How a Substance Becomes a Controlled One

Consider a compound that has existed, unchanged, in the chemical literature for thirty years. On a Monday it is entirely lawful to possess, sell and ship. On a Thursday it is a controlled substance carrying a custodial sentence. Between Monday and Thursday nobody ran an experiment. Nobody discovered a toxicity, published a case series, or learned anything about the molecule that was not already known. What happened was that a document was signed, or a notice appeared in a government gazette, or a committee met and voted. The molecule was not reclassified by nature. It was reclassified by an administrator.

This is the single most useful thing to understand about drug control, and it is routinely lost. People talk about scheduling as though it were a property of a substance, discoverable in principle by looking hard enough at the substance. It is not. It is a decision made by named institutions under statutory procedures, informed by evidence but not determined by it, and shaped throughout by history, politics, timing and precedent. Everything that follows — the oddities, the inconsistencies, the maps that disagree at a border — falls out of that one fact.

The procedure, and what it actually asks

The American statute is the most-copied template, so it is worth reading closely. Scheduling begins on the initiative of the enforcement authority, or on petition from any interested party — a manufacturer, a medical association, a member of the public. The authority must then request a scientific and medical evaluation and a recommendation from the health department, and that evaluation carries genuine weight: if the health department recommends against control on scientific and medical grounds, the enforcement authority may not proceed 1.

The criteria themselves are eight in number, and only some of them are about the molecule. The statute asks about actual and relative potential for abuse; the scientific evidence of pharmacological effect; the state of current scientific knowledge; the history and current pattern of abuse; the scope, duration and significance of that abuse; the risk to the public health; psychic or physiological dependence liability; and whether the substance is an immediate precursor of something already controlled 1. Four of those eight are questions about human behaviour and social circumstance rather than pharmacology. Two of them — history and current pattern of abuse — are, strictly, questions about whether people have already got hold of it.

And the central term is undefined. The statute is built around abuse potential without saying anywhere what abuse is. That is not sloppiness; it is a deliberate delegation, leaving the substance of the judgement to the agencies and the courts. But it means the criteria cannot function as a test that a molecule passes or fails. They function as a set of considerations a decision-maker must be seen to have weighed.

When speed matters more than certainty

Full scheduling is slow — evaluations, recommendations, notice, comment, sometimes years. In the early 1980s that pace collided with a market that had learned to move faster than paperwork, and legislatures responded by adding emergency powers. The American version permits temporary placement in the most restrictive schedule where the authority finds it necessary to avoid an imminent hazard to public safety, on a much shorter procedure and for a limited period, extendable while the full assessment runs 1. Equivalent mechanisms exist across Europe and elsewhere under different names.

Emergency scheduling is honest about what it is. It does not claim to have established anything. It says: we think this is dangerous enough to stop now and find out properly afterwards. That is a defensible position for a public authority to take, and it is also an unambiguous admission that a substance's legal status can be settled well before its pharmacology is.

The moment chemistry became a legal category

Through the early 1980s legislators watched the same sequence repeat: a controlled substance, a structurally related compound not on the list, a market, a scheduling action, another related compound. Lists are enumerations, and an enumeration can always be stepped around. The response, in the United States, was the analogue provision of 1986.

Its construction is worth spelling out because it is genuinely novel as legal drafting. A controlled substance analogue is defined as a substance whose chemical structure is substantially similar to that of a substance in the two most restrictive schedules, and which either has, or is represented or intended to have, a substantially similar or greater effect on the central nervous system. Such a substance is then treated, to the extent it is intended for human consumption, as though it were itself in the most restrictive schedule 2. No listing, no notice, no gazette. The status attaches by operation of the definition.

Substantially similar is nowhere defined, and that is the design rather than a defect in it. A precise definition would have created a precise edge, and a precise edge is exactly the thing the provision existed to abolish. The cost has been decades of litigation in which expert chemists testify in opposite directions about the same pair of molecules, and courts have repeatedly had to grapple with whether the standard gives adequate notice of what is prohibited. The practical upshot for anybody reading a schedule is straightforward: in systems with an analogue provision, the published list is a floor, not a boundary, and a compound's absence from it settles nothing on its own.

The United Kingdom reached similar ground by a different route. Its 1971 statute had long used generic definitions — descriptions of whole chemical families by their common scaffold and permitted substitutions, so that a single provision captures compounds nobody has yet made. Then in 2016 Parliament inverted the logic entirely. Rather than defining what is prohibited, the Psychoactive Substances Act prohibits the production, supply, import and export of any substance capable of producing a psychoactive effect in a person who consumes it, and then exempts a defined list — medicines, alcohol, tobacco, caffeine, food 3. The default flipped from permitted-unless-listed to prohibited-unless-exempted, which is about as complete a reversal of a regulatory philosophy as one can find.

Three registers, not one

A further source of confusion is that people speak of a compound being banned as though there were a single list somewhere. There are at least three distinct registers operating in parallel, they have different authors and different consequences, and a substance can appear on any combination of them.

The first is drug-control law proper: the schedules described above, backed by criminal sanction, concerned with abuse and dependence. The second is medicines law, which is not about abuse at all — it governs whether something may be placed on the market or administered as a medicine, and its instrument is the marketing authorisation rather than the schedule. Most biologically interesting peptides have no relationship whatever with the first register; they are entirely a medicines-law question, and to describe such a compound as controlled or uncontrolled is to ask the wrong question of the wrong statute. The third register is not law at all. Sporting bodies maintain their own prohibited lists, revised annually, which include whole classes of peptide hormones and growth factors. Those lists have no criminal force, are drafted by a non-governmental organisation, and can nevertheless end a career.

One last asymmetry is worth noting, because it shapes the lists over time. The statutory procedures run in both directions on paper — substances can be removed or moved to a less restrictive schedule as readily as added — but in practice removal is vanishingly rare. Adding a substance requires only that a hazard be plausible; removing one requires somebody with standing to petition, resources to assemble an evidential case, and an institution willing to be seen relaxing a control 1. The incentives are wildly asymmetric, and the consequence is a set of lists that grows almost monotonically. Whatever was added during a panic in 1986 is, in most jurisdictions, still there.

Why the lists do not match the evidence, or each other

If classification were an evidential exercise, the resulting order would broadly track measured harm. It has been tested, and it does not. A group of researchers built a nine-category matrix of harm — physical damage, dependence liability, and social harm, each broken down further — and had it applied to twenty substances by two independent expert panels. The rankings that emerged correlated poorly with the legal classification then in force, with several legal drugs placing above substances in higher-penalty classes 4. A larger study three years later used a formal multicriteria decision analysis across sixteen harm criteria and reached the same conclusion more sharply still, with alcohol scoring as the most harmful substance overall once harm to others was included 5.

Neither study argued that classification should simply be replaced by a harm score. Both made the narrower point this article is making: a legal classification is not a measurement, and treating it as one produces confusion in both directions. It leads people to assume that anything unlisted is safe, and to assume that anything listed has been shown to be uniquely dangerous. Neither inference holds.

The divergence between countries has the same root. The international conventions of 1961 and 1971 establish a common floor: parties agree to control the substances in the treaty annexes. But a floor is not a ceiling, and every signatory is free to add above it. Each does so according to its own history, its own prompting incidents, its own committees, its own emergency actions taken during whichever local panic happened to arrive first. The result is a patchwork in which a compound may be squarely controlled in one member state, unlisted in the one beside it, and captured in a third not by name but by a generic definition that never mentions it. Borders are where the arbitrariness becomes visible, and it is visible there because it was there all along.

None of this is an argument that the lists lack authority. They have complete authority; they are law, and a court applying them will not be interested in whether their internal logic is elegant. It is an argument for reading them as what they are — accumulated administrative decisions, each with a date and an author and an occasion — rather than as a table of natural properties that somebody merely wrote down.