The Dominant Route

I. The Weekly Credit Card

In June 2019, a single sentence travelled further than any environmental statistic of its decade.

A research synthesis commissioned by WWF International, prepared by the consultancy Dalberg Advisors, and built on a meta-analysis of prior studies by the University of Newcastle in Australia offered the finding that the average person was ingesting, each week, by weight, the equivalent of a credit card's worth of plastic.¹

On 11 June the figure crossed the Atlantic on CNN International.² By the morning of 12 June, The Irish Times was carrying it on its environment page.³ Within forty-eight hours it had passed through the UK trade press.⁴ Within a month it was in parliamentary questions and broadcast lead-ins; within a season it was on the side of a Pret A Manger collection box in a motorway car park I happened through in late July, the WWF logo sun-bleached already, the words credit card still bright. I stood there with a coffee in one hand and my own card in the other and tried to do what the poster was asking. It is a thing you can do. You can hold the plastic you are supposedly eating. You can weigh it in your palm.

On its own terms, the sentence is unfair. It is a back-of-envelope — a synthesis of particle counts from water, shellfish, salt, beer, honey and sugar, converted to mass by assumptions about particle density, and then rendered as a single everyday object. The WWF report was upfront about the uncertainty; fact-checkers followed within months to say the figure was very likely inflated, and later commentary put the median human ingestion many orders of magnitude lower.⁵ None of that mattered. What mattered was that the sentence fit inside a pocket. It fit inside a wallet. It fit inside a mouth. A statistic, to travel, needs a small object in a large room — and this one had found exactly the right small object.

A good statistic is a small object in a large pocket. A large object, however true, will wait for a pocket that fits it.

What I did not know, standing in that car park in July 2019, was that the room I was standing in — the air between my face and the windscreens, the dust at the edges of the tarmac, the upholstery inside every parked car, the polyester inside every coat hung in every driver's back seat — was probably delivering a second dose of the same substance by a different route, and that the second dose may have been larger than the first, and that no instrument on earth that summer could count it at the size class that would eventually matter. There was no sentence yet for what the room was doing. The room would have to wait six more years for its sentence, and when the sentence came, it would arrive in a quieter voice, on a July afternoon in 2025, in a French peer-reviewed journal, with no credit card on its cover.

This is a report about the six years the room had to wait, and about the five other rooms in the last hundred years that waited in exactly the same way, for exactly the same reason. It is a report about a pattern with a name that feels, at first, too abstract to carry the weight of what it is describing: the pattern of the instrument shadow — the long interval between a hazard beginning and the microscope becoming fine enough to render it as a number that can be said aloud. We will come back to the credit card. The credit card has a second job to do, and the pattern has to be in the room before the Talisman can do it.

The question underneath the whole piece is this: how did the route we could picture come to define the microplastic story — and what does it say about us that we tend to solve only the problems our instruments are ready to measure?

It may be that the answer is bracing. It may also be that, once you have seen it before, it is the shape of a thing you can carry.

II. The Food Frame Was Correct, It Was Just Early

The strongest version of the case against this piece is not that the microplastic-in-food story was foolish. It is that the microplastic-in-food story was correct for the evidence it had, and that a decade of work inside that frame produced real regulatory wins that are still paying down exposure today — and that the later air evidence of 2025 does not retroactively invalidate any of them.

That version of the case deserves to be said in full, in the voice of the people who did the work.

The food frame landed the UK microbead ban in rinse-off personal care products under the Environmental Protection (Microbeads) (England) Regulations 2017.⁶ It underwrote the Plastic Packaging Tax that took effect on 1 April 2022 at £200 per tonne for packaging containing less than 30% recycled content.⁷ It was the substrate for the BPA shelf reset in infant bottles and the decade-long reformulation pressure that is now folding phthalates out of food-contact plasticisers.⁸ The canonical anchor for the food frame — Cox et al., 2019, in Environmental Science & Technology — estimated annual microplastic consumption from eight representative food groups, bottled and tap water, and an inhalation term, and identified bottled water as a dominant sub-route within the frame it was using. It was a careful paper. It was a careful decade.⁹

Crediting the instruments by name is not politeness. It is precision. The Plastic Packaging Tax is an instrument. The microbead ban is an instrument. The BPA shelf reset is an instrument. Each arrived because the food frame had the evidence, the rhetoric, the legal fulcrum, and — crucially — the number that could be written into statute and pinned to a product category. Each is a working demonstration that when we can count a thing and name its route, we can pull the lever that reduces it.

The food frame was the right instrument for the evidence in 2018, and the air evidence of 2025 does not retroactively invalidate a decade of correct work. It names a second problem the first instrument could not see.

That is the sentence a food-frame advocate signs. Everything that follows is about the second clause.

A map is not wrong when it stops at the coastline. It is wrong when it insists the coastline is the edge of the world.

Here is what the instruments could not yet do at scale across residential indoor air in 2018. The dominant microplastic analytical workhorse for the decade of the food frame was Fourier-transform infrared spectroscopy — FTIR — whose practical polymer-identification floor for routine particle-by-particle work sits in the low tens of micrometres for unattended throughput, and descends lower only with patient single-particle technique and accepting a steep drop in counts per hour. The sub-10 µm particles were not undetected in principle; the filters caught them, and the air-science groups already pointing at the room knew they were there. What could not be done was routine, sample-wide polymer identification of the 1–10 µm fraction of ordinary residential indoor air at enough throughput to produce a dose estimate a committee that writes policy could pin a number to. And a small cohort of researchers was already building the scaffolding of the air frame before the instrument had caught up. Stephanie Wright and Frank Kelly's 2017 review in Environmental Science & Technology was not a witness statement; it was a research-priority document that argued, in terms the field had not yet accepted, that airborne microplastic should be treated as a distinct exposure route demanding its own methods.²³ Ana Catarino and co-authors' 2018 work on household-fibre fallout during meals was not a bystander observation; it was an early claim that the domestic-fibre route might already be exceeding the food route at the per-meal scale, and it named the measurement gap explicitly.²⁴ Steve and Deonie Allen's 2019 paper in Nature Geoscience on atmospheric microplastic deposition in a remote Pyrenees catchment was not a nature note; it was the atmospheric anchor that forced the field to accept that plastic moved through air at distance, a precondition for treating indoor air as a reservoir rather than a background.²⁵ Wright, Catarino and the Allens were not the witnesses of the air frame. They were its architects, and they were building the frame for years before the microscope that could scale it arrived. And then in 2025, a team at Géosciences Environnement Toulouse, led by Jeroen Sonke, published an automated Raman pipeline with a stated limit of detection of 1 µm, and pointed it at the air in three residential apartments and two inter-city cars in France, sampled across seven and five separate runs respectively.¹⁰

The results are the reason this report exists.

Median residential concentration: 528 microplastic particles per cubic metre of indoor air, with an inter-quartile range that ran up past 2,487. Car cabin median: 2,238 particles per cubic metre, with variation that tracked how recently the cabin had been driven. Polymer classes identified: the synthetic-household fingerprint — polyamide (nylon), polyethylene terephthalate (PET, the workhorse polyester of fabrics and fibres), polyethylene, polypropylene, and other synthetic fractions typical of domestic furnishings. Adult inhalation dose estimate: 68,000 ± 40,000 particles per day in the 1–10 µm fraction alone. Child estimate: 47,000 ± 28,000 per day. And the verbatim structural finding that is the reason this piece is a pattern report and not a product report: 94% of detected particles were smaller than 10 µm.¹¹

Ten micrometres is, by convention, the thoracic threshold — the size below which inhaled particles are considered capable of penetrating beyond the upper airways into the bronchi and deeper lung.²⁷ Finer still (roughly 1–4 µm) and the particles reach the alveoli, the gas-exchange surface where the body's defences are thinnest. Ninety-four per cent of the particles the Toulouse team counted are below the thoracic threshold, and the majority sit in the size band that reaches further in. They are not being swallowed. They are being breathed into the part of the lung the body is least able to push out.

The sub-micron fraction — the particles smaller than the pipeline's 1 µm floor — is extrapolated by power-law from the measured 1–10 µm distribution. It is not directly polymer-identified. It is, in the paper's careful phrase, an approximate estimate. Remember that word. It will come back.

An independent Raman study of indoor-outdoor ratios in Wenzhou, China, in 2021 had already reported indoor concentrations in the same order of magnitude as — and in fact above, by a factor of roughly three — the Toulouse median, using a different pipeline in a different geography.¹² The magnitude does not appear to be an artefact of one French lab. What is new in 2025 is not that the particles are there. What is new is that they can be counted at polymer resolution. The microscope caught up.

If the 2025 indoor figures hold under replication, then the adult inhalation dose of 1–10 µm microplastic particles alone appears to be larger, by a substantial margin, than the 1–10 µm particle dose attributable to food and water in the canonical food-frame synthesis. May be is doing load-bearing work in that sentence; the literature is young, the hedging is earned, the replication list is short. But the shape of the evidence has tipped, and the question the next three sections of this report are about is not whether the tip happened. It is why it took ten years. It is why the ten years felt, from inside, like science proceeding in the right order. And it is why, if you had asked a food-frame scientist in 2018 what we should be looking at, they would not — honestly could not — have said start with the sofa.

The food frame did not beat the air frame in a fair fight. The air frame was not yet a number.

III. The Shape of What We Notice

Try, for a second, to picture your own bedroom inhaling.

Not the air in it. The room itself — the walls, the duvet, the polyester curtain, the carpet, the mattress pad, the blackout lining, the vacuum bag under the stairs, the cheap rug inside the wardrobe. Picture the room as a slow-motion lung.

You can picture a meal. You have been rehearsing the image of a meal your whole life. You know what a plate looks like. You know what a fork looks like. You know the exact embarrassed half-second between noticing a black fleck on a lettuce leaf and deciding what it is. You have cultural machinery for contaminated food going back to the bronze age. You have grandmothers.

But you cannot picture a bedroom breathing in. Which is why, for ten years, nobody has asked you to.

What we can imagine is what we can act on. What we cannot imagine is what waits.

There is a reason environmental risks travel in particular shapes. The risk-perception literature is clear that affect — the felt weight of a hazard, its visibility, its visceral image — does far more work on our attention than probability or magnitude.¹³ What Paul Slovic called the affect heuristic is not a failure mode; it is how the apparatus is wired. Risks that can be pictured get fought. Risks that cannot be pictured wait for the instrument that makes them picturable. This is not a character flaw. It is the species acting to type.

There is a narrower version of this claim that will not do the work this report needs. You could call what I am describing availability bias — we attend to what comes readily to mind — and stop there. But availability bias predicts only direction: that any risk that lacks a vivid image gets underweighted. What it does not predict is timing — why the same risk, unchanged in underlying magnitude, moves from underweighted to overweighted inside eighteen months the moment a new instrument makes it countable at a new resolution. Availability bias has no clock. The pattern this report is prosecuting is not a property of one brain; it is a property of four apparatuses acting in sequence — an instrument, a number, a story, and a policy — and each link in that sequence has a date on it. Availability bias explains why the story of the room took its shape once it began. The Instrument Shadow is about why it did not begin until the microscope arrived.

But the literature does not yet name the thing I want to name here, which is that the route by which a hazard enters the public imagination appears to require a particular narrative shape — a protagonist (the consumer, making a choice), an instrument of harm (the product, the meal, the fish), a moment of contamination (the bite, the swallow), an adversary (the corporation, the supplier), and a route to rescue (the ban, the substitute, the cleaner shelf). A five-part shape. If the story does not fit that shape — if there is no moment, no bite, no choice, no adversary the reader can see through a photograph — it does not travel, even when the exposure is larger. It waits in the draft folder of the species' conversation with itself until some new instrument gives it a moment somebody can photograph.

It turns out "follow the science" also means "wait for the microscope."

The food-plus-plastic story had all five parts. The shopper picks up a bottle of water. The bottle sheds microplastic into the water. The shopper drinks the water. The brand is the adversary. The reusable bottle is the rescue. You can photograph every one of those steps, and the WWF campaign did. There is a reason that sentence travelled in 48 hours from Geneva to a Pret in a motorway car park. The sentence had a photograph for every verb.

The air-plus-plastic story has none of the five parts. The protagonist is sitting still. The instrument of harm is the upholstery they are sitting on, and also the rug, and also the ceiling tile, and also the curtain, and also every cushion in a hundred million rooms. There is no moment; the moment is every moment. There is no bite; there is only breathing, and breathing does not look like a decision. There is no single adversary; the sofa does not come from one factory and the curtain does not come from one mill and the ceiling tile does not come from one supplier. And there is no rescue that fits on the same scale as the harm, because replacing the polyester curtain and the upholstery and the rug and the ceiling tile is not a shopping decision; it is a renovation the household cannot afford. A five-part story reduces, in the telling, to a room you cannot leave that has no fault you can pin.

The conclusion is not that the food-frame scientists were wrong, or the WWF campaigners were cynical, or the 2018 journalists were lazy. The conclusion is that the air story did not yet have an instrument that could make it a five-part story. The moment it had one — the moment Raman microspectroscopy at 1 µm resolution made it possible to point at a cushion in a specific apartment in Toulouse and say this fibre, this polymer, this count, this dose — the shape the story needed began, for the first time, to be available. The instrument was the precondition for the five-part shape. And the five-part shape was the precondition for the story. And the story was the precondition for the room becoming, in our heads, a room.

Nothing about us is unusual here. The story has been the same shape the whole time.

IV. The Same Sofa, Three Times

We have been in this piece of furniture before.

The sofa in the room around you — or the mattress, or the rug, or the curtain, or the ceiling tile — has been rebuilt, inside its frame, three times in the lifetime of anyone reading this, each time to solve the previous version's problem, each time by a chemistry whose cost was invisible until the instrument for measuring it arrived.

First, the sofa caught fire. The 1988 Furniture and Furnishings (Fire Safety) Regulations, responding to real domestic deaths, required every foam-filled piece of upholstered furniture sold in the UK to pass an open-flame ignition test. The regulations did their job; fire deaths from sofa ignition fell steeply over the following two decades. The instrument that solved the problem was, at core, a match. You could see a match. You could count what it did. The problem was visible the whole time, and when we finally addressed it we addressed it through chemistry — specifically, brominated flame retardants, most prominently the PBDE family, layered into the foam by weight.

Then the sofa became the chemistry. Over the 1990s and into the 2000s, gas chromatography–mass spectrometry instruments matured to the point where PBDEs could be measured in the blood and breast milk of people who had never touched a sofa factory in their lives. The body-burden studies arrived; the regulations followed; the major PBDE congeners were phased out over the following decade. The problem had been there since the late 1980s. The instrument that could see it arrived roughly a decade and a half later.

Then the sofa became the particles. Brominated flame retardants were swapped for alternative chemistries, or for less-flammable upholstery fabrics, or for foams that resisted ignition at higher density. What the substitutions left behind — the slow shedding of synthetic fibre from the fabric, the slow abrasion of the foam into the air of the room, the slow polyester bloom of the carpet and the curtain around the sofa — was a problem nobody had an instrument for until the automated Raman pipelines came online in the 2020s.

One sofa. Three risks. Three instruments. Each instrument the precondition for the next round of solving. Each round genuinely reduced the problem it was pointed at — the 1988 sofa really did burn, the 2004 PBDE really did accumulate, and both reductions saved real bodies. But each round installed, in the place of the old hazard, a new hazard whose shape the current instrument could not resolve. The generation of the home that caught fire was regulated by the match; the generation that caught body-burden was regulated by GC-MS; the generation that is turning up in our lung surveys is being regulated, for now, by silence.

Every generation of the home is regulated by the previous generation's microscope.

The substitution grammar underneath that clock — fire became chemistry became particles — has been laid out elsewhere in this series (report 020, The Progress Paradox). This report is not rewriting that grammar. It is interested in the clock: when each instrument arrived, and what was invisible until it did. The two reports are siblings looking at the same family from different angles.

Take secondhand smoke. Of the five historical cases this report has in its pocket, SHS is the cleanest structural twin to microplastic — one instrument, one narrative breakout, one policy landing, and roughly a decade between each. The biochemical assay for cotinine — the metabolite of nicotine, and the biomarker that made somebody else's smoke inside somebody else's body into a countable quantity — matured into routine use around 1980. In 1985, Martin Jarvis and colleagues at the Institute of Psychiatry in London used it on the saliva of 569 non-smoking schoolchildren, in the British Medical Journal, and reported measurable cotinine in the overwhelming majority of children whose parents smoked at home.²⁶ The following year, in December 1986, the US Surgeon General's report The Health Consequences of Involuntary Smoking named passive smoking as a cause of disease in non-smokers. Twenty-one years after that, on 1 July 2007, the UK indoor smoking ban took effect — in pubs and workplaces, but not, significantly for this report, the living room.¹⁷ One instrument (1980). One breakout (1985–1986). One policy (2007). Twenty-seven years from instrument to policy. That clock is the shape of the room we are standing in now.

Here is the same clock laid out across the five historical cases this report carries, so the pattern can be read end-to-end rather than in prose:

Asbestos is the instructive exception. Doll's 1955 paper in the British Journal of Industrial Medicine and the Wagner, Sleggs, and Marchand paper in the same journal in 1960 were both based on clinical counting — cases, lungs, and deaths on mortuary tables — rather than on fibre counts in ambient air. The disease arrived before the instrument that could count its dose. The phase-contrast microscopy fibre-counting protocol did not reach industrial routine until 1969, and the UK chrysotile ban did not take effect until 24 November 1999 — forty-four years after Doll, thirty years after the instrument. And the 1999 ban addressed new exposure; the asbestos already in the walls of pre-2000 housing was left in the walls.¹⁶ Asbestos is the exception that clarifies the rule: when the disease arrives before the instrument, the instrument still has to arrive before the policy can — and when it arrives, the policy still arrives thirty years late.

Five rooms. Five instruments. Five waits.

We are, it turns out, exactly as clever and exactly as foolish as we have always been.

V. The Atmosphere

The word we use for the air in a room is, underneath, a word for the fact that we could not see its walls.

Atmosphere is a seventeenth-century neologism — Greek atmós, vapour or breath, compounded with sphaîra, sphere. It entered English in 1638, in John Wilkins.¹⁹ The astronomers who adopted it were not describing the weather; they were describing the height of the vapours that cause twilight, a region they could calculate but not touch. The word, from its very first breath, was the name of a room we could not see the walls of. We inherited it four hundred years before we inherited the instrument to measure what was inside it.

The last time the instrument made a serious move toward the interior of that room, it was on 30 July 2025, in a French laboratory, at 1 µm resolution, and it revealed that the air of three apartments, measured across seven sampling runs, contained a median of 528 particles of identifiable plastic per cubic metre.

On 2 October 2025, the UK Committee on the Medical Effects of Air Pollutants issued a brief statement. The verbatim sentence, from the GOV.UK summary text, reads:

COMEAP currently advises against trying to assess population health risks from airborne NMPs until the evidence base has improved.²⁰

The committee was not being evasive. It was doing its job. Its job is to say when a number is firm enough to carry a policy, and the airborne microplastic number, in October 2025, was not firm enough for that weight. What is extraordinary is not the deferral. What is extraordinary is that the deferral is the pattern acting on itself in real time. Read the committee's sentence again, slowly, with the five historical rooms in the back of your head. It is the sound a microscope makes when it has not yet caught up.

And the next instrument is already visible. In January 2024, a group at Columbia University published, in the Proceedings of the National Academy of Sciences, a hyperspectral stimulated Raman scattering platform with a single-particle detection floor of 60 nanometres, confirmed in the paper's methods — sensitive enough to count single nanoplastic particles an order of magnitude below the floor of the Toulouse pipeline. They pointed it at three brands of bottled water. They reported a total micro-nano plastic concentration of about 2.4 × 10⁵ particles per litre, approximately 90% of which were nanoplastic, and the authors noted that this was one to two orders of magnitude above the microplastic counts reported in earlier bottled-water surveys by older methods.²¹

Note the careful wording. The 100× figure is not a within-sample comparison of two measurement techniques. It is a comparison of the new counted burden against the previous literature's counted burden. The burden was there the whole time; the instrument was not. The instrument dropped the floor from approximately 10 µm to approximately 60 nm, and the number the field thought it knew moved by two orders of magnitude. A 2024 PNAS Letter by Dušan Materić, of the Helmholtz Centre for Environmental Research in Leipzig, raised methodological concerns about the blank-contamination protocols in the Qian study — a caveat that constrains how confidently the field should treat the precise 240,000-per-litre figure.²² Qian and co-authors published a "Reply to Materić" the same year; the paper has not been withdrawn; the precise point estimate should be read as reported rather than as a settled number. Fine. The precise count is not the point. The shape of the result is the point. Every time the instrument floor drops, the counted burden rises by roughly the same ratio the floor dropped.

No equivalent airborne experiment has yet been published. The instrument exists; the airborne measurement has not yet been made. That absence is the shadow as it is forming.

Here, then, is the sixth case — a prediction put on the record in April 2026 by a report that wants to be quoted back to itself in 2028 or 2030 when the next number arrives. If an airborne stimulated-Raman scattering, or single-particle ICP-MS, survey of the sub-1 µm fraction of indoor air is conducted between now and 2030 in a residential or vehicular setting, and the count it reports is not at least one order of magnitude above the 1–10 µm burden Yakovenko measured, then the framework this report is prosecuting is weakened and the piece should be read with that in mind. If, as the pattern of the last hundred years predicts, the count is one to two orders of magnitude higher, the headline will say worse than thought — and it will again be the wrong sentence. The count will not be a revision. It will be a new particle class that the previous decade's apparatus could not count. We have been told worse than thought so many times this century that we have forgotten to ask why the previous thought was an underestimate by exactly the ratio the previous instrument's floor was off.

What we cannot count, we cannot tell. What we cannot tell, we cannot choose. What we cannot choose, we are standing inside of.

Come back, for a moment, to the credit card.

The credit card is still a figure of speech. It is still, physically, the shape it was in the summer of 2019 — a small rectangle that fits in a pocket and weighs roughly five grams and can be held in the palm of a hand. That is what made it travel. The sentence had a photograph for the weight; the photograph had a small object; the small object had a wallet; the wallet had a pocket; the pocket was in a coat; the coat was hung in a room. The sentence did its five-part job. It found the moment. It found the bite. It found the photograph. It travelled across the world in forty-eight hours because it gave a species that thinks in pockets a thing to put in one.

The room around the pocket did not travel. The room around the pocket had no sentence. For ten years the room had no sentence because the room had no instrument — because the polymer-identification floor of the dominant analytical method could not see the size class the room was shedding — and the species does not, for reasons that have been true since the bronze age, tell stories about things it cannot see. So it held the thing it could hold. It held the credit card.

What the credit card was the whole time, underneath the photograph, was the shape of what we are able to notice. It was a figure the size of our attention. It was exactly as large as the pocket our instruments had built for it. And the room around the pocket — the room that was standing there the whole time with its upholstery breathing out polyester the way a lung exhales — was, and is, the larger thing the figure of speech was shadowing. The reader is already inside the larger thing, and has been since before the first sentence of this report.

We have repeated this shape five times in a century. We are still here. The microscope is finally wide enough to see the room, and the next microscope is already being built to see below the floor of this one. That is not a conclusion. It is a beginning we have been late to, the way we are always late — exactly as late as the instrument, and no later.

Hope is not the evidence that we will catch up. Hope is the practice of noticing, each time the floor drops, that the pattern has a name now, and that a pattern with a name can be carried. The credit card is the shape of what we have been able to say. The room is the shape of what we are learning to.

The Levers

This report is about a pattern, not a purchase. The levers it points at are structural rather than consumer-facing, and they belong to institutions before they belong to households.

• Publish the airborne sub-micron fraction. The most valuable single contribution any research group can make in the 2026–2030 window is an airborne analogue of the Qian et al. bottled-water SRS survey, in a residential or vehicular setting, with full polymer identification below 1 µm. The prediction this report puts on the record is testable only if that instrument is pointed at that air.
• Replicate Yakovenko et al. 2025 in a different lab, a different country, and a different housing stock. Independent Raman-pipeline replication of the residential and car-cabin figures is the single cheapest defence of the result against the "another panic" framing.
• Extend the UK domestic-dust regulatory regime past asbestos. The 1999 chrysotile ban protected new exposure and left the existing stock in the walls. A purpose-built domestic-interior regulatory regime — for fibre shedding, particulate release, and volatile load from furnishings — is the missing instrument for the room that every historical case in this report points at.
• Ask COMEAP for a revisit date. The October 2025 deferral is reasoned. It is also open-ended. A committed revisit window — two years, three years — converts "not enough data yet" from a resting position into a schedule.

No product is recommended. No brand is named as the fix. The gap this report maps is not something a shopper can close.

What Would Change This Analysis

This framework is falsifiable. Five things would require it to be revised or withdrawn:

1. Adult ingestion burden found to clearly exceed adult inhalation burden on the weight of evidence after 2026 — for example, if a systematic review of post-2025 particle-counting studies finds the food route exceeds the air route on a mass or particle basis. The piece ends at Section II, with the food frame restored.
2. The Yakovenko/Sonke 2025 residential figures shown to be a methodological artefact of the Toulouse Raman pipeline — for example, by a multi-lab intercomparison study finding non-replicable particle counts. Dimension 1 of the framework collapses.
3. Any historical case of a dominant-domestic-inhalation hazard publicly campaigned on before its food or contact equivalent. The instrument-shadow pattern is falsified by a counter-example. I could not locate one in the five cases I examined; the literature may contain one I missed.
4. Sub-micron airborne plastic counts found to be comparable to or smaller than 1–10 µm counts in a direct SRS or single-particle ICP-MS airborne survey published between now and 2030. The sixth-case forecast fails, although the historical pattern stands.
5. Peer-reviewed evidence that narrative story-shape does not, independent of magnitude, predict the mobilisation of a public risk. Dimension 3 of the framework collapses into a retrospective observation rather than a mechanism.

This report should be read with those five collapse conditions open on the desk. May be the dominant route is the exact weight of the evidence currently available, and the report's force comes from the pattern, not from the adjective.