “should not view himself or herself as a mere rubber stamp, but should take a close look at the material,” granting authorization “only as far as need is demonstrated” — and the applicant must “set out the facts fully and frankly.” That is the standard. The question is whether the structure lets the gate meet it.primary

The structure does not. The application is ex parte — only the applicant is in the room; there is no one to argue the other side. The information is drafted by the very party seeking the power. And the only large-scale Canadian look behind the curtain found the brake leaking badly: a 2017 review for the Manitoba Law Journal — a panel including a judge, lawyers, and a senior RCMP officer — examined 125 warrant applications and found that 23% should not have been authorized, and 20% would not have survived a challenge in court. Roughly one in five gates that were supposed to hold, did not. (In Ontario, a justice of the peace need not even be legally trained.) The cascade does not require anyone to act in bad faith. It only requires each gate to assume the last one did the real looking.primary

Four hands on the lever, and the lever only moves one way. Each gate inherits the prior signature and adds its own. Nobody decides; everybody signs.

This is the same move Case 20 · The Echo found in an intelligence file — the verification step doing the laundering — except here it is stacked. Place the checks in series and each becomes a reason for the next to relax: the physician sees a justice’s order already issued; the second physician sees the first’s Form 1 already signed. The legitimacy of “four independent safeguards” is claimed for a process in which the number of stamps grows while the quantum of independent scrutiny shrinks toward the first one — which was ex parte to begin with.

Counter: defence in depth becomes deference in depth. The more gates you add in a line, the more each can lean on the others, until the chain’s real output is not scrutiny but a thicker stack of signatures.

§03 — No brake on the first two weeks

Here is where the defender raises the strongest-sounding reply, and it is worth stating plainly: there is a review. A detained person can have their involuntary status heard by an independent tribunal — in Ontario, the Consent and Capacity Board. So the chain is not the end of it. But look at how that review is built, because its construction is the point.

It is downstream: it reviews a detention that is already in force; it never authorizes one in advance. It is opt-in: in Ontario you may apply to the Board each time a Form 3 or renewal is signed, but you must know the right exists, ask for it, and obtain counsel — from inside detention. The first hearing that happens automatically, with no application, does not arrive until the fourth certificate of renewal, by which point the detention has run for months. And the counsel who would make the opt-in real is the next starved gate: legal-aid representation that cannot always move at the pace the regulations allow. So the only adversarial brake in the whole structure is downstream of the detention, available only if summoned, and throttled by the resources of the body meant to summon it.primary

For the person held during those first two weeks, the distinction between “a safeguard that exists but is downstream, opt-in and starved” and “no safeguard” is not a distinction at all. And this is not only an Ontario quirk: in British Columbia, mental-health-law experts have called the province’s involuntary-detention regime “uniquely problematic,” warning that detained patients “lack legal protection” because the law deems their consent rather than testing it. The brake is present on the org chart and absent at the moment it would have to bite.primary

And the opt-in presupposes the one thing the setting can quietly remove: a patient well enough to use it. This is a medicated environment. Detention and treatment are governed separately — a capable involuntary patient keeps the right to refuse medication, as the Supreme Court held in Starson v. Swayze (2003) — but capacity to consent to treatment is assessed by the same institution, a finding of incapacity hands the decision to a substitute decision-maker, and emergency chemical restraint is permitted to control immediate risk. The right to seek a detention review never formally disappears; what can disappear is the clarity it runs on. The single brake that depends on the patient’s own initiative — to know the right, form the intent, instruct counsel — assumes a patient in a condition the ward is empowered to alter. And the asymmetry runs to the family, too: a relative can open the first door, because the Form 2 is sworn by “anyone,” and a relative may even become the substitute decision-maker for treatment — but no relative can open the exit. There is no family lever on the detention itself. The door that let them in does not run in reverse for the people who love them.primary

Counter: a review you have to know about, ask for, and fund a lawyer to pursue — from inside a locked ward, clear-headed enough to try — is not a check on the decision to lock the ward. It is a check on whether the person the ward can sedate can still find the door.

§04 — The two detentions

The state has two ways to take your liberty against your will. One is the criminal charge. The other is the civil committal this case has been walking. They look like cousins: both are detention, both ordered through the state, both reviewable by a tribunal. Then you set the protections side by side, and the resemblance collapses.

The Charter’s strongest procedural shields are triggered by five words: “a person charged with an offence.” Section 11 hangs on that phrase, and from it flow the right not to be compelled to testify against yourself (s.11(c)), the presumption of innocence (s.11(d)), and the right not to be denied reasonable bail without just cause (s.11(e)). The accused gets the rest of it too: the right to silence before trial (R. v. Hebert; s.7), counsel before questioning (s.10(b)), and a state that must prove its case beyond a reasonable doubt to an adversary. None of those five words describe a psychiatric patient. There is no charge and no offence. So by their own terms, the section 11 protections never attach.primary

Put the interrogation room and the ward side by side. In the room, the officer must caution you, may not compel you to speak, and cannot turn your silence into evidence; the state builds its case without your help, or it does not build it. On the ward, the assessment is the case. The diagnostic interview is not classified as interrogation, so the right to silence does not reach it: there is no caution, you cannot decline it the way a suspect declines to talk, and what you say informs the decision to hold or renew you. So, often, does what you refuse to say, written down as “guarded” or “lacking insight.” The accused is warned that anything he says may be used against him, and handed the power to say nothing. The patient is given neither the warning nor the power.

A man accused of an indictable offence can be home on bail by morning. A man committed on an allegation has no bail to ask for, and no right to stay silent in the room that decides whether he leaves.

Two limits, kept in plain sight, because the point is sharp enough without them. First, “used against you” here means against your liberty, the decision to commit or renew, not to convict you of a crime: section 13 of the Charter bars that later criminal use, a protection the committal context never has to overcome, because it is not chasing a conviction. It is chasing your detention. Second, this is not a claim that a court has struck any of it down. Courts have largely upheld civil committal as a justified, non-penal exercise. The narrower observation is the one that holds: the same loss of liberty, often longer than a criminal remand, is reached with markedly fewer protections, because it is filed under care rather than offence. The classification, not the severity, decides which shields appear.analysis

Push the asymmetry one step further and it inverts into something worse. The civil detainee has no bail hearing to ask for, but suppose that, mid-apprehension, the person resists, as someone in acute crisis often does. The instant they lay a hand on the officer executing the warrant they have, technically, committed a criminal offence: resisting or obstructing a peace officer (Criminal Code s.129), or, with any force, assaulting one “with intent to resist or prevent the lawful… detention of themselves” (s.270). And a criminal charge brings exactly what the committal withheld — a bail hearing before a justice, counsel, the presumption of innocence, the right to silence. The only way to earn the audience is to commit the crime of resisting the detention.primary

But the protections attach to the offence, not to the detention. The bail hearing is about the scuffle; the committal still has none, and the two run in parallel — a person can be released on the assault charge and held the same hour on the certificate. Worse, the one act of resistance is spent twice: on the criminal track it is the charge, and on the civil track it is a symptom — “agitated,” “combative,” “lacking insight” — the very evidence used to justify keeping them. One act, two doors, both closing. And for some assaults on police the bail is reverse-onus, presumed against, so the stream that was supposed to grant release can instead deny it. The protections are real; they simply never reach the thing that took the liberty.

Civil committal is the only lengthy detention the state offers with no bail hearing of its own. The one way to earn one is to commit a crime resisting the detention, which is then filed as proof you needed detaining.

Counter: the accused is presumed innocent and may go free on bail; the committed person is presumed unwell and must talk his way out, to an examiner whose notes are the evidence. The same loss of liberty, with the burdens reversed.

§05 — The on-ramp that wasn’t there

Now the second door, and the fact that proves what the machine is for. Everything above describes a system that authorizes detention with great procedural energy when the state initiates it. Watch what the same system does when the initiative comes from the person in crisis instead.

On 21 May 2020, Samwel Uko — twenty years old, a former University of Saskatchewan Huskies football player, visiting family in Regina — sought help twice that day at the Regina General Hospital. On the first visit he was seen: a physician assessed him, recorded depression, referred him to a mental-health clinic, and released him before eleven that morning. On the second visit, returning in worse condition, he was removed by four security guards before he was registered or seen by a triage nurse, telling them, on video the public has since watched, “No, I have mental issues.” About an hour later he was found dead in Wascana Lake. The Saskatchewan Health Authority later apologized — “We failed” — acknowledged it had failed to provide follow-up care, and paid his family $81,000. A 2022 coroner’s inquest returned twenty recommendations. (The inquest itself, and what it did with those recommendations, is the subject of Case 08; the point here is narrower.)primary

Set the two doors side by side. Through the first, a person the state wants held can be carried past a justice of the peace and two physicians into a two-week detention with no adversary in any room and no automatic review. Through the second, a person who walked in asking — presenting the exact danger-to-self the Mental Health Act exists to catch — met no justice, no form, no authorized pathway in, and a faster response from security than from triage. The machine that manufactures authorization on demand for the state-initiated case produced, for the self-initiated one, no on-ramp at all.

The same machinery that will authorize your detention in an afternoon could not produce one “yes” for a man asking to be let in.

That is the asymmetry, and it is the whole argument. If the apparatus measured danger, the self-presenting person in acute crisis would be the easiest case it ever sees. Instead the elaborate procedure attaches to the file the state opens, and the bare threshold — unreviewable front-desk discretion, no form, no record of a refusal — governs the file you try to open yourself. The recommendations that followed name what works in that unguarded space: the inquest jury urged anti-racism and de-escalation training and a psychiatric nurse on staff around the clock, and Uko’s family has said he was turned away because he was Black. Carry that as the record carries it — the family’s account, and a jury’s formal call for training against institutionalized racism and unconscious bias. The structural reading does not depend on settling it; it only notes where the bias would operate. It operates exactly where there is no JP, no form, and no review: the one decision in the system that all the procedural machinery does not touch.

Counter: a procedure dense enough to detain you for two weeks, and a threshold bare enough to eject you before triage, are not an oversight in the design. They are the design — calibrated to the direction of the request, not its urgency.

§06 — The strongest case for the chain

The case is only worth making if it survives the best version of the other side, so here it is, without strawmanning. The ex parte character of the first gates is not a flaw but a necessity in much of criminal procedure: you cannot tell the subject of a search that it is coming. Justices and physicians do refuse and do narrow — Araujo not only permits the questioning, it demands it, and many weak applications never reach a gate because the applicant screens them out first; a high authorization rate can reflect that screening rather than abdication. The downstream review board is a real tribunal that really discharges people. And in the medical context the gates are not merely legal formalities: a physician completing a Form 1 or Form 3 is exercising an independent clinical judgment the law obliges them to form themselves, and the 72-hour and two-week limits are real ceilings, not open-ended powers.

All true, and the case grants it. But notice what the strongest defence argues for. Every point above is a reason the gates could brake — that the law equips them to. None of it answers the structural facts: that the only Canadian audit found a fifth of warrants shouldn’t have issued; that the front end has no adversary and no automatic review; that the review that exists is opt-in and downstream; and that the entire apparatus has no on-ramp for the person who asks. The defence describes the system as designed. The case describes the system as it runs — and, in the Uko door, as it failed. The gap between the two is the laundering: the legitimacy of the design is spent on the behaviour of the practice.

Counter: “the gates are empowered to refuse” is not the same claim as “the gates refuse.” The first is the brochure. The second is the question — and the one audit we have answered it one in five times the wrong way.

§07 — The receipt printer

There is a sharp way to put all this, and because it is sharp the case carries it as someone else’s words, not its own verdict. The charge in circulation runs: “Call it independent judicial oversight when a justice of the peace signs what police put in front of them. If approval is the default outcome, the JP isn’t acting as a brake — they’re the receipt printer.” It is a polemic, and the case does not adopt its tone or its certainty. A justice who signs is not, as a matter of fact established here, a printer; some refuse, and the data to prove a near-universal grant rate across Canada does not exist.

What the case adopts instead is the court’s version of the same standard, which cannot be dismissed as polemic because the Supreme Court wrote it: the authorizer must not be “a mere rubber stamp.” That is the yardstick. The polemic and the precedent agree on what a real safeguard would be; they disagree only on how often the real thing is present. And the honest answer — given the ex parte structure, the absent front-end review, the opt-in remedy, the one audit’s one-in-five, and the door that had no on-ramp — is that the structure makes the rubber stamp the path of least resistance and the refusal the act of courage. A safeguard that depends on courage to function is not, reliably, a safeguard. It is a hope wearing a safeguard’s uniform.

Counter: you do not need to prove every justice is a printer. You need only show the machine is built so that printing is easy and refusing is hard — and then ask who built it that way.

§08 — The on-ramp, named

Strip it to the structure. The state advertises a chain of independent safeguards around the power to detain a person against their will — a justice of the peace, physicians, a review board — and the legitimacy of that word, independent, is what the chain trades on. But the gates sit in series, each ex parte or deferring to the last; the front end has no adversary and no automatic review; the only true review is downstream, opt-in, and starved of the counsel that would trigger it. The redundancy manufactures legitimacy in proportion to the number of stamps, while the independent scrutiny it advertises trends toward the first, ex parte signature. That is authorization laundering: the form of oversight kept while its function drains out.

And then the fact that names the whole thing. The same apparatus that runs so smoothly when the state pushes a person in had no door at all for Samwel Uko when he asked to come in. A system that measured danger would find the self-presenting man in crisis its clearest case; this one found him a security matter. So the procedure is not a measure of need. It is a measure of who opened the file — dense where the state acts on you, bare where you ask for help. The elaborate machinery is not, at bottom, a safeguard against wrongful detention. It is a processing line for state-initiated committal, and it is structurally indifferent — at the threshold, hostile — to need that originates with the patient.

The chain is not a brake on the state. It is an on-ramp for the state — and there is no on-ramp for you.

The case keeps its limits in plain sight. It does not say any particular committal was unlawful, or any particular gate corrupt, or any clinician unqualified; the failure it describes is in the geometry, not the people. It does not relitigate Samwel Uko’s death, which the inquest examined and Case 08 records; it borrows from that record only the one fact the structure needs. And it does not resolve whether racism turned him away — it reports that his family says so and that a coroner’s jury called for training against institutionalized bias, and observes only that the threshold is precisely where such bias would meet no procedural check at all. So when a state shows you a row of safeguards around its power to hold you, count the doors. Ask which way each one opens. Ask whether the person they were built to protect could ever make one open from the outside. The On-Ramp is the answer to that last question, and the answer, for the man who asked, was no.