Food · Misleading · 谁在害怕化学名
Food Additives & the Zero-Added Label
防腐剂挡住的是肉毒杆菌 · 剂量决定毒性的道理在肝和肾里: 挂把手、焊水球、往外滤, 产线一饱和就开始堆 · 三聚氰胺骗过的是一台只会数氮的仪器, 伤的是肾小管 · 天然的河豚毒素照样致命 · 身体不查一个分子的来路, 只读它的形状和剂量
Story path
- 1What a preservative is holding backWhat a preservative is holding back
- 2Dose makes the poison · your liver and kidneys decideDose makes the poison · your liver and kidneys decide
- 3Melamine fooled a machine that only counts nitrogenMelamine fooled a machine that only counts nitrogen
- 4Natural = safe? Three counterexamplesNatural = safe? Three counterexamples
- 5So what is the real problemSo what is the real problem
- 6Your body cannot read the labelYour body cannot read the label
Chapter 1
What a preservative is holding back
What a preservative is holding back
The small amount of nitrite in a slice of cured pork is holding back a bacterium that can stop you breathing.
It's called Clostridium botulinum, and its spores are everywhere — in soil, on the surface of meat — normally doing nothing. But it hates oxygen, and cured meat is exactly the home it wants: no air, moist, room temperature, protein everywhere. Once a spore wakes up in that environment, it starts releasing a neurotoxin into the meat. The toxin travels to the junction between nerve and muscle and jams the switch that lets the nerve fire, so the muscle receives no instructions at all: first drooping eyelids and slurred speech, then paralysis working down to the breathing muscles.
What nitrite does is concrete: in the meat it converts to a reactive small molecule, gets inside the bacterium, and dismantles several enzymes it needs to make energy — so the spore never wakes (Sofos 1979; EFSA 2017). As a side effect it also fixes that pink color and holds off the rancid taste of aging fat.
So what happens if you take it out? You don't get a cleaner piece of meat. You reopen the door for botulism.
That's the first thing this story is about: the real choice is never between 'additives' and 'no additives'. It is between this risk and that risk — which one is smaller and more controllable.
And 'smaller' is something you can actually measure. The place where it gets measured is concrete too — not a meeting room, but your liver and your kidneys. The next scene goes to look at that machine.
It's called Clostridium botulinum, and its spores are everywhere — in soil, on the surface of meat — normally doing nothing. But it hates oxygen, and cured meat is exactly the home it wants: no air, moist, room temperature, protein everywhere. Once a spore wakes up in that environment, it starts releasing a neurotoxin into the meat. The toxin travels to the junction between nerve and muscle and jams the switch that lets the nerve fire, so the muscle receives no instructions at all: first drooping eyelids and slurred speech, then paralysis working down to the breathing muscles.
What nitrite does is concrete: in the meat it converts to a reactive small molecule, gets inside the bacterium, and dismantles several enzymes it needs to make energy — so the spore never wakes (Sofos 1979; EFSA 2017). As a side effect it also fixes that pink color and holds off the rancid taste of aging fat.
So what happens if you take it out? You don't get a cleaner piece of meat. You reopen the door for botulism.
That's the first thing this story is about: the real choice is never between 'additives' and 'no additives'. It is between this risk and that risk — which one is smaller and more controllable.
And 'smaller' is something you can actually measure. The place where it gets measured is concrete too — not a meeting room, but your liver and your kidneys. The next scene goes to look at that machine.
The other half of the trade-off
Nitrite is not innocent. In the meat and in your stomach it can take part in forming nitrosamines — one of the mechanisms behind IARC classifying processed meat as Group 1. That is real, and the processed-meat island covers the grading and the numbers properly; no need to repeat them here.So the real approach isn't a binary of 'allow' or 'ban'. It squeezes the dose into the narrow gap where it is just enough to suppress the bacterium while forming as little nitrosamine as possible. After re-evaluating in 2017, the EU set the acceptable daily intake for nitrite ion at 0.07 mg per kg of body weight (EFSA 2017).
One more thing worth noticing, and scene six will come back to it: cured meats labelled 'no nitrite added' often use cultured celery powder instead. Celery is naturally rich in nitrate (vegetables are the largest dietary source of nitrate, EFSA 2008), and a bacterial culture reduces it to nitrite. The industry line is blunt: nitrite is nitrite regardless of the source (Mermelstein 2018). The label changed; the molecule didn't — and the bacterium reads the molecule.
bouvard-2015-processed-meat-iarc
Chapter 2
Dose makes the poison · your liver and kidneys decide
Dose makes the poison · your liver and kidneys decide
The same molecule is harmless in a small amount and lethal in a large one. There is no magic in between. The difference comes down to one thing: whether the speed at which you clear it keeps up with the speed at which it arrives.
First, where it has to go
Any foreign molecule that gets into your blood has only two exits: the kidney, which filters it into urine; and the liver, which dumps it into bile to leave with your stool.
The trouble is that most foreign molecules are fat-loving — they dissolve in oil. And fat-loving things have a fatal property: they can pass straight through cell membranes. The kidney tubule filters such a molecule into the urine, and it promptly dissolves back through the tubule wall into the blood. Filtered once, returned once. On the kidney alone, a molecule like that never leaves.
The liver's two steps
So the liver first reshapes it into something excretable. Two steps, and the order cannot be reversed:
Step one is drilling a handle onto it. A large family of enzymes in liver cells does exactly this (cytochrome P450, usually written CYP450): it presses an oxygen atom onto the foreign molecule, creating a socket that other things can attach to (Guengerich 2008).Step two is welding a water balloon onto that handle. Another set of enzymes hangs on a big, water-loving, electrically charged tag — glucuronic acid, sulfate, or glutathione (Jancova 2010).
After that the molecule has a different personality: heavy, charged, water-loving. Charged things cannot slip back through a membrane. So once the kidney filters it, it can never sneak back — it can only leave with the urine. The bile route works the same way.
That is the entirety of what 'detoxification' means: not destroying the thing, but reshaping it into something excretable. For the whole system, dive to hepatic and renal; for how powerless detox products are against it, see detox-cleanse.
Step one sometimes makes things worse
An honest note: drilling the handle is done by oxidation, and the intermediate that oxidation produces is often far more reactive than the original molecule. Almost always step two seals it immediately and nothing happens. But if step one runs too fast and step two can't keep up, that reactive intermediate gets a window in which to crash into things. This is exactly the road aflatoxin takes to cause cancer: it arrives at the liver fairly harmless, and it is the liver's own enzymes that activate it into a highly reactive molecule, which then lodges itself into DNA (IARC 2012). Scene four meets it again.
Only now does dose enter
Enzymes are physical objects: finite in number, each with a ceiling on how fast it can work. The whole of 'the dose makes the poison' rests on that one fact.
At low amounts there are idle enzymes to spare. Whatever arrives is grabbed immediately. Clearance speed tracks concentration — double the concentration, double the clearance rate. The result is that a fixed fraction leaves every hour, and it can never pile up.Past a certain amount, every enzyme is occupied. The production line is at capacity. More arriving changes nothing: only a fixed quantity can be cleared per hour.Beyond that point, input exceeds capacity and the remainder starts queuing. Blood concentration climbs, and the reactive intermediate that should have been sealed instantly now has time to linger.
So the relationship between dose and harm isn't a straight slope. It's a hinge: almost nothing happens below saturation, and it rises steeply above it.
Is this molecule toxic is therefore the wrong question. The question is: does this amount exceed the speed at which my body clears it?
Two examples that point opposite ways
Alcohol is the easiest to grasp: the enzyme that breaks it down saturates at a very low concentration. So blood alcohol doesn't fade proportionally — a fixed amount disappears per hour. That's also why drinking twice as much doesn't take twice as long, but longer. For the detail, see alcohol-metabolism.
Vitamin C is the other end. However much you take, the transporters in your intestinal wall are finite in number, so the absorbed fraction actually falls; and of whatever does reach your blood, anything above a certain concentration gets tipped straight into the urine by the kidney (Levine 1996). Which is why, when the EU re-evaluated ascorbic acid (E300 — vitamin C), it set no acceptable daily intake at all: at normal dietary levels it is regarded as physiologically harmless (EFSA 2015).
The same logic yields a strict limit at one end and no limit at all at the other. And what decides the outcome is never whether the molecule is natural or synthetic. It's how fast your body clears it.
First, where it has to go
Any foreign molecule that gets into your blood has only two exits: the kidney, which filters it into urine; and the liver, which dumps it into bile to leave with your stool.
The trouble is that most foreign molecules are fat-loving — they dissolve in oil. And fat-loving things have a fatal property: they can pass straight through cell membranes. The kidney tubule filters such a molecule into the urine, and it promptly dissolves back through the tubule wall into the blood. Filtered once, returned once. On the kidney alone, a molecule like that never leaves.
The liver's two steps
So the liver first reshapes it into something excretable. Two steps, and the order cannot be reversed:
Step one is drilling a handle onto it. A large family of enzymes in liver cells does exactly this (cytochrome P450, usually written CYP450): it presses an oxygen atom onto the foreign molecule, creating a socket that other things can attach to (Guengerich 2008).Step two is welding a water balloon onto that handle. Another set of enzymes hangs on a big, water-loving, electrically charged tag — glucuronic acid, sulfate, or glutathione (Jancova 2010).
After that the molecule has a different personality: heavy, charged, water-loving. Charged things cannot slip back through a membrane. So once the kidney filters it, it can never sneak back — it can only leave with the urine. The bile route works the same way.
That is the entirety of what 'detoxification' means: not destroying the thing, but reshaping it into something excretable. For the whole system, dive to hepatic and renal; for how powerless detox products are against it, see detox-cleanse.
Step one sometimes makes things worse
An honest note: drilling the handle is done by oxidation, and the intermediate that oxidation produces is often far more reactive than the original molecule. Almost always step two seals it immediately and nothing happens. But if step one runs too fast and step two can't keep up, that reactive intermediate gets a window in which to crash into things. This is exactly the road aflatoxin takes to cause cancer: it arrives at the liver fairly harmless, and it is the liver's own enzymes that activate it into a highly reactive molecule, which then lodges itself into DNA (IARC 2012). Scene four meets it again.
Only now does dose enter
Enzymes are physical objects: finite in number, each with a ceiling on how fast it can work. The whole of 'the dose makes the poison' rests on that one fact.
At low amounts there are idle enzymes to spare. Whatever arrives is grabbed immediately. Clearance speed tracks concentration — double the concentration, double the clearance rate. The result is that a fixed fraction leaves every hour, and it can never pile up.Past a certain amount, every enzyme is occupied. The production line is at capacity. More arriving changes nothing: only a fixed quantity can be cleared per hour.Beyond that point, input exceeds capacity and the remainder starts queuing. Blood concentration climbs, and the reactive intermediate that should have been sealed instantly now has time to linger.
So the relationship between dose and harm isn't a straight slope. It's a hinge: almost nothing happens below saturation, and it rises steeply above it.
Is this molecule toxic is therefore the wrong question. The question is: does this amount exceed the speed at which my body clears it?
Two examples that point opposite ways
Alcohol is the easiest to grasp: the enzyme that breaks it down saturates at a very low concentration. So blood alcohol doesn't fade proportionally — a fixed amount disappears per hour. That's also why drinking twice as much doesn't take twice as long, but longer. For the detail, see alcohol-metabolism.
Vitamin C is the other end. However much you take, the transporters in your intestinal wall are finite in number, so the absorbed fraction actually falls; and of whatever does reach your blood, anything above a certain concentration gets tipped straight into the urine by the kidney (Levine 1996). Which is why, when the EU re-evaluated ascorbic acid (E300 — vitamin C), it set no acceptable daily intake at all: at normal dietary levels it is regarded as physiologically harmless (EFSA 2015).
The same logic yields a strict limit at one end and no limit at all at the other. And what decides the outcome is never whether the molecule is natural or synthetic. It's how fast your body clears it.
Where the 100 comes from, in one pass
How much of an additive is allowed follows a short algorithm. Run long-term feeding studies, escalating the dose until you find the boundary where measurable abnormalities begin. The tier just below it is the no-observed-adverse-effect level (NOAEL). Divide that by 100 and you have the acceptable daily intake (ADI), in milligrams per kilogram of body weight per day.The 100 is two tens multiplied. What's worth seeing clearly is that each of those tens corresponds to the very machine this scene has been describing (WHO/FAO 2009):
The first 10 · you are not a rat. The enzyme set in a rat's liver is not yours, and the same molecule can be cleared at speeds an order of magnitude apart in the two species.The second 10 · you are not the average person either. Among humans, how much of each liver enzyme you carry is set by your genes; a child's production line isn't finished; the elderly and people with poor liver function have less capacity. That spread also reaches an order of magnitude.
So the 100 isn't administrative caution. It is the quantified margin for one specific thing: how much clearance speed can differ. The arithmetic belongs to a committee. The reasoning belongs to your liver and kidneys.
The ADI's definition rewards slow reading too: eating this much every day for a lifetime is expected to carry no appreciable health risk. It is not a red line where harm begins; it is a floor below which a lifetime of daily intake is fine. Concretely: the ADI for aspartame is 40 mg per kg of body weight (JECFA 2023) — 2,400 mg a day for a 60 kg person, roughly a dozen-plus cans of diet soda. Individual sweetener ADIs live on the artificial-sweeteners island.
And the thinnest link in the framework, stated honestly: it assesses each substance separately, and handles the combined effect of many additives eaten together far less solidly than any one of them alone. Put that back into this scene's mechanism and you can see why the combination is hard to compute — different molecules may crowd onto the same production line. Compete for the same enzyme family and together they might push it to saturation, while individually not one of them exceeded its limit. That gap is widely acknowledged and still being worked on. It is not evidence that additives are toxic, but it is genuinely the weakest link as things stand.
who-fao-2009-ehc240-risk-assessmentjecfa-2023-aspartame-adi
Chapter 3
Melamine fooled a machine that only counts nitrogen
Melamine fooled a machine that only counts nitrogen
The routine method for testing milk protein never sees protein at any point. It counts nitrogen atoms and then does a multiplication.
Why nitrogen can stand in for protein
Start with a question about the body: when you drink a mouthful of milk, what does your body actually take from it?
Not the word 'protein'. Protein is a long chain. It enters the stomach and small intestine, gets cut into pieces by enzymes that work like scissors, and finally transporters in the intestinal wall carry amino acids into your blood one at a time. What the body wants is those amino acids — above all the handful it cannot build itself and must get from food.
And amino acids share a feature: every one of them carries a nitrogen-containing head. That's where the multiplication comes from — nitrogen makes up roughly 16% of protein by weight, so measure total nitrogen, multiply by about 6.25, and call it protein.
That conversion rests on exactly one premise: all the nitrogen in the sample sits on amino acids.
That premise is precisely what melamine stepped on
Melamine is an industrial feedstock for plastics and coatings. Its molecule is small yet crammed with nitrogen — two-thirds nitrogen by weight, far more than any protein. Throw a handful into watered-down milk and the nitrogen count is restored, so the 'protein' figure on the lab sheet looks good again (Gossner 2009).
But what the instrument read and what the infant received are two unrelated things. Melamine is not an amino acid: the scissors in the stomach have nothing to cut, the transporters in the intestinal wall don't recognize its shape, and not a single amino acid gets carried in. The number was real. The thing the number stood for was not there at all.
Then it went to the kidney
Carry the previous scene's question forward: once this molecule is inside, how does it get cleared?
It doesn't need to trouble the liver — it is already water-loving enough that the kidney can filter it straight into urine, no handle required. That sounds like good news. It's exactly where the trouble is.
The kidney tubule's day job is taking water back. As filtrate runs down the tube, water is drawn back into the blood and the remaining fluid gets steadily more concentrated. So the melamine gets steadily more concentrated too. And the same industrial feedstock often carries a relative alongside it — cyanuric acid, which travels the same route.
Once the two molecules are concentrated enough inside the tubule, they begin to interlock: the hydrogen of one against the nitrogen of the other, layer on layer, into a crystal that will not dissolve in water (WHO 2008). What is now in the tube is no longer a molecule. It's a stone. The tube blocks, urine can't get out, pressure backs up, and filtration stops.
Look closely at the causation here: the harm isn't toxicity in the poisoning sense. It's physical obstruction. And it was the kidney doing its job that concentrated it into existence — the very act that turns a low-toxicity molecule into a stone is what this organ does all day.
That also explains why it was infants who collapsed: their tubes are narrower, their urine more concentrated, and formula was their only food — per kilogram of body weight, more went in than for any adult. In the 2008 incident the WHO counted roughly 300,000 cases, 51,900 hospitalizations, and 6 deaths (WHO 2008).
What this scene really wants to leave you with is that 6.25
An instrument never measures the thing you care about. It measures a proxy — a shadow that usually moves in step with the thing you care about. Break that 'usually' deliberately and the number survives while its meaning does not.
This tool is portable. Amino spiking in the protein powder trade is the second generation of the same move: cheap free amino acids — glycine, taurine — are blended into the powder. This time what's added really is amino acids, so the nitrogen passes easily. But they happen to be the ones your body can already build itself, and you get no more of the ones you actually lack. The nitrogen test still can't tell the difference, because it was never measuring what you wanted in the first place.
So the next time you see any number, it's worth asking first: is the thing this instrument directly measured the same as the thing I care about? And if there's a conversion in between, what does that conversion assume?
And one sentence set straight
Melamine is not a food additive. It was never approved as one, never evaluated as one, and nobody ever gave it an ADI. So 'the melamine scandal proves additives aren't safe' is broken grammatically, not politically: it is equivalent to saying someone robbed a bank with counterfeit notes, therefore currency is unsafe.
Conflating the two actively helps the counterfeiters. Your vigilance gets spent on the long chemical names on the ingredient list, while the thing you actually need to guard against isn't on that list at all — the ingredient list is filled in by the people who follow the rules.
Why nitrogen can stand in for protein
Start with a question about the body: when you drink a mouthful of milk, what does your body actually take from it?
Not the word 'protein'. Protein is a long chain. It enters the stomach and small intestine, gets cut into pieces by enzymes that work like scissors, and finally transporters in the intestinal wall carry amino acids into your blood one at a time. What the body wants is those amino acids — above all the handful it cannot build itself and must get from food.
And amino acids share a feature: every one of them carries a nitrogen-containing head. That's where the multiplication comes from — nitrogen makes up roughly 16% of protein by weight, so measure total nitrogen, multiply by about 6.25, and call it protein.
That conversion rests on exactly one premise: all the nitrogen in the sample sits on amino acids.
That premise is precisely what melamine stepped on
Melamine is an industrial feedstock for plastics and coatings. Its molecule is small yet crammed with nitrogen — two-thirds nitrogen by weight, far more than any protein. Throw a handful into watered-down milk and the nitrogen count is restored, so the 'protein' figure on the lab sheet looks good again (Gossner 2009).
But what the instrument read and what the infant received are two unrelated things. Melamine is not an amino acid: the scissors in the stomach have nothing to cut, the transporters in the intestinal wall don't recognize its shape, and not a single amino acid gets carried in. The number was real. The thing the number stood for was not there at all.
Then it went to the kidney
Carry the previous scene's question forward: once this molecule is inside, how does it get cleared?
It doesn't need to trouble the liver — it is already water-loving enough that the kidney can filter it straight into urine, no handle required. That sounds like good news. It's exactly where the trouble is.
The kidney tubule's day job is taking water back. As filtrate runs down the tube, water is drawn back into the blood and the remaining fluid gets steadily more concentrated. So the melamine gets steadily more concentrated too. And the same industrial feedstock often carries a relative alongside it — cyanuric acid, which travels the same route.
Once the two molecules are concentrated enough inside the tubule, they begin to interlock: the hydrogen of one against the nitrogen of the other, layer on layer, into a crystal that will not dissolve in water (WHO 2008). What is now in the tube is no longer a molecule. It's a stone. The tube blocks, urine can't get out, pressure backs up, and filtration stops.
Look closely at the causation here: the harm isn't toxicity in the poisoning sense. It's physical obstruction. And it was the kidney doing its job that concentrated it into existence — the very act that turns a low-toxicity molecule into a stone is what this organ does all day.
That also explains why it was infants who collapsed: their tubes are narrower, their urine more concentrated, and formula was their only food — per kilogram of body weight, more went in than for any adult. In the 2008 incident the WHO counted roughly 300,000 cases, 51,900 hospitalizations, and 6 deaths (WHO 2008).
What this scene really wants to leave you with is that 6.25
An instrument never measures the thing you care about. It measures a proxy — a shadow that usually moves in step with the thing you care about. Break that 'usually' deliberately and the number survives while its meaning does not.
This tool is portable. Amino spiking in the protein powder trade is the second generation of the same move: cheap free amino acids — glycine, taurine — are blended into the powder. This time what's added really is amino acids, so the nitrogen passes easily. But they happen to be the ones your body can already build itself, and you get no more of the ones you actually lack. The nitrogen test still can't tell the difference, because it was never measuring what you wanted in the first place.
So the next time you see any number, it's worth asking first: is the thing this instrument directly measured the same as the thing I care about? And if there's a conversion in between, what does that conversion assume?
And one sentence set straight
Melamine is not a food additive. It was never approved as one, never evaluated as one, and nobody ever gave it an ADI. So 'the melamine scandal proves additives aren't safe' is broken grammatically, not politically: it is equivalent to saying someone robbed a bank with counterfeit notes, therefore currency is unsafe.
Conflating the two actively helps the counterfeiters. Your vigilance gets spent on the long chemical names on the ingredient list, while the thing you actually need to guard against isn't on that list at all — the ingredient list is filled in by the people who follow the rules.
The paperwork layer, briefly
On paper the cut is clean, and it uses two entirely separate lists. The permitted list is GB 2760, the Standard for Uses of Food Additives, where each entry is pinned down: its name, which food categories it may enter, the maximum level (GB 2760-2024). The forbidden list is the Catalogue of Non-Food Substances That May Be Illegally Added to Food, published by the Ministry of Health in batches from December 2008; melamine is on it, and so is Sudan red. Their category name is 'non-food substances'.Sudan red is the same sentence in another form: an industrial dye for shoe polish and floor wax, mixed into chilli powder only because chilli powder is priced on redness (EC 2005/402). It shares one feature with melamine, and that feature says more than the lists do — neither was added to make the food better. Both were added to make a number or a look better. That is the real definition of an illegal additive: it serves the instrument and the shelf, not the person eating it.
WHO later set a tolerable daily intake for melamine of 0.2 mg per kg of body weight (WHO 2008). The point of that number isn't to permit adding it. It draws a line for trace contamination — migration from plastic containers, say — so that deliberate adulteration and trace contamination can be told apart.
gb-2760-2024-additive-usemoh-china-2008-illegal-additives-listec-2005-402-sudan-dyes
Chapter 4
Natural = safe? Three counterexamples
Natural = safe? Three counterexamples
Tetrodotoxin is 100% natural, contains no additives whatsoever, and one small mouthful can stop you breathing within hours.
This scene does one job: prise the words 'natural' and 'safe' apart. Three examples, none containing anything man-made.
Tetrodotoxin: it plugs, with precision, the little gate in the nerve cell membrane that lets sodium ions through. With the gate blocked the nerve can't fire, and muscles get no instructions — first numb lips and tongue, then paralysis working down to the breathing muscles. It doesn't cook out; ordinary heat won't destroy it. And there is no antidote: the hospital can breathe for you with a machine while your body metabolizes it (Lago 2015).
Aflatoxin: what Aspergillus flavus releases when peanuts, corn, or nuts get damp and mouldy. It takes exactly the road from scene two — the liver's enzymes drill a handle onto it, the resulting intermediate is wildly reactive, and before step two can seal it away it has already lodged itself into DNA and damaged p53, the gene that acts as a brake. IARC lists it as Group 1 — the same tier as processed meat and tobacco: sufficient evidence of carcinogenicity in humans (IARC 2012). It is heat-stable too; frying won't remove it.
Bongkrekic acid: in wood-ear fungus soaked too long, home-fermented corn flour, or spoiled wet rice noodles, a bacterium called Burkholderia gladioli releases it. It gets into your mitochondria and jams the machine that makes energy; starved of fuel, cells die in sheets. Among the poisonings recorded in China between 2010 and 2020, the case fatality rate approached one in three (Zhang 2023). Heat-stable, and no antidote. How to avoid natural toxins is covered on the foodborne-illness page.
Notice what these three share, picking up from the last scene: 'no antidote' means there is no clearance route anyone can speed up. All a hospital can do is keep you going while your own machine grinds through it. The dose-makes-the-poison rule still holds for them perfectly well — their capacity is simply terrifyingly low.
'Natural' describes origin, not safety. Some of the most toxic molecules on earth are made by living things — they were synthesized precisely in order to kill something. And 'synthetic' likewise only describes origin: the vitamin C you take, whether pressed from a fresh pepper or made in a factory, is the same molecule by the time it reaches your small intestine, and the transport protein in the intestinal wall can't tell them apart and doesn't care.
Honestly: 'natural things are on average more worth eating' is often true. But it's true because whole foods arrive with fibre, potassium, and phytochemicals attached — not because the word 'natural' holds any magic. The reason has to match the conclusion, or the next person selling a pure natural detox herb will win you over with the identical sentence.
This scene does one job: prise the words 'natural' and 'safe' apart. Three examples, none containing anything man-made.
Tetrodotoxin: it plugs, with precision, the little gate in the nerve cell membrane that lets sodium ions through. With the gate blocked the nerve can't fire, and muscles get no instructions — first numb lips and tongue, then paralysis working down to the breathing muscles. It doesn't cook out; ordinary heat won't destroy it. And there is no antidote: the hospital can breathe for you with a machine while your body metabolizes it (Lago 2015).
Aflatoxin: what Aspergillus flavus releases when peanuts, corn, or nuts get damp and mouldy. It takes exactly the road from scene two — the liver's enzymes drill a handle onto it, the resulting intermediate is wildly reactive, and before step two can seal it away it has already lodged itself into DNA and damaged p53, the gene that acts as a brake. IARC lists it as Group 1 — the same tier as processed meat and tobacco: sufficient evidence of carcinogenicity in humans (IARC 2012). It is heat-stable too; frying won't remove it.
Bongkrekic acid: in wood-ear fungus soaked too long, home-fermented corn flour, or spoiled wet rice noodles, a bacterium called Burkholderia gladioli releases it. It gets into your mitochondria and jams the machine that makes energy; starved of fuel, cells die in sheets. Among the poisonings recorded in China between 2010 and 2020, the case fatality rate approached one in three (Zhang 2023). Heat-stable, and no antidote. How to avoid natural toxins is covered on the foodborne-illness page.
Notice what these three share, picking up from the last scene: 'no antidote' means there is no clearance route anyone can speed up. All a hospital can do is keep you going while your own machine grinds through it. The dose-makes-the-poison rule still holds for them perfectly well — their capacity is simply terrifyingly low.
'Natural' describes origin, not safety. Some of the most toxic molecules on earth are made by living things — they were synthesized precisely in order to kill something. And 'synthetic' likewise only describes origin: the vitamin C you take, whether pressed from a fresh pepper or made in a factory, is the same molecule by the time it reaches your small intestine, and the transport protein in the intestinal wall can't tell them apart and doesn't care.
Honestly: 'natural things are on average more worth eating' is often true. But it's true because whole foods arrive with fibre, potassium, and phytochemicals attached — not because the word 'natural' holds any magic. The reason has to match the conclusion, or the next person selling a pure natural detox herb will win you over with the identical sentence.
Chapter 5
So what is the real problem
So what is the real problem
The real problem with a bag of crisps isn't the chemical names at the end of the ingredient list. It's the first few words at the top: oil, salt, sugar.
Ingredient lists are ordered by weight, highest first. What decides what that bag does to you is the top three, not the antioxidant at the bottom added at a tenth of a percent.
Degree of processing and additives are two different things. They often show up together, which is why they keep getting treated as one:
A sugary drink can be genuinely free of added preservatives and still deliver thirty-odd grams of sugar per bottle.A soy sauce marketed as zero-added can carry exactly the same sodium as the ordinary bottle beside it.Meanwhile plain yogurt, canned tomatoes, and frozen vegetables all carry additives on the label, and all are good things.
Here's the honest part: ultra-processed food genuinely is a real problem. That claim is not one we're here to dismantle — it has evidence. In the Hall 2019 inpatient controlled trial, with macronutrients matched and both groups eating ad libitum, the ultra-processed arm took in about 500 kcal more per day and gained weight within two weeks. Umbrella-review epidemiology points consistently at a range of adverse outcomes (Lane 2024) — that's association, not causation, but Hall's trial gives it a genuine causal leg. NOVA and that trial are covered properly on the ultra-processed-foods island; no need to repeat them.
What matters is why it's a problem. The evidence currently points at: high energy density, little fibre, lots of sugar and sodium, and a texture so soft you barely chew, so you eat fast. Those are all facts about the recipe and the structure. 'Because it contains additives' is not on that list. Additives are a marker of ultra-processed food, not its cause.
Mistaking the marker for the cause has a concrete cost: you pick the zero-added cookies, eat the same sugar and fat, and — because the label reassured you — quite possibly eat a bit more of them.
Ingredient lists are ordered by weight, highest first. What decides what that bag does to you is the top three, not the antioxidant at the bottom added at a tenth of a percent.
Degree of processing and additives are two different things. They often show up together, which is why they keep getting treated as one:
A sugary drink can be genuinely free of added preservatives and still deliver thirty-odd grams of sugar per bottle.A soy sauce marketed as zero-added can carry exactly the same sodium as the ordinary bottle beside it.Meanwhile plain yogurt, canned tomatoes, and frozen vegetables all carry additives on the label, and all are good things.
Here's the honest part: ultra-processed food genuinely is a real problem. That claim is not one we're here to dismantle — it has evidence. In the Hall 2019 inpatient controlled trial, with macronutrients matched and both groups eating ad libitum, the ultra-processed arm took in about 500 kcal more per day and gained weight within two weeks. Umbrella-review epidemiology points consistently at a range of adverse outcomes (Lane 2024) — that's association, not causation, but Hall's trial gives it a genuine causal leg. NOVA and that trial are covered properly on the ultra-processed-foods island; no need to repeat them.
What matters is why it's a problem. The evidence currently points at: high energy density, little fibre, lots of sugar and sodium, and a texture so soft you barely chew, so you eat fast. Those are all facts about the recipe and the structure. 'Because it contains additives' is not on that list. Additives are a marker of ultra-processed food, not its cause.
Mistaking the marker for the cause has a concrete cost: you pick the zero-added cookies, eat the same sugar and fat, and — because the label reassured you — quite possibly eat a bit more of them.
So where should your eyes go on the package
Look first at the nutrition table: energy, sodium, sugar. Those numbers are what you actually accumulate day after day (from March 2027 sugar and saturated fat become mandatory entries too, GB 28050-2025).Then read the top three ingredients. If sugar or refined oil is first, nothing further down the list changes that.Don't spend your energy decoding the chemical names at the end. Ascorbic acid is vitamin C — a legal additive numbered E300 in the EU, used as an antioxidant, and the same molecule you get from an orange. After re-evaluation the EU didn't even set an ADI for it, since at normal dietary levels it is regarded as physiologically harmless (EFSA 2015). A chemical name you don't recognize tells you that chemical nomenclature and everyday speech are different vocabularies. It tells you nothing else.Why sodium and sugar earn that attention is covered on the salt and sugar-honey islands; the randomized evidence on sodium and blood pressure is the DASH-Sodium trial (Sacks 2001), and the ceiling on free sugars is WHO's recommendation (WHO 2015).
gb-28050-2025-nutrition-labelssacks-2001-dash-sodiumwho-2015-free-sugars
Chapter 6
Your body cannot read the label
Your body cannot read the label
There is no part of you that can tell whether a molecule is natural or synthetic. That isn't a debating trick — it's structural. The molecule has no such field on it.
What the body actually reads
When a molecule arrives at your intestinal wall, only two kinds of things are waiting for it: transporters that recognize shape, and enzymes that recognize shape. They work in an intensely physical way, like a lock reading the teeth of a key. Whether the molecule gets carried in, or gets cut open, depends on what it looks like — not on where it came from.
The vitamin C example is the cleanest: ascorbic acid pressed from a fresh pepper and ascorbic acid made in a factory are the same molecule — same atoms, same arrangement. The transporter in your intestinal wall treats them identically, because it cannot do otherwise: all it can feel is the teeth, and the teeth are identical.
So the body can read exactly three things:
What shape this molecule is — which decides what lock it opens. Scene one's nitrite dismantles a bacterium's enzymes; scene four's tetrodotoxin plugs the sodium gate.How much, and how fast it arrives — which decides whether it outruns your clearance speed, the production line from scene two.What it arrived with — the fibre, water, and potassium in whole food change how fast it gets absorbed.
Natural and synthetic are not on that list. Those two words describe a molecule's résumé, and the body never checks résumés.
So what does 'zero added' promise?
It promises a category that does not exist in your body. That's the root of the whole pitch: it moves your attention from which molecule, and how much to where the molecule came from — the one dimension the body happens not to read.
For how concrete that is, scene one's example suffices. A cured meat labelled 'no nitrite added' has often switched to cultured celery powder. Celery is naturally rich in nitrate (vegetables are the largest dietary source of nitrate, EFSA 2008), and a bacterial culture reduces it to nitrite. So: what lands on the botulism bacterium's enzymes is the same nitrite ion; and what takes part in forming nitrosamines in your stomach is the same nitrite ion. The industry line is blunt: nitrite is nitrite regardless of the source (Mermelstein 2018).
The label changed; the molecule didn't. The bacterium can't read the label, and neither can your stomach.
So what is the phrase actually selling?
A contrast that doesn't exist. 'This product has zero added preservatives' implies the others added them, and the others are worse. But the preservative the others use has been evaluated, numbered, and capped; whereas this 'zero added' version may simply have a shorter shelf life (so the door from scene one opens a crack again), or may just have switched to a molecule that doesn't count as an 'additive' on the label and is identical inside your body. You paid more and bought reassurance, not safety.
Regulators have seen this too: from March 2027, claims like 'zero added' may no longer be printed on prepackaged food in China (GB 7718-2025). But that only takes the words off the shelf. What actually protects you is knowing why they never promised anything in the first place.
Finally, the whole picture
This page is not a brief for the food industry:
Ultra-processed food is a real problem — and the real problem is the recipe and the structure, scene five.The sugar and sodium in the recipe are a real problem — and those two numbers are printed plainly on the nutrition table.People really do put things into food that don't belong there — and those things appear on no ingredient list at all, scene three.
All of that vigilance is warranted.
What's worth dismantling is the bad reasoning: chemical name equals danger, natural equals safe, contains additives equals toxic. Its flaw isn't excess vigilance — it's vigilance aimed at the one dimension the body doesn't read, a molecule's origin, and therefore missing the three it does: which molecule, how much, and what it came with.
Next time you see those words on a package, you don't need to remember a single regulation. One question is enough: did the molecule change, or only its origin story?
Related: the closest precedent is MSG — dive to msg-glutamate; nitrite in cured meat and the IARC grading are at processed-meat; sweetener ADIs at artificial-sweeteners; NOVA and Hall's trial at ultra-processed-foods; avoiding natural toxins at foodborne-illness; whether organic is safer at organic-food; the full picture of the liver's two steps at hepatic.
What the body actually reads
When a molecule arrives at your intestinal wall, only two kinds of things are waiting for it: transporters that recognize shape, and enzymes that recognize shape. They work in an intensely physical way, like a lock reading the teeth of a key. Whether the molecule gets carried in, or gets cut open, depends on what it looks like — not on where it came from.
The vitamin C example is the cleanest: ascorbic acid pressed from a fresh pepper and ascorbic acid made in a factory are the same molecule — same atoms, same arrangement. The transporter in your intestinal wall treats them identically, because it cannot do otherwise: all it can feel is the teeth, and the teeth are identical.
So the body can read exactly three things:
What shape this molecule is — which decides what lock it opens. Scene one's nitrite dismantles a bacterium's enzymes; scene four's tetrodotoxin plugs the sodium gate.How much, and how fast it arrives — which decides whether it outruns your clearance speed, the production line from scene two.What it arrived with — the fibre, water, and potassium in whole food change how fast it gets absorbed.
Natural and synthetic are not on that list. Those two words describe a molecule's résumé, and the body never checks résumés.
So what does 'zero added' promise?
It promises a category that does not exist in your body. That's the root of the whole pitch: it moves your attention from which molecule, and how much to where the molecule came from — the one dimension the body happens not to read.
For how concrete that is, scene one's example suffices. A cured meat labelled 'no nitrite added' has often switched to cultured celery powder. Celery is naturally rich in nitrate (vegetables are the largest dietary source of nitrate, EFSA 2008), and a bacterial culture reduces it to nitrite. So: what lands on the botulism bacterium's enzymes is the same nitrite ion; and what takes part in forming nitrosamines in your stomach is the same nitrite ion. The industry line is blunt: nitrite is nitrite regardless of the source (Mermelstein 2018).
The label changed; the molecule didn't. The bacterium can't read the label, and neither can your stomach.
So what is the phrase actually selling?
A contrast that doesn't exist. 'This product has zero added preservatives' implies the others added them, and the others are worse. But the preservative the others use has been evaluated, numbered, and capped; whereas this 'zero added' version may simply have a shorter shelf life (so the door from scene one opens a crack again), or may just have switched to a molecule that doesn't count as an 'additive' on the label and is identical inside your body. You paid more and bought reassurance, not safety.
Regulators have seen this too: from March 2027, claims like 'zero added' may no longer be printed on prepackaged food in China (GB 7718-2025). But that only takes the words off the shelf. What actually protects you is knowing why they never promised anything in the first place.
Finally, the whole picture
This page is not a brief for the food industry:
Ultra-processed food is a real problem — and the real problem is the recipe and the structure, scene five.The sugar and sodium in the recipe are a real problem — and those two numbers are printed plainly on the nutrition table.People really do put things into food that don't belong there — and those things appear on no ingredient list at all, scene three.
All of that vigilance is warranted.
What's worth dismantling is the bad reasoning: chemical name equals danger, natural equals safe, contains additives equals toxic. Its flaw isn't excess vigilance — it's vigilance aimed at the one dimension the body doesn't read, a molecule's origin, and therefore missing the three it does: which molecule, how much, and what it came with.
Next time you see those words on a package, you don't need to remember a single regulation. One question is enough: did the molecule change, or only its origin story?
Related: the closest precedent is MSG — dive to msg-glutamate; nitrite in cured meat and the IARC grading are at processed-meat; sweetener ADIs at artificial-sweeteners; NOVA and Hall's trial at ultra-processed-foods; avoiding natural toxins at foodborne-illness; whether organic is safer at organic-food; the full picture of the liver's two steps at hepatic.