The Congealed List — Statecraft

A pharmacist around 1900

In 1898 the German chemical concern Bayer brought a new cough remedy onto the market under a name the company itself had coined: Heroin. The substance was marketed as a powerful, supposedly non-addictive successor to morphine, suitable for the treatment of cough and respiratory complaints, including in children. In the same period the soft drink an American pharmacist had compounded a few years earlier contained an extract of coca leaves, and that cocaine component was removed from the recipe only around 1903. Whoever entered a European or American pharmacy in those years could buy laudanum over the counter, a tincture of opium on an alcoholic base, freely available for pain, diarrhoea and insomnia, and administered on a large scale to infants to keep them quiet.
This is no story about a dark period of ignorance that the modern state has happily left behind. It is the starting point of a puzzlement. The substances that today stand in the heaviest category of the Dutch Opium Act (Opiumwet), with enforcement, investigation and imprisonment as their endpoint, circulated a hundred and twenty years ago as medicine and pleasure substance without any special legal status. The substance that the same Opium Act does not know at all, alcohol, stood on the shelf then and stands there still. Between those two states lies no linear progress of scientific insight towards rational policy. There lies a succession of regimes, each with its own claim to reasonableness, each built on the congealed outcome of the previous round.
The historical variation is no anecdote. It is the instrument through which thinking detaches itself from the assumption that the present classification is self-evident. The moment it becomes visible that the same society has treated the same substance successively as medicine, as danger and as object of enforcement, the naturalness drops out of the categories with which we now work. What remains is a question the policy itself does not ask: why does this substance stand on this list, and can that answer be reconstructed from the criteria the policy claims to apply? This Compass argues that the answer is negative, that the inability to account for the classification is no incident but a structural feature, and that this structural feature is the same dissociation described elsewhere in this corpus at organisational and state level, here applied to a domain in which the congealed category has taken the form of a statutory list.
The international architecture that carries the present classification is younger than one is inclined to think. The first treaty attempt to curb the trade in opium and cocaine was the International Opium Convention of The Hague of 1912, born not of a public-health analysis but of a complex of trade-policy and colonial interests around the Asian opium market.¹ The modern worldwide prohibition structure dates from the Single Convention on Narcotic Drugs of the United Nations of 1961, which introduced a system of schedules into which substances are classified and from which national legislation, the Dutch Opium Act included, derives its categories.² The decisive point is that these treaties did not find the classification but made it, at a moment, under a particular constellation of interests, and that all later policy formation has treated that classification as a given rather than as a product.

The test the policy does not carry out

When a classification claims to rest on the harmfulness of substances, a test imposes itself: lay the substances side by side on the dimensions that rationally matter, and see whether the classification can be reconstructed from that ranking. The dimensions are no mystery. They concern the harm a substance does to the user, physically and psychologically, the addiction potential, and the harm to third parties and to society, from violence and traffic accidents to care costs and social disruption. This is precisely the test that was published in 2010 in The Lancet by a panel led by the British psychopharmacologist David Nutt, which scored twenty substances on sixteen harm criteria, nine for harm to the user and seven for harm to others, with the criteria weighted by relative weight.³
The outcome of that exercise is well known and is rarely taken in its full implication. When harm to the user and harm to others are added together, alcohol is the most harmful substance of the twenty, more harmful than heroin and crack. Alcohol scores almost three times as high as cocaine or tobacco, and ecstasy comes in at about an eighth of the harm score of alcohol. Heroin, crack and methamphetamine are the most harmful for the individual who uses them, but once harm to the environment is weighed in, alcohol shifts to the top because the social footprint of alcohol, in violence, accidents, family damage and care burden, exceeds that of almost every other substance.⁴ This is no marginal finding of one idiosyncratic researcher. The ranking has since been repeated and confirmed in different jurisdictions, with the same fundamental structure each time: the statutory classification and the harm ranking do not run in parallel.
For the Dutch reader the relevant study is not the British one but the Dutch one, and that one is older and more explicit than one would expect. In 2009 the National Institute for Public Health and the Environment (RIVM) published a report, Ranking van drugs, in which a panel of nineteen experts ranked nineteen substances, seventeen illegal drugs plus alcohol and tobacco, according to their harmfulness on the basis of the scientific state of the art.⁵ The report did something rarely so straightforwardly done by a national institute: it tested explicitly whether the statutory classification of drugs in the Netherlands corresponds with the scientifically grounded harm ranking. The finding was that the correspondence is missing. For legislative purposes, the researchers established, drugs are largely classified by addiction potential, and such classifications lack a scientific basis. Alcohol, tobacco, heroin and crack scored high, while psilocybin mushrooms, LSD and khat scored low, and cannabis and ecstasy fell in the middle range.
Five years later the test was repeated at European scale. Forty drug experts from across the European Union scored twenty substances again on sixteen criteria, weighing the criteria by what would hold on average across the Union. Alcohol, heroin and crack emerged as the most harmful substances, with weighted harm scores of seventy-two, fifty-five and fifty respectively on a scale up to a hundred. All other substances stayed below thirty-eight, making them far less harmful than alcohol. The conclusion of the European experts ran that drug policy should focus on the substances with the highest total harm, alcohol and tobacco included, and that substances such as cannabis and ecstasy deserve a lower priority and a lower statutory classification.⁶
The pattern that becomes visible here is no detail that can be polished away by refining the criteria. It is robust across panels, years and countries. Whoever ranks the substances on their harm profile and lays that ranking beside the statutory classification finds no correspondence but a break. The substance with the highest total harm is freely available, taxed by excise and culturally anchored. The second substance in social harm, tobacco, is likewise freely available, albeit increasingly discouraged. Substances that cause a fraction of that harm stand in the heaviest prohibition category. The classification cannot be reconstructed from the criteria on which it claims to rest. That is, in the most literal sense of the word, an unaccountable classification: a classification that cannot give an account of itself.

Where the reasoning points, and where it does not stop

First principles lead here to a conclusion the policy would rather not make explicit. If the separate treatment of substances rests on their harmfulness, and the harm ranking does not bear out the separate treatment, then the ground for the differentiation falls away. And the moment the ground for differentiation falls away, there is no principled defensible distinction between the substance one places freely on the shelf and the substance for which one prosecutes. The reasoning then points to liberalisation, not as an ideological preference but as logical endpoint: whoever sells alcohol and prosecutes cannabis applies a distinction that on inquiry does not exist, and the only consistent way out of that incoherence seems to be lifting the prohibition on the substances that do less harm than the substance he permits.
Here, however, the reasoning must be pulled back, for at this point the indifferent plea for liberalisation makes the same mistake as the regime it criticises. Liberalisation as endpoint assumes that the absence of a rational ground for the present distinction automatically supplies a rational ground for release. That does not follow. What first principles really deliver is no prescription for a particular regime, but a demand on every regime: that it be reconstructible from its own criteria. The absent ratio dissolves the existing differentiation, but it does not prescribe that all substances should henceforth lie on the same shelf. It prescribes that the treatment of a substance must stand in proportion to its harm profile, consistently applied. And consistent application of the harm profile does not lead to a free market in everything. It leads to a regime in which the most heavily harmful substances, alcohol included, are most heavily controlled, and in which the more lightly harmful substances receive a lighter regime that reflects their actual risk.
The difference between those two endpoints is not academic. A non-dissociated regime can mean for one substance that it is supplied under strict medical control, for another that it is regulated and available with quality control and excise, and for a third that it is discouraged or restricted, but each time on the basis of its actual harm and not on the basis of the accident of its historical status. The common feature is not the degree of freedom but reconstructibility: in a non-dissociated regime one can point to, for every substance, which criterion leads to which treatment, and that derivation can withstand the question why a comparably harmful substance receives a comparable treatment. That is a higher and at the same time more modest demand than liberalisation. Higher, because it also brings the established substances under the test and does not spare alcohol. More modest, because it prescribes no particular regime, only the duty to account.
The stake of this Compass has therefore shifted from a position on drugs to a diagnosis of a regime. The question is no longer whether substances should be free. The question is why a state that bases its classification on harmfulness maintains a classification that contradicts its own harm research, and what it says about the nature of that regime that it can carry this contradiction for decades without resolving it.

How it works now: the congealed list in practice

The Dutch implementation of drug policy is a case apart, and precisely for that reason a sharp self-portrait. Since the revision of 1976 the Opium Act has two schedules: Schedule I for substances with, in the statutory formulation, an unacceptable risk, and Schedule II for cannabis and related products, for which a lighter regime applies.⁷ That division is in itself already an acknowledgement that differentiation by harmfulness is possible and desirable, for it rests on the insight of the commissions of the time that cannabis has a different risk profile from heroin. But the division has stalled halfway. It draws one dividing line, between cannabis and the rest, and leaves the rest of the incoherence untouched, the most striking blind spot being that the most harmful substance of all falls outside both schedules because it stands under an entirely different statutory framework.
The toleration model (gedoogmodel) built on this division is the Dutch contribution to the world history of dissociation. Coffee shops are allowed to sell cannabis to the consumer without criminal prosecution, provided they observe a set of conditions. But the purchase and the cultivation, the supply of the front door via the back door, have remained illegal throughout. The state thereby compelled an entire branch of industry to a daily illegal act in order to make a tolerated act possible. This is no pragmatic compromise that came about by accident. It is the institutional form of a state that admits a reality it does not wish to legitimise, and that does not solve the tension between admitting and legitimising but displaces it to the back door, where it has sat for decades as organised crime.
Only in 2025 has the state begun to close that back door, and the way it does so lays the dissociation bare. On 7 April 2025 the experimental phase of the Closed Coffee-Shop Chain Experiment began, in which coffee shops in ten participating municipalities may sell only regulated, quality-controlled cannabis sourced from specially licensed growers.⁸ The experiment must, in the words of its own explanatory note, investigate whether regulated production, distribution and sale of cannabis is possible and what its effects are on public health, safety and crime. The trial lasts four years. That a state allows four years of research to establish whether legal supply is possible of a product it has tolerated at the front door since 1976 is the purest illustration of the congealed outcome as new given: the outcome of an earlier policy round, the toleration model with its illegal back door, is treated as a given fact whose regulation constitutes a new and uncertain experiment, instead of as a self-imposed contradiction the state could always have lifted.
The fine details confirm the pattern. Within the experiment, enforcement on illegal hashish had to be postponed in the first months because there was insufficient regulated hashish available to supply the coffee shops and a shortage would have led to illegal street dealing.⁹ In other words: even within the regulated experiment the state had to tolerate temporarily what it had come to regulate, because the regulated chain could not yet carry the reality that the illegal chain had carried for years. The experiment reproduces in miniature the structure it claims to overcome. This is the Blue grammar in its purest form, with baseline measurements, process evaluations and track-and-trace systems around a phenomenon that society has allowed in practice for half a century.
The dissociation that the toleration model institutionalises is no incongruity without victims. The refusal to regulate the back door sustained for decades an illegal cultivation economy with its own violence, exploitation and power theft, and the same underlying pattern, a prohibited market alongside a demand that does not disappear, has produced in the broader hard-drug trade a shadow economy that reaches into the heart of the rule of law. In September 2019 the lawyer Derk Wiersum was assassinated, counsel of the chief witness in the Marengo trial, the first time a defence lawyer was murdered in the Netherlands. In July 2021 the crime journalist Peter R. de Vries was shot dead, confidant of that same chief witness. The violence no longer targeted the underworld alone, but the core actors of the administration of justice, the bar and journalism.¹⁰ And the response was once again the Blue reflex. After the Wiersum murder the cabinet allocated a hundred million euros, among other things to step up the fight against drug criminals, while a leading legal psychologist called that hardening a Pavlovian reaction: the harder the government fights the trade, the riskier and therefore the more lucrative that trade becomes, and the larger the money flows that feed the violence.¹¹ The fight produces the profit that makes the fight necessary.
The Dutch case is no exception but an instance. Other states have wrestled with the same incoherence and have come to other arrangements, and the comparison between those arrangements shows precisely where the dissociation sits and where it is broken. Three cases deserve attention, not because any one of them supplies the best model, but because they together demonstrate that the decisive difference lies not in the degree of freedom a regime allows, but in the degree to which the regime can account for itself and can be embedded.

Portugal: the shift of domain

In 2001 Portugal took the step that breaks most fundamentally with the criminal-law logic. With Law 30/2000, in force from 1 July 2001, the possession and use of all drugs for personal use was decriminalised, not legalised.¹² The distinction is essential and is often missed in the Dutch discussion. The substances remained illegal; what changed was the nature of the offence and the domain in which it is handled. Whoever is found with a quantity for personal use commits an administrative offence, not a crime, and is not brought before a court but referred to a Commission for the Dissuasion of Drug Addiction, a multidisciplinary panel of usually a lawyer and two experts from the medical or social sphere.
The analytically decisive feature of the Portuguese model is not the decriminalisation in itself, but the fact that the whole system was placed under the Ministry of Health instead of under Justice or the Interior.¹³ That is a shift of domain, and in the language of this corpus it is a de-dissociation. The substance is brought back under the regime that fits its actual nature. A phenomenon that is in essence a public-health question is institutionally placed with public health, and the criminal-law machinery is withdrawn to where it does belong, with the trade and the production. The Portuguese regime is therefore not liberal in the sense that it releases the substances, but coherent in the sense that it makes the treatment of the user fit the diagnosis of the problem.
About the outcomes of the Portuguese model an extensive and sometimes politicised literature has arisen, and honesty requires saying that causal attribution is difficult, partly because reliable prevalence figures from before 2001 are missing.¹⁴ What is firm is that the feared explosion of use has not materialised, that the share of problematic users among persons referred to the commissions has fallen, and that use among schoolchildren has lain below the European average for two decades.¹⁵ More important than the precise effect is an observation about time: the Portuguese model needed almost a decade to make its positive effects visible. That slowness is no weakness of the model but a characteristic of the kind of intervention it is, and it becomes crucial as soon as one lays the Portuguese case beside the Canadian.

Switzerland: embedded differentiation

Switzerland supplies the sharpest contrast with the Dutch half-heartedness, precisely because it made explicit the differentiation that the Netherlands does not dare complete and then embedded it through the heaviest conceivable instrument of legitimation. After the heroin epidemic that concentrated around the Platzspitz park in Zurich in the 1980s and 1990s, Switzerland developed the so-called four-pillar policy: prevention, treatment, harm reduction and law enforcement, alongside each other and in coherence.¹⁶ Within that architecture, in 1994, the most controversial innovation came about, the medical supply of heroin on prescription to chronic, treatment-resistant addicts.
The results of that heroin supply are well documented. The physical and mental health of participants improved, their social reintegration increased, their use of illegal drugs fell significantly and their illegal activities decreased sharply, while there was no increase in new heroin users, a fear the opponents had voiced and one the facts refuted.¹⁷ But the decisive point for this argument is not the efficacy. It is the embedding. The heroin supply did not remain an experiment that could be abolished at the first political counter-wind. A prohibitionist movement forced a national referendum in 1997 against the four-pillar policy, and seventy per cent of the Swiss voted to maintain the policy. In a second referendum, on 30 November 2008, the population voted by an overwhelming majority to anchor the heroin supply permanently in the narcotics legislation.¹⁸
In that same referendum of 30 November 2008, however, the Swiss voter rejected a proposal to replace the cannabis market with a regulated market with tolerated production and points of sale.¹⁹ This is no incoherence. It is the opposite. Switzerland in one vote legitimised the medical heroin supply and rejected cannabis liberalisation, and with that established a differentiated regime in which every component has an explicit democratic ground. The substance with the heaviest addiction profile received the heaviest, medically controlled regulation; the attempt to release a lighter substance was not blocked by an administrative reflex but by a pronounced majority. What distinguishes the Swiss case is not that it made the right choice, but that it made the choice in a way that can carry it. The Yellow reality, the political and moral struggle over what should happen with these substances, was not evaded and not hidden in Blue procedure, but fought out in the arena where legitimacy is conferred, and that is why the regime holds.

Canada: divergence as symptom

It is tempting to place a third successful model alongside Portugal and Switzerland, and Canada is often advanced in that role on account of the federal cannabis legalisation in 2018. That presentation is misleading, and that is precisely why Canada is the most instructive case. Canada delivers no model but a natural experiment in federal divergence, in which three administrative layers under one and the same overdose crisis move in opposite directions. Whoever presents Canada as a best case misses what is really to be seen: that the moment the shared rational ground for one regime is missing, the outcome does not converge towards the reasonable optimum, but falls apart along political fault lines.
At federal level Ottawa is the statutory enabler of rationalisation. The Cannabis Act of 2018 legalised cannabis for recreational use at national scale, and it was the federal health authority Health Canada that granted in January 2023 the exemption to the federal drug law that allowed the province of British Columbia to decriminalise possession of small quantities of hard drugs. The federal track creates the legal room, but it delegates the execution to the provinces, and at provincial level the paths diverge radically.
British Columbia initially chose the most pronounced harm-reduction line. From January 2023 adults could carry up to two and a half grams of hard drugs, including heroin, fentanyl, cocaine and methamphetamine, for personal use without criminal prosecution. But the regime did not hold. Under pressure from public indignation at visible drug use in parks, hospitals and public transport, and with an election in sight, the province asked the federal government in spring 2024 to make possession in public spaces a criminal offence again, which was granted in May 2024.²⁰ In September 2024 the province moreover announced that it would extend compulsory care and open high-security facilities. In January 2026 British Columbia chose not to extend the decriminalisation pilot.²¹ Within one electoral cycle the province moved from one paradigm to another.
The contrast with Portugal is diagnostic here. The Portuguese model needed almost a decade to work; British Columbia got three years, cut through by continuous changes of course between decriminalisation and renewed criminalisation. This is no refutation of the idea behind decriminalisation. It is a failure of embedding. An intervention that by definition needs years to take effect was broken off before it could work, because the Yellow reality of visible disturbance and electoral pressure overwhelmed the Blue rationale of harm reduction. What in this corpus is called the embedding test, the question of what remains standing when no one is thinking about the intervention any longer, has been mercilessly answered in British Columbia: nothing, for the intervention was never anchored in anything more durable than a political conjuncture.
Alberta chose the opposite pole. With the Compassionate Intervention Act, which received royal assent on 15 May 2025, the province created a track in which a family member, guardian, care professional or police officer can have someone with a severe addiction who poses a danger to himself or others compulsorily placed in treatment, with a quasi-judicial commission as gatekeeper and an investment of a hundred and eighty million dollars in compulsory-treatment centres in Edmonton and Calgary.²² It is the first law of this type for adults in Canada, a parallel system alongside the existing mental-health legislation. Alberta claims results: a fall of thirty-eight per cent in opioid-related deaths in 2024 against 2023, more than double the national average fall of seventeen per cent.²³ The medical literature notes alongside it that the compulsory-treatment track has insufficient evidence of effectiveness and raises substantial ethical and legal objections.²⁴
The decisive insight from the Canadian case lies in the composition of these three movements. Ottawa rationalises legally, British Columbia swings from harm reduction to coercion, Alberta institutionalises coercion as principle, and all this under the same crisis and with an appeal to the same figures. British Columbia and Alberta both refer to mortality figures, to the evidence, to public health, and arrive at opposing regimes. That painfully undermines the naive first-order conclusion that first principles lead by themselves to one rational endpoint, or that towards liberalisation. The Blue grammar of evidence and mortality figures is deployed in every jurisdiction to legitimise a Yellow reality that is, per jurisdiction, opposed. The figures do not choose; politics chooses, and uses the figures as cover.

Two opioids, one schedule

No pair of substances tests this argument more sharply than oxycodone and fentanyl, and it begins with a fact that carries the whole incoherence within itself. Oxycodone, fentanyl and heroin all stand on Schedule I of the Opium Act, the schedule of the unacceptable risk, and they are all three opioids, pharmacologically the same family of substances that act on the same receptors.²⁵ Yet in 2021 around four hundred and fifty thousand people received oxycodone from their general practitioner, and fentanyl is administered daily in the operating theatre and in palliative care, while heroin is the archetype of the prohibited drug for which one is investigated and prosecuted. Here the dissociation sits not between schedules, as with cannabis and heroin, and also not between the prohibition and the open shelf, as with heroin and alcohol. Here it sits within the heaviest category itself. The schedule of unacceptable risk contains at once the substance the general practitioner prescribes hundreds of thousands of times a year and the substance society treats as evil par excellence, and the regime cannot account for the difference between the two on grounds of harm, for the harm profile of the two opioids lies in the same order. The distinction rests not on the molecule but on the route through which it enters: the doctor’s prescription pad as against the dealer’s bag.
At this point the naive reading of this Compass would say: medicalise everything, then, bring every substance under the regime of the prescription, and the incoherence disappears. The American opioid crisis closes that exit, and it closes it in a way that should bring everyone who supposes the medical channel to be naturally safe to their senses. The deadliest drug crisis in modern history unfolded in three waves. The first began in the late 1990s with prescription opioids, led by the powerful oxycodone variant introduced in 1996 that the manufacturer marketed aggressively with the claim that the addiction risk was low. The second wave, from around 2010, was heroin, to which users turned when prescription opioids became more expensive and harder to obtain. The third wave, from 2013, was fentanyl. Since the start of the epidemic more than seven hundred and twenty-seven thousand people in the United States have died of an opioid overdose.²⁶ The decisive point is where that catastrophe began. Not on the street, but at the prescription counter, in the most Blue, the most regulated, the most medically legitimised channel imaginable, captured by a commercial interest that presented itself as clinical practice. Medicalisation is no de-dissociation when the medical channel itself conceals a Yellow reality behind a Blue façade.
The structural turn that follows also closes the opposite exit, that of harder prohibition. When the United States began to close down the prescription channel, with re-examination of prescribing practice and reformulation of the best-known oxycodone variant, the number of deaths did not fall. It tripled. Between 2010 and 2023 the number of opioids prescribed per hundred inhabitants fell sharply, while the total number of opioid overdose deaths in the same period more than tripled, with fentanyl taking an ever-larger share.²⁷ This is the working of what has come to be called the iron law of prohibition: as enforcement pressure rises, the market shifts to more potent and more easily concealable forms, because higher potency per unit of volume pays off the moment concealment becomes the problem.²⁸ In the prohibition era that law drove the market from heroin to fentanyl and is now driving it further to the nitazenes. It is the same law that, in the Dutch hard-drug trade, drove the money flows and the violence up as the fight against it hardened, to the point of the assassination of a lawyer and a journalist. Fentanyl is fifty to a hundred times as potent as morphine, and a quantity of two milligrams can be lethal. The regime points to the fentanyl deaths as proof that drugs must be fought with force, while the dominance of fentanyl is itself the product of that fight. That is the congealed outcome that feeds the next round: every cycle treats the lethality produced by the enforcement of the previous cycle as a given fact justifying still more enforcement.
With that, fentanyl closes the exit that oxycodone left open. Fentanyl really does belong to the most harmful substances per unit, and its harm profile, consistently applied, justifies heavy regulation. First principles do not prescribe a free market in fentanyl. This is precisely the landing of paragraphs eleven and twelve, now made concrete on the hardest material. The absent ratio dissolves the existing distinction, the question of why oxycodone is medicine and heroin a crime while both opioids stand on Schedule I, but it prescribes no regime whatever. It prescribes that the treatment of every substance be reconstructible from its actual harm, consistently applied. And oxycodone and fentanyl show that this consistency cuts both ways: it would bring alcohol under heavier regulation and it would keep fentanyl heavily regulated, but each time on the basis of harm and not on the basis of the channel through which the molecule reached the user. The opioids are the case in which both the libertarian and the technocratic reading fail, and only reconstructibility remains standing.
For the Dutch reader this story is not in the first place a present but a known future. Europe did not have the American fentanyl wave, and the reason is telling: precisely the absence of an American culture of mass medical opioid prescribing has kept the risk group for synthetic opioids in Europe limited for a long time to heroin users and buyers of counterfeit medicines.²⁸ But the Afghan opium ban that the Taliban announced in 2022 sets, with the documented precedent of the 2000-01 ban that drove synthetic opioids into the Baltic states, the conditions for what the EU drugs agency calls a perfect storm. The heroin supply in Europe is for now still stable, but as soon as supply contracts, the market fills with more potent synthetic opioids, and in the last years the nitazenes, a group of very powerful substances, have already overtaken the fentanyl derivatives among the new synthetic opioids appearing in Europe.²⁹ These substances appear not only in heroin but also in counterfeit oxycodone and benzodiazepine pills sold online, and with that in a risk group without any opioid tolerance: not the stereotypical addict, but the pain patient who no longer received a prescription or could not bear its price, and who dies of a pill indistinguishable from a genuine medicine. In the Netherlands counterfeit oxycodone containing the nitazene isotonitazepyne has already been found, and the documented presence of Mexican networks in the Dutch synthetic-drugs industry makes a transfer of production know-how to synthetic opioids not unthinkable.³⁰
The regime’s response to this known future is the same dissociation, now in advance. The EU drugs agency calls for vigilance and preparation, develops warning systems and threat analyses and speaks of acting today and anticipating tomorrow.³¹ That is Blue grammar over a Yellow reality it does not name. The predictable outcome of curbing a supply without addressing the demand is the shift to a more lethal substitute, an outcome demonstrated twice in the Baltic states, in 2001 and again after 2019. A regime that could read its own history would treat the coming shift to synthetic opioids as the endogenous product of its own supply-oriented enforcement. Instead it treats every escalation as a new external threat that demands more of precisely the enforcement that produced it. That is blindness to a known future, and it is the same blindness that institutionalises the congealed list: the inability to see that the catastrophe of the next round is the design of the current.

Back to the line

The diagnosis that emerges from the foregoing is not that society has chosen the wrong regime. It is that the regime, in its Dutch and in its international guises, cannot explain why it does what it does, and that this inability is no accidental shortcoming but a structural feature with the same shape as the dissociation described elsewhere in this corpus. The statutory classification of substances is a congealed outcome. It was fixed at a moment in history, under a constellation of trade-political, colonial and moral interests, and since then treated by every policy round as a given fact one can tinker with, not as a product one can revise. The substance stands on the list because it stands on the list. That is the congealed outcome as new given in its most literal appearance: not a metaphor from organisation studies but a statutory text with two columns, whose content cannot be derived from its own criteria, and whose heaviest column contains two opioids that the state at the same time prescribes and combats.
The mechanism that sustains this contradiction can be read, in the language of the change colours, as a Yellow reality that has clothed itself in a Blue grammar. The classification of substances is at its core a Yellow phenomenon: it is the product of moral panic, of international power relations, of path dependency and of the electoral value of toughness against drugs. But it presents itself in Blue language, as a rational classification by risk, with schedules, harm assessments and scientific committees. The danger of that configuration is precisely what this corpus has named elsewhere as the most dangerous institutional state: a Blue institution structurally unable to address a Yellow reality, because the Blue vocabulary makes the Yellow power question invisible. As long as the debate is conducted in terms of harm rankings and evidence, the real question, the one about power, morality and legitimacy, remains out of view, and the regime can continue its incoherence indefinitely because it evades the only question that could resolve it.
Here it also becomes visible why the strategic triangle reads the cases differently. The drug regime in its dominant, prohibitionist form maximises one corner, political legitimacy in the specific guise of moral and electoral toughness, at the expense of the other two. Public value, measured as the actual reduction of harm to users and society, is sacrificed, as the harm rankings show. Operational capacity is wasted on the enforcement of a distinction that has no harm ground, while the most heavily harmful substance falls outside enforcement. The Swiss case distinguishes itself in that it did not play the three corners against each other but interrogated the legitimacy corner explicitly, through the referendum, and could thereby build a regime in which public value and political legitimacy did not work against each other. British Columbia failed because it chased public value without having acquired the legitimacy corner, and at the first collision between the two it let public value fall.
It is tempting to draw from this analysis the conclusion that the solution is a Swiss referendum, or a Portuguese shift of domain, or some other importation. That conclusion would repeat the mistake this Compass has dissected in paragraphs ten and eleven. The point is not which regime is the right one. The point is that no regime can be embedded as long as the choice between regimes is Yellow and the accounting Blue, for then the real decision remains unspoken and therefore unprotected against the next political wind. Switzerland did not embed because it chose medical heroin supply, but because it carried that choice through the Yellow arena and had it confirmed there. British Columbia failed not because decriminalisation was wrong, but because it had never won the Yellow decision and had cast it in Blue pilot language, where it proved defenceless.
Behind every property of the dissociated regime this Compass has described stands a person who pays the bill. The congealed list, the Yellow reality behind the Blue façade, the failed embedding: these are not abstractions but mechanisms with a residue, and the residue is a life. It is the user who dies of a dose he could not see, taste or smell, because the prohibition drove the substance into an unregulated market. It is the pain patient who, when the pendulum swung from over-prescription to sudden withdrawal, ended up on the street or online and there bore the price of both errors. It is the addict who got a criminal record instead of a treatment plan, because the substance stood in the wrong domain. It is the lawyer and the journalist who were murdered by the economy the regime sustains, and the young person recruited into the shadow economy because that was where the money was. This is whom it affects, and it strikes hardest among those with the least voice in the Blue grammar in which the debate is conducted. A diagnosis that names this is therefore not without obligation. It calls for no particular policy, but it does ask the regime to give up a self-description, the one in which the list is about harm, for that self-description is paid in the lives enumerated here.
The substances behind the pharmacist’s counter around 1900 have not become more harmful or more innocent in the hundred and twenty years that followed. What changed was not the substance but the list, and the list changed not because the harm changed but because the power and the morality changed, while the policy kept insisting that it was about harm. A state that publishes its own ranking of harm and refutes its own classification with it, that prescribes and combats two opioids on the same list at the same time, and that meets the predictable next round of its own supply policy as an external threat, tells something about itself that reaches further than the drug file. It shows that it is capable of knowing, documenting and enforcing an evident error at the same time. That is no drug problem. That is the dissociation, legible on a list with two columns.

Notes

Colophon

The Congealed List · Drug policy, the absent ratio, and the regime that cannot account for itself Statecraft Compass · v0.3 · May 2026

Author: Jacob Huibers Publisher: HOUSE OF VIRIDIAN OÜ · Tallinn · Lisbon Contact: jacob@statecraft.nl · statecraft.nl

Related series: Series III Dissociated Organisations (in compilation). Series IV The Dissociated Union (in preparation). The Dissociated Government. A Diagnosis (manuscript in preparation). Manuscript reference: De Richting van de Beweging: Interim Management in the Public Sector (manuscript in preparation).

International Opium Convention, The Hague, 23 January 1912. The treaty was the first multilateral instrument on drug control and arose from the international opium conferences that went back to the Shanghai Commission of 1909. ↩
Single Convention on Narcotic Drugs, United Nations, New York, 1961, as amended by the 1972 Protocol. The treaty introduced the system of schedules from which national legislation derives its classifications. ↩
Nutt DJ, King LA, Phillips LD, on behalf of the Independent Scientific Committee on Drugs. Drug harms in the UK: a multicriteria decision analysis. The Lancet 2010;376:1558-1565. ↩
Ibid. The modelling showed that heroin, crack and methamphetamine are the most harmful for the individual, while alcohol, heroin and crack are the most harmful for others; on adding both, alcohol is the most harmful substance, with a score almost three times that of cocaine or tobacco. ↩
Van Amsterdam JGC, Opperhuizen A, Koeter MWJ, Van Aerts LAGJM, Van den Brink W. Ranking van drugs. Een vergelijking van de schadelijkheid van drugs. RIVM, Bilthoven, 2009. See also Van Amsterdam J, Opperhuizen A, Koeter M, Van den Brink W. Ranking the harm of alcohol, tobacco and illicit drugs for the individual and the population. European Addiction Research 2010;16:202-207. ↩
Van Amsterdam J, Nutt D, Phillips L, Van den Brink W. European rating of drug harms. Journal of Psychopharmacology 2015;29(6):655-660. Compare Van Amsterdam J, Van den Brink W. The high harm score of alcohol. Time for drug policy to be revisited? Journal of Psychopharmacology 2013;27:248-255. ↩
Opium Act (Opiumwet), originally 1919, substantially revised in 1928 and 1976. The division between Schedule I (hard drugs with an unacceptable risk) and Schedule II (cannabis products) dates from the 1976 revision, arising from the work of the Hulsman and Baan commissions. ↩
Closed Coffee-Shop Chain Experiment Act (Parliamentary Papers 34.997). Start of the experimental phase on 7 April 2025; see Rijksoverheid.nl, Experiment gesloten coffeeshopketen (wietexperiment). Ten participating municipalities: Almere, Arnhem, Breda, Groningen, Heerlen, Hellevoetsluis/Voorne aan Zee, Maastricht, Nijmegen, Tilburg and Zaanstad. ↩
Parliamentary letter from the Minister of Justice and Security and the State Secretary for Youth, Prevention and Sport, 2 April 2025, on the progress of the Closed Coffee-Shop Chain Experiment. Enforcement on non-regulated hashish was temporarily suspended on account of supply shortage; the transitional period expired on 1 September 2025. ↩
The murder of the lawyer Derk Wiersum took place on 18 September 2019 in Amsterdam; he was counsel to the crown witness Nabil B. in the Marengo trial. Crime journalist Peter R. de Vries, confidant of the same crown witness, was shot on 6 July 2021 and died on 15 July 2021. See WODC, Justitiële verkenningen 2021, no. 4, on organised crime and subversion, which establishes that the violence targets the core actors of the administration of justice, the bar, the judiciary, law enforcement and journalism. ↩
After the Wiersum murder the cabinet made a hundred million euros available, among other things for the protection of judges and lawyers and for intensifying the fight against drug crime (NOS, July 2021). The characterisation of that hardening as a Pavlovian reaction, in which harder enforcement makes the trade riskier and more lucrative, is from the legal psychologist Peter van Koppen (NOS, July 2021). Compare the work of Pieter Tops and Jan Tromp on the undermining shadow economy and the recruitment of young perpetrators. ↩
Lei n.º 30/2000, in force from 1 July 2001. See EUDA (formerly EMCDDA), Drug Policy Profile: Portugal. ↩
The system of Comissões para a Dissuasão da Toxicodependência sits under the Ministry of Health, initially through the Instituto da Droga e da Toxicodependência (IDT), now through SICAD. ↩
Moury C, et al. Understanding successful policy innovation: the case of Portuguese drug policy. Addiction 2023. For the methodological limits on causal attribution, see also the discussion in the Portuguese policy literature on the absence of prevalence figures from before 2001. ↩
Transform Drug Policy Foundation, Drug decriminalisation in Portugal: setting the record straight. In 2018, ninety per cent of individual cases before the commissions showed no problematic use. ↩
Uchtenhagen A. Heroin-assisted treatment in Switzerland: a case study in policy change. Addiction 2010;105(1):29-37. For the broader context, see Csete J, From the Mountaintops: What the World Can Learn from Drug Policy Change in Switzerland, Open Society Foundations, 2010. ↩
Ibid.; see also the review of medical heroin supply, in which a retention rate of around seventy-six per cent over twelve months was established, alongside improvements in health and social reintegration and a sharp fall in illegal activities. ↩
The 1997 referendum, forced by the movement Jugend ohne Drogen, yielded a majority of about seventy per cent for maintaining the four-pillar policy. The referendum of 30 November 2008 permanently anchored the heroin supply in the narcotics legislation. ↩
At the same vote of 30 November 2008 a separate popular initiative to regulate the cannabis market was rejected. ↩
Health Canada granted in January 2023 an exemption under the Controlled Drugs and Substances Act for British Columbia. On 7 May 2024 the federal government approved the province’s request to make possession in public spaces a criminal offence again. ↩
In January 2026 British Columbia announced that it would not extend the decriminalisation pilot. Already in September 2024 the province had announced that it would extend compulsory care. ↩
Compassionate Intervention Act (Bill 53), Legislative Assembly of Alberta, royal assent 15 May 2025. 2025 budget: one hundred and eighty million dollars over three years for two centres of a hundred and fifty beds each in Edmonton and Calgary. ↩
Alberta reported for 2024 a fall of thirty-eight per cent in opioid-related deaths against 2023, against a national average of seventeen per cent. The attribution of this fall to the Compassionate Intervention Act has not been established; the act only took effect in May 2025, while the fall concerns the year 2024. ↩
Wilson, Larson, Colizza. Involuntary treatment for substance use: application of Kass’ ethical framework to Alberta’s Compassionate Intervention Act. Canadian Medical Association Journal 2025;197(40):E1346-E1348. The authors point to the absence of sufficient evidence of effectiveness of compulsory treatment. ↩
Oxycodone, fentanyl and heroin (diacetylmorphine) all stand on Schedule I of the Opium Act. Oxycodone and fentanyl are available on prescription as medicines; see Jellinek and the Drugs Information Line (Trimbos Institute), which place oxycodone explicitly as an opiate in the same family as heroin and morphine. The supply figure of around 450,000 persons in 2021 is taken from figures of the Stichting Farmaceutische Kengetallen; oxycodone supplies rose from around 70,000 (2003) to around 420,000 (2008) and further in the years that followed. ↩
The three-wave structure and mortality figures follow the Centers for Disease Control and Prevention (CDC) and the National Institute on Drug Abuse (NIDA). The first wave (prescription opioids) set in around 1999 and peaked in 2017; the second wave (heroin) from 2010, peak 2016; the third wave (synthetic opioids, mainly illegally manufactured fentanyl) from 2013. The powerful oxycodone variant (OxyContin) was introduced in 1996 and marketed aggressively with under-estimation of the addiction risk. The cumulative number of opioid overdose deaths since the start of the epidemic is estimated at more than 727,000. ↩
USAFacts, on the basis of CDC data: from 2010 to 2023 the number of opioids prescribed per hundred inhabitants fell sharply, while the total number of opioid overdose deaths in the same period more than tripled. In 2023 fentanyl was involved in around sixty-nine per cent of all overdose deaths. ↩
The iron law of prohibition, a concept going back to Richard Cowan (1986): as enforcement becomes stricter, the potency of the prohibited substance rises, because concealment favours a higher efficacy per unit of volume. The related concept is the balloon effect, in which enforcement does not remove the problem but displaces it. The observation that the absence of an American culture of mass opioid prescribing kept the European risk group limited is taken from Global Initiative against Transnational Organized Crime, The looming threat of synthetic opioids in Europe? (2024). ↩ ↩²
EUDA, EU drug market reporting on heroin and other opioids (2024) and the Call to action on new synthetic opioids (2024); Council of the European Union, Opioids: a new drug threat for Europe. The Afghan opium ban of the Taliban dates from 2022; the precedent of the 2000-01 ban drove synthetic opioids, among them fentanyl, into the Baltic markets. In the past five years nitazenes (benzimidazole opioids) have overtaken fentanyl derivatives among the newly reported synthetic opioids in the EU Early Warning System. In Estonia nitazenes were involved in around thirty-nine per cent of drug-related deaths in 2022. ↩
Drugs Information Line (Trimbos Institute), warning concerning counterfeit oxycodone containing the nitazene isotonitazepyne, repeated in 2025 and 2026 as still current. Global Initiative against Transnational Organized Crime, Highly potent synthetic opioids are already in Europe’s drug supply chains (2024), on the presence of Mexican networks in the Dutch synthetic-drugs industry and the possible transfer of production know-how, by analogy with the earlier industrialisation of methamphetamine production in the Netherlands. ↩
EUDA, Call to action. New synthetic opioids: European preparedness and response (2024), with the formulation about acting today and anticipating tomorrow. ↩