Anti-Vegan Arguments
In a public comment section the visible opponent is rarely the real audience. For every person arguing there are many more reading quietly, and those readers are the ones still deciding, which makes the useful move a single clean correction with sources rather than the same argument rebuilt in thread after thread. This page is that correction, kept in one place. The deeper work sits elsewhere: my advocacy page takes 347 of these thoughts and objections apart in far greater depth, and lays out what to actually do with each one.
What follows are a lot of objections to plant-based eating I have been handed lately, collected from comment sections, livestreams, farm pages, carnivore forums and my own inbox, and answered one at a time. Two kinds of objection run through it, and they are not the same animal. A thought-terminating cliché is a phrase whose job is to end the conversation before it starts: circle of life, plants feel pain too, everything in moderation. A functional thought is a real cost the person is actually weighing: the cooking, the money, the family dinner, the bloodwork. The first kind deserves dismantling. The second deserves respect and a real answer. I have marked which is which as we go, and named the logical fallacy where there is one, because learning to spot the shape of an argument is more useful than memorising any single rebuttal. (The distinction, and what to do with each, is the subject of Veganism Rewired.)
The fad frame: “veganism is the unnatural one”
This first cluster is almost entirely cliché. Every item works by asserting that plant-based eating is a recent invention and meat-eating is the human default, so the burden of proof lands on you before a single fact is discussed.
Is veganism a fad diet?
“Vegan is the actual only fad diet here. It has never existed in history.” A fad diet is defined by novelty, rapid social spread, a restrictive rule invented by one personality, and a break with scientific consensus. Vegetarianism appears in Jain, Buddhist and Pythagorean practice thousands of years ago; the word vegan was coined in 1944; and the diet holds a formal position statement from the Academy of Nutrition and Dietetics that runs through 2032. The carnivore diet dates to roughly 2018, spread through podcasts and social media, is built on a single elimination rule, and has no position statement from any national health body anywhere on earth. The accusation is an exact description of the accuser’s own diet.
Fallacy: projection · false equivalence
Has the carnivore diet existed throughout history?
“Carnivore diets have existed and been studied throughout history.” Neither half is true. No human population has ever eaten zero plant food where plant food was obtainable, including the Arctic peoples usually cited, who ate berries, seaweed, roots, lichen and the stomach contents of the animals they killed. And “studied throughout history” collapses entirely on inspection: the first substantial data collection on carnivore eaters is a self-selected online survey from 2021. Age would not settle the question anyway, since bloodletting is older than both diets. The whole carnivore evidence base, counted
Fallacy: appeal to antiquity
If a vegan diet needs supplements, is it really a complete diet?
“It’s impossible without supplements, so it isn’t a real human diet.” B12 is made by bacteria, not by animals. Cattle are dosed with cobalt so their gut microbes can produce it, and pigs and poultry are fed industrially fermented B12 directly, because they cannot make it in confinement. So the choice is not supplement versus no supplement; it is taking the supplement yourself or paying a farmer to route it through an animal first. Meanwhile iodised salt, fortified flour, vitamin D milk and folic acid are so normal nobody calls them a crack in the omnivore diet. The B12 post
Fallacy: appeal to nature · special pleading
Is a vegan diet unnatural?
“It is not how we are meant to eat.” Meant is doing enormous work in that sentence, and it smuggles in a designer. Naturalness is not a health claim: arsenic, tobacco, hemlock and childhood mortality are all perfectly natural, while insulin, clean water and eyeglasses are not. The honest version of the question is whether the diet produces good outcomes, which is an empirical matter with a large and consistent literature behind it. And the ancestral record, when you actually read it, describes a diet that was plant-heavy for most of our lineage’s existence. The flexible ape
Fallacy: appeal to nature
Animals eat other animals, so why shouldn’t we?
“The lion eats the deer. That’s nature.” Lions also kill rival cubs, eat their prey alive, and have no access to a supermarket, a moral vocabulary, or a choice. Nothing an animal does establishes what a creature capable of deliberation ought to do; if it did, we would have to defend infanticide and cannibalism on the same evidence. The relevant difference is not strength or status but necessity: obligate carnivores must kill to survive, and we demonstrably do not.
Fallacy: appeal to nature
Isn’t eating meat just accepting that life feeds on life?
“Meat eaters honour the cycle. Vegans are in denial about it.” This is a false choice between gratitude and refusal, and it quietly assumes those are the only two options. Nobody claims a harm-free diet exists; the claim is that harm can be reduced, and reducing it is not denial but arithmetic. Plant-based eating causes fewer deaths per calorie by every honest accounting, including the field deaths people raise as a counterargument. Reverence is a feeling; it does not alter what happens in the slaughterhouse.
Fallacy: false dichotomy
Can ethics override human biology?
“I support kinder farming, but morality can’t change human biology.” Nothing about this diet asks biology to change, which is the quiet error in the sentence. Human biology already accommodates an extraordinary range of diets, from the nearly all-marine to the nearly all-tuber, and that flexibility is the single most distinctive thing about how our species eats. The question was never whether we can digest plants; it is whether a diet built from them supports health, and that question has been answered repeatedly in the affirmative. Ethics is not being asked to overrule physiology. It is being asked to choose between options physiology already permits. The flexible ape
Fallacy: is-ought conflation · appeal to nature
Isn’t everything fine in moderation?
“Everything in moderation.” This is the purest thought-terminating cliché in the collection: it sounds like wisdom, resolves nothing, and is never applied consistently. Nobody recommends moderate smoking or moderate asbestos, which means the phrase is not a principle but a way of declining to look at a specific dose-response curve. For processed meat and red meat those curves are published, graded, and monotonic in the range most people eat.
Fallacy: thought-terminating cliché
Didn’t even vegetarian cultures always eat some animal products?
“They kept dairy and eggs because it is required for life.” Persistence is not proof of necessity; slavery persisted in most cultures too, and nobody reads that as a nutritional finding. Traditional vegetarian populations retained dairy for reasons of economics, ritual and calorie density, in an era with no fortified foods and no way to isolate a vitamin. The one nutrient those cultures genuinely obtained by accident (B12, from bacterial contamination of water and produce) is now available in a tablet costing pennies. That is a solved problem, not a law of nature.
Fallacy: appeal to tradition
Evolution, anatomy and the archaeological record
Here the objections stop asserting and start arguing, which makes them more interesting and no more correct. Every one of them reads a genuine piece of evidence as though it settled a question it was never designed to answer.
Humans are omnivores, so shouldn’t we eat meat?
“We’re omnivores. It’s what we evolved to do.” Omnivore describes a capability, not an obligation, and the gap between those two words is where the whole argument lives. We are also capable of digesting insects, bone marrow and each other, and nobody treats that capability as a dietary prescription. What evolution equipped us with was flexibility, and flexibility means the choice is genuinely ours rather than dictated. Plants over animals
Fallacy: appeal to nature · capability treated as obligation
Were humans apex predators?
“We were apex predators for two million years.” This comes from a single 2021 hypothesis paper built on indirect proxies (stomach acidity, fat storage, tooth wear) and it was contested by other researchers immediately on publication. Even taken at face value, it describes an ecological niche in a specific glacial environment, not a physiological requirement, and the same literature documents heavy plant use wherever the climate allowed it. Trophic position is a description of what a population ate under constraint, never a prescription for what a supermarket-era human should eat. The flexible ape
Fallacy: cherry-picking · overreach from indirect evidence
Does human anatomy prove we are built for meat?
“Acidic stomach, short colon, tiny cecum, canine teeth: we aren’t built for plants.” Each item is selected and the rest of the body ignored. Our stomach acidity sits within the range of other omnivores and primates; our colon is proportionally long compared with true carnivores; and our salivary amylase gene is present in far more copies than in chimpanzees, which is a starch adaptation written directly into the genome. Canines prove nothing (gorillas and hippos have far larger ones and eat plants), and the appendix is lymphoid tissue and a microbiome reservoir, not a failed cecum. Appendicitis is not caused by plant food; fibre intake is associated with lower risk. The flexible ape
Fallacy: selective evidence
Did meat make our brains big?
“Meat made us human, and brains have shrunk since we started farming.” The expensive-tissue hypothesis is one candidate explanation among several, and the strongest competitors are cooking and starch, both of which make plant calories dramatically more available and both of which are supported by the amylase evidence. Brain volume did decline over the last several thousand years, but body size declined alongside it, and the same pattern appears in every domesticated species, which is why the leading explanations are social and developmental rather than dietary. Neanderthals had larger brains than we do and ate a great deal of meat, which is awkward for anyone using brain size as a scoreboard. The smaller brains claim, in full
Fallacy: single-cause reasoning · post hoc
Doesn’t the archaeological record show a meat-based diet?
“The atlatl and the spear point weren’t for making salad.” This is survivorship bias baked into the ground itself. Stone and bone survive for a hundred thousand years; tubers, greens, nuts and fruit rot within weeks, so a record assembled from what preserves will always look carnivorous regardless of what was eaten. When researchers examine the evidence that does capture plants (starch granules trapped in dental calculus, phytoliths, grinding stones), they find plant processing everywhere they look, including on Neanderthal teeth. Hunting tools survive because they are made of rock, not because salad did not exist. The flexible ape
Fallacy: survivorship bias (taphonomic)
Didn’t modern fruits and vegetables not exist 10,000 years ago?
“Bananas and apples are human inventions. Modern fruit is bred to be a sugar bomb.” True, and it applies with far more force to the other side of the plate. The aurochs became the Angus; the jungle fowl became a bird that reaches slaughter weight in six weeks and cannot walk properly; wild boar became an animal bred for intramuscular fat. Every food in a modern supermarket is domesticated, so “it’s been bred” cannot be an argument against one column and silent about the other. Selective breeding also raised yields and reduced toxicity, which is why we can eat almonds and potatoes at all.
Fallacy: special pleading
Are humans genetic carnivores?
“Humans are genetic carnivores. What we want doesn’t change that.” Our genome argues the opposite at several loci. We carry an expanded salivary amylase gene family for digesting starch, a lactase-persistence mutation that swept through multiple populations in the last ten thousand years, and, decisively, a broken vitamin C gene, meaning we must obtain a vitamin from food that true carnivores manufacture internally or extract from fresh blood and organs. Obligate carnivores are also protected against dietary cholesterol in ways we plainly are not. A genuine genetic carnivore does not develop atherosclerosis from its own natural diet.
Fallacy: assertion presented as settled science
Can’t the brain run on ketones instead of glucose?
“Ketones fuel the brain, so you don’t need dietary carbohydrate.” Ketones can substitute for a substantial share of the brain’s fuel during prolonged fasting, but not all of it, and several tissues (red blood cells, the renal medulla, parts of the brain itself) have an absolute glucose requirement that ketones cannot meet. The body covers that requirement through gluconeogenesis, which means dismantling protein and glycerol to manufacture the very molecule the diet claims is unnecessary. That the body owns an emergency generator is not an argument for running the house on it permanently. Carbohydrate quality, in full
Fallacy: overextension of a starvation adaptation
Ancestral and traditional populations
These are the exhibits offered as proof for the section above, and they share one structural flaw: a whole population is compared against a different population living in a different century under different conditions, and the diet is credited with whatever difference appears.
Were the Inuit healthy on an all-meat diet?
“Coastal Alaskan natives were the healthiest people in North America, and it wasn’t on spring salad.” The claim rests on Bang and Dyerberg’s 1970s fieldwork, whose cardiovascular data came from a registry that systematically undercounted deaths in remote settlements; when researchers went back and checked the records properly, Arctic populations showed coronary disease at rates comparable to or higher than Western ones, along with markedly elevated stroke. The Horus study CT-scanned mummified remains from four preindustrial populations, including Unangan hunter-gatherers, and found atherosclerosis in all of them. Traditional Arctic diets also carried high rates of osteoporosis, and the CPT1A variant that spread through these populations as a fat adaptation has been linked to elevated infant mortality. They were adapted to a hard environment, which is not the same as optimised. The ancestral diet argument, in full
Fallacy: ecological fallacy · cherry-picking
Didn’t the Maasai stay healthy on milk, meat and blood?
“The Maasai ate meat and blood and had clean arteries.” George Mann, the researcher most often cited for the low cholesterol figures, also performed the autopsies, published in the early 1970s, and found extensive atherosclerosis in Maasai aortas, comparable to American men, which is the half of his work that never gets quoted. The men in question also walked enormous distances daily on a modest calorie intake, and few lived into the decades when coronary disease becomes lethal. A population can carry plaque and still die of something else first.
Fallacy: ecological fallacy · selective citation
What about Kitava, Tokelau, Weston Price and the Indian railway workers?
“These traditional populations ate high-fat animal diets and had no heart disease.” Kitava is the strangest one to deploy, because the Kitavans ate a tuber-based, high-carbohydrate diet in which the bulk of the calories came from starch and fruit rather than animal fat; it supports the opposite case. Tokelau’s relevance collapses once you follow the migrants: Tokelauans who moved to New Zealand and shifted diet developed exactly the disease profile the argument claims coconut protects against. Malhotra’s 1967 railway-worker comparison, still circulated as proof that ghee is protective, compared two regions differing in physical labour, smoking, income, and virtually everything else, with no attempt to control for any of it. Coconut oil versus animal fat
Fallacy: cherry-picking · ecological fallacy
Didn’t traditional peoples die young from infection rather than diet?
“They died of infections, not their food. Vegans blame the diet.” This is a fair correction, and it cuts straight through the argument it was raised to defend. If short lifespans in these populations cannot be attributed to diet, then the absence of chronic disease cannot be attributed to diet either: you have removed the ability to draw any dietary conclusion from the population at all. Heart disease and most cancers are diseases of later decades; a population that rarely reaches those decades tells us very little either way. And where researchers have imaged the arteries directly, the plaque is there.
Fallacy: special pleading
Don’t archaeologists agree we ate mostly meat?
“Other fields outside nutrition show we ate a heavily meat-based diet.” No field says this as a consensus, and the evidence from the field most often invoked points the other way. Dental calculus studies, starch-grain analysis and grinding-stone finds have pushed the documented use of processed plant foods steadily earlier, including well into the Neanderthal record. What archaeology actually shows is enormous dietary variation by latitude, season and period, which is the flexibility argument, not the carnivore one. The flexible ape
Fallacy: vague appeal to authority
Nutrient adequacy
This is the most common real-world objection and the one most worth answering carefully, because underneath the sloganeering there is a genuine functional concern: people want to know they will not damage themselves. The short answer is that two nutrients need deliberate attention and the rest of the list dissolves on contact with the evidence.
Where do vegans get protein, and is plant protein incomplete?
“Plant protein is incomplete and you can’t build muscle on it.” Every plant food contains all nine essential amino acids; the differences are in proportion, not presence, and the “complementary protein” rule was retracted by the author who popularised it decades ago because the liver maintains an amino acid pool across the whole day. Soy, seitan and lentils score well on modern digestibility measures, and controlled trials show no difference in strength or lean-mass gains when total protein and training are matched. The practical adjustment is modest: aim a little higher on total protein, lean on legumes, soy and nuts, and the problem disappears. The protein post
Fallacy: outdated premise (the complementation myth)
Are nutrients in plants poorly absorbed?
“How bioavailable is any of that? You’re delusional and malnourished.” Absorption differs by nutrient and mostly in ways that favour regulation rather than deficiency: non-heme iron uptake is controlled by your own stores, so you absorb more when you need more, whereas heme iron is absorbed largely unregulated, which is precisely why it is associated with oxidative stress and colorectal cancer. Zinc absorption adapts upward over weeks on a higher-phytate diet. Across cohort biomarker data, vegans match or exceed omnivores on most micronutrients, with B12 and vitamin D the genuine exceptions. Plants over animals
Fallacy: overgeneralising from a single nutrient
Doesn’t the absence of B12 in plants prove veganism is incomplete?
“B12 doesn’t exist in plant life. That settles it.” It does not exist in animal flesh either, in the sense that matters: no animal synthesises it. B12 is made exclusively by bacteria, and animals accumulate it from their gut microbes or from supplemented feed. In modern agriculture the supplement is standard (cobalt for ruminants, fermented B12 dosed directly into pig and poultry feed), so the modern omnivore is taking a B12 supplement with an animal in the middle of the supply chain. The nutrient is not an argument for eating animals; it is an argument for eating bacteria’s product, and you can do that directly. The B12 post
Fallacy: incomplete comparison
Is B12 a pharmaceutical scam?
“B12 was invented in a Merck lab. Vitamins are refined petroleum. Follow the money.” B12 was isolated in 1948 by teams at Merck and at Glaxo working independently, and the structure was solved by Dorothy Hodgkin, who won a Nobel Prize for it. This is one of the best-documented episodes in twentieth-century chemistry, not a marketing exercise. Commercial B12 is produced by bacterial fermentation, which makes “refined petroleum” chemically impossible. As for following the money: the largest purchaser of the world’s B12 supply is the animal feed industry, which turns the conspiracy theory into an argument against its own conclusion. The B12 post
Fallacy: genetic fallacy · conspiracy reasoning
Can you go without B12 if you feel fine?
“I’m 72, never supplemented, never tested, never sick.” Liver stores buffer B12 for years, and recycling efficiency varies enormously between people, so feeling fine is exactly what early depletion feels like. The damage that matters is neurological and can become irreversible without ever producing anaemia, which is why the nutrient is uniquely suited to this particular error. Someone who is confident they are the exception has a cheap way to prove it, a serum B12 and methylmalonic acid test, and refusing the test while asserting the result is not confidence, it is a coin flip with the outcome face down. The advice to seek out unfortified nutritional yeast is the single most harmful thing in this genre. The B12 post
Fallacy: anecdote · survivorship bias
Do only vegans need to supplement B12?
“Needing a B12 pill is proof the diet is broken.” Then turning fifty is proof the human body is broken, because since 1998 the Institute of Medicine has advised everyone over 50, meat-eaters included, to get most of their B12 from fortified foods or a supplement. Its stated reason is that 10 to 30 percent of older people may be unable to absorb the B12 naturally present in food. B12 in food is bound to protein and needs stomach acid to be released, many older adults make too little, and one common cause, atrophic gastritis, affects about 8 to 9 percent of adults 65 and older. The free B12 in fortified foods and supplements needs no such release, so in most cases it still absorbs. Deficiency is not a vegan peculiarity either: national survey data from 2007 to 2018, as summarised by the NIH, put about 3.6 percent of US adults in deficiency and 12.5 percent in insufficiency, and with vegans a small minority, most of those adults necessarily eat animal foods. The honest difference is timing: vegans need a reliable source from the start, and everyone else is advised to rely on one from fifty. The B12 post
Fallacy: special pleading
Is cyanocobalamin dangerous because it contains cyanide?
“They’re literally telling you there’s cyanide in it.” The cyanide content of a standard tablet is measured in micrograms and is smaller than what you would absorb from ordinary foods, which is why it has never produced toxicity at supplemental doses in the eighty years it has been used. The chemistry runs the opposite way to the fear: cyanocobalamin is what the body produces while clearing cyanide, and the related compound hydroxocobalamin is used in emergency medicine as an actual cyanide antidote. People with specific conditions (some kidney disease, certain rare genetic disorders, heavy smokers) reasonably use methylcobalamin instead, which is a clinical nuance, not a scandal. The B12 post
Fallacy: chemophobia · equivocation on dose
Can vegans get enough DHA without fish?
“Conversion from flax is negligible, vegans have low omega-3 status, and supplements can’t reach the brain anyway.” Conversion from ALA is inefficient but not negligible, it improves when omega-6 intake is moderate, and it is measurably higher in long-term vegans and in women of reproductive age. Where it falls short, algal oil supplies preformed DHA and EPA directly, from the organism fish get theirs from, without the mercury, PCBs or fishing fleet. The claim that only phospholipid-bound DHA can enrich the brain comes from rodent data with no completed human cognitive trial behind it, and, critically, vegans show lower cardiovascular disease rates despite the lower omega-3 index, which is the outcome the marker was supposed to predict. A small algal supplement is cheap insurance; the deficiency narrative is not supported by what happens to the people. The DHA post
Fallacy: mechanism prioritised over outcome
Are DHA supplements a scam or actively harmful?
“DHA supplements are a hundred percent a scam, and they raise LDL, prostate cancer risk and atrial fibrillation.” This argument comes from inside the plant-based community rather than from carnivores, and it follows a consistent pattern: a real citable finding is stretched into an absolute the cited researchers do not endorse. The prostate figure refers to incidence of high-grade disease in an observational analysis, not mortality from supplementation; the atrial fibrillation signal is real but dose-dependent and concentrated at high-dose prescription formulations; and “if you take DHA that’s all you get” is contradicted by the same isotope work it cites, where DHA supplementation raised plasma EPA substantially. At 250 mg a day none of the risk arguments apply. The DHA post
Fallacy: overreach from a citable kernel
Are there nutrients found only in animal foods?
“Creatine, carnitine, carnosine, taurine, retinol, heme iron, D3, choline, DHA and B12: you can only get them from animals.” The list conflates two different categories: nutrients that are essential, and nutrients that are merely present in meat. Your body synthesises creatine, carnitine, carnosine and taurine, which is why no deficiency syndrome exists for any of them in people eating plants; choline is in soy, legumes, cruciferous vegetables and quinoa; D3 comes from lichen or sunlight. Retinol and heme iron deserve a longer look, because both illustrate the same underlying pattern: a preformed nutrient arrives past the gate that would normally regulate it. Eat beta-carotene, and an enzyme in your gut lining cleaves it into retinol only as fast as a feedback loop allows, which is why hypervitaminosis A from carrots is essentially unheard of. Eat retinol directly, from liver or a supplement, and it bypasses that gate entirely, which is why liver-related vitamin A toxicity is a real clinical entity. Heme iron works the same way in reverse: non-heme iron from plants is metered by hepcidin, a hormone that shuts absorption down once your stores are full, while heme iron enters through its own route and barely responds to that signal at all. Iron your body does not need does not sit quietly; it catalyses reactions that damage cell membranes and oxidise LDL, which is the mechanistic link between heme iron intake and higher type 2 diabetes risk in cohort after cohort. The foods that carry heme, red and processed meat, are also linked to colorectal cancer, though whether heme itself is the reason is less settled than people assume: the newest, longest-followed US cohort analysis found its data do not support a strong role for heme iron specifically. That leaves B12, which is bacterial and supplemented into livestock anyway. Two genuine items on a list of ten, both explained by the same principle rather than by animal foods being special. Essential nutrients, in full · How preformed nutrients work, and why the conversion step exists · The colorectal cancer nuance
Fallacy: category error (essential versus merely present)
Does grass-fed beef solve the omega-3 problem?
“Grass-fed beef is a good source of omega-3.” The ratio improves; the absolute amount does not become meaningful. Grass-fed beef contains on the order of tens of milligrams of long-chain omega-3 per serving, against the hundreds of milligrams used in any trial showing an effect, so you would need implausible quantities of beef to reach a dose anyone studies, along with all the saturated fat that comes with it. A quarter-teaspoon of algal oil settles the question at a fraction of the cost and none of the trade-off. Essential nutrients, in full
Fallacy: relative improvement mistaken for adequacy
Do expert bodies refuse to endorse vegan diets for children and pregnancy?
“The Academy narrowed its 2025 position to adults, and some countries advise against it in pregnancy.” The 2025 scope reflects what that particular systematic review covered, not a finding of harm in other groups, and reading a narrowed review scope as a reversal is an argument from silence. The British Dietetic Association, in a formal agreement with The Vegan Society, states that a well-planned vegan diet supports healthy living in people of all ages including pregnancy and breastfeeding, and Dietitians of Canada and Australia’s NHMRC take the same position. A 2004 review in the American Academy of Pediatrics’ journal Pediatrics in Review concluded that multiple experts had independently found vegan diets can be followed safely by infants and children without compromising nutrition or growth; that is a review article rather than an AAP policy statement, and it is two decades old. Not every body agrees, and the disagreement is worth stating rather than hiding: the European paediatric society ESPGHAN is markedly more cautious, urging medical and dietetic supervision for vegan infants and young children and arguing that the North American statements rest on thin evidence. Germany’s DGE, usually cited as the body that advises against, updated its position in 2024 and no longer does; it now says the data are too limited to recommend either for or against in pregnancy, childhood and old age, asks for qualified nutritional counselling because mistakes in those groups can be irreversible, and endorses a well-planned, B12-supplemented vegan diet for healthy adults. That is a caution about execution, which everyone agrees with, not a claim that the diet is inherently unsafe. Essential nutrients, in full
Fallacy: argument from silence
Can you get enough calcium without dairy?
“You need milk for calcium.” You need calcium; the cow is optional. Low-oxalate greens (kale, bok choy, collards, broccoli) are absorbed at around 50 percent against milk’s 32, and fortified plant milk delivers roughly the same absorbable calcium per cup as dairy provided you shake the carton, since most of the added calcium settles out. Calcium-set tofu is the other heavy hitter, and the label tells you which one you bought: calcium sulfate means it counts, nigari means it largely does not. Spinach, chard and beet greens are excellent foods and poor calcium sources, and counting them is the most common way vegans end up short. The bone density post · The dairy post
Fallacy: false necessity
Doesn’t milk build strong bones?
“Milk does a body good. Drink it or your bones pay.” Calcium matters, and EPIC-Oxford suggests low intake is part of why vegans fracture more. Whether milk itself prevents fractures is a separate question, and it has been asked directly. A meta-analysis of prospective cohorts covering 195,102 women and 3,574 hip fractures found no association between milk intake and hip fracture, a relative risk of 0.99 per daily glass (95 percent confidence interval 0.96 to 1.02), with no apparent benefit even at three or four glasses a day (Bischoff-Ferrari 2011). The data in men were thinner, three cohorts and 195 fractures, and pointed to a possible small benefit (0.91, 0.81 to 1.01) that the authors say needs more data; leaving out one large Swedish study, the women’s estimate became a marginal 5 percent lower risk per glass. All of it is observational, as the authors note. The fair summary is that the cow is one calcium source among several, not a fracture shield, and the fix for vegans is the calcium, vitamin D and fortified foods covered above. The dairy post · The bone density post
Fallacy: conflating a nutrient with one food
Isn’t a vegan diet a form of protein poisoning?
“You could grant that she’s promoting rabbit starvation, but so are all vegan diets.” Rabbit starvation is protein poisoning: what happens when someone eats large amounts of very lean protein without enough fat or carbohydrate to cover their energy needs, and the liver hits its ceiling for processing nitrogen. By definition it is an all-meat, low-fat, low-carbohydrate failure mode; the classic cases are trappers eating lean game and nothing else. A vegan diet is moderate in protein and fuelled by carbohydrate, which is the exact physiological opposite of the condition. This one is aimed at the only diet that structurally cannot produce it. The protein post
Fallacy: inverted mechanism
Plant toxins, antinutrients and fibre
The mirror image of the previous section: not “plants lack things” but “plants contain things that are hurting you.” Almost every item here is a real molecule doing real chemistry at a dose nobody eats.
Are plants full of toxins designed to harm you?
“Plants evolved chemicals to poison whatever eats them, and the inflammation almost killed me.” Plants do produce defence compounds, and the observation that they exist tells you nothing about whether they matter at dietary doses. Many of them are precisely the compounds associated with benefit (sulforaphane, polyphenols, allicin), acting through hormesis, a mild stress that upregulates your own antioxidant and detoxification machinery. The “Brussels sprouts are full of carcinogens” line traces to Bruce Ames’s natural-pesticides work, which found that roughly half of all chemicals tested at maximum tolerated doses in rodents come up carcinogenic, synthetic and natural alike, and which Ames himself said should not be applied to what the general population is actually exposed to. The population data run in exactly the opposite direction to the mechanism story. Plant toxins and antinutrients
Fallacy: dose-free mechanism argument
Do oxalates cause kidney stones, brain fog and mitochondrial damage?
“Oxalates caused my stones, fatigue, joint pain and hair loss, and quitting greens turned my health around.” Oxalate matters for one well-defined group: people who form calcium oxalate stones, who are advised to moderate high-oxalate foods and to take calcium with meals so the oxalate binds in the gut rather than the kidney. Outside that group the sprawling symptom list has no clinical literature behind it, and the Organic Acids Test used to diagnose it is a direct-to-consumer panel adapted from newborn screening for inborn errors of metabolism, with no independent validation for the uses it is being sold for; the clinical measure of oxalate status is a 24-hour urine collection. “Oxalate dumping” is the tell: a claim constructed so that feeling worse after quitting plants counts as evidence that plants were the problem, which makes it unfalsifiable. Note also that only a handful of foods are genuinely high-oxalate (spinach, chard, beet greens, rhubarb, almonds); kale, bok choy and broccoli are not, and are routinely eliminated by people following this advice for no reason. Plant toxins and antinutrients
Fallacy: equivocation on dose · unfalsifiable claim · anecdote
Do lectins and phytates damage your gut and block minerals?
“Lectins are the new gluten and phytates are mineral thieves.” The lectins capable of doing harm (concentrated in raw kidney beans) are destroyed by ordinary cooking, and canned or properly boiled legumes contain negligible amounts. Phytate does reduce absorption of iron and zinc in a single meal, and it is also an antioxidant associated with lower colorectal cancer risk, while soaking, sprouting, fermenting and simply eating this way for a few weeks all shrink the effect. The decisive point is outcome rather than mechanism: legume consumption is one of the most consistent predictors of longevity in the entire literature, which is difficult to reconcile with a food that steals your minerals. Plant toxins and antinutrients
Fallacy: mechanism prioritised over outcome
Will kale and soy wreck your thyroid?
“Goitrogens suppress thyroid function.” Goitrogenic compounds interfere with iodine uptake, which means the risk is conditional on iodine deficiency rather than on the vegetable. With adequate iodine (and vegans do need a deliberate source, since plant-based diets without iodised salt or seaweed run low), intervention data show no change in thyroid hormone levels and at most a tiny rise in TSH of unclear significance, concentrated in people whose thyroid was already underactive. The documented harms from cruciferous vegetables involve extraordinary raw intakes: the one published coma case followed more than a kilogram of raw bok choy a day for months. Fix the iodine, and the goitrogen question resolves itself.
Fallacy: dose-free mechanism argument
Do nightshades cause inflammation?
“Tomatoes, peppers and potatoes are inflammatory.” I am not aware of any controlled human trial showing that nightshade vegetables raise inflammatory markers, and the arthritis organisations that have examined the question say the same; the only experimental evidence is decades-old mouse work that newer mouse studies contradict. The food group also includes some of the most consistently protective items in the diet, with tomato intake associated with lower cardiovascular and prostate cancer risk. The claim circulates on self-report and has never been substantiated in a trial. If you individually react to something, eliminate it and reintroduce it deliberately; that is a personal intolerance, not a property of the food group.
Fallacy: anecdote generalised to a food group
Does soy mess with your hormones?
“Soy feminises men.” Isoflavones are structurally similar to estrogen and functionally quite different, binding preferentially to a receptor subtype that often produces opposing effects in tissue. Meta-analyses of controlled trials in men find no effect on testosterone, free testosterone or estrogen at ordinary and even high intakes, though the largest of them comes from authors affiliated with a soy-industry body, which you should know before weighing it, and a 2025 dose-response analysis suggests estradiol may shift at some intakes. The handful of case reports driving the folklore involve men drinking extraordinary daily volumes of soy milk. Meanwhile soy intake is associated with lower breast cancer recurrence, which is the opposite of what an estrogenic food would do. Plant toxins and antinutrients
Fallacy: mechanism prioritised over outcome
Is fibre necessary, or is it a “natural junk food”?
“There’s no essential fibre, breastfed babies don’t eat it, and the fibre trials failed.” No nutrient is essential by fiat; what matters is what happens without it, and the largest synthesis to date found the highest fibre consumers had substantially lower all-cause and cardiovascular mortality along with less heart disease, stroke, diabetes and colorectal cancer. “We don’t digest it” also does not mean “we don’t use it.” Undigested fibre reaches the colon, where resident bacteria ferment it into short-chain fatty acids, chiefly butyrate, which is the preferred fuel source for the cells lining your colon and is separately linked to lower colon cancer growth and reduced inflammation. Skip the fibre and you are not sparing your gut a job it couldn’t do anyway; you are starving the colon of the fuel it runs on. The failed trials tested isolated fibre supplements against polyp recurrence: a reductionist test of a molecule, not of the foods, which is exactly the error the low-carb movement criticises elsewhere. Burkitt himself insisted the protection came from high-fibre foods, and the resistant-starch workaround misses the same point: you cannot bolt one fraction onto a meat-based diet and reproduce a whole-food package. Breastfed infants receive human milk oligosaccharides, which feed the microbiome exactly as fibre does later. Carbohydrate quality, in full
Fallacy: reductionism (nutrient tested in place of food)
Doesn’t your body make its own antioxidants?
“I make glutathione. Antioxidant supplements were debunked anyway.” Both halves are true and neither supports the conclusion. Endogenous antioxidant systems exist and have finite capacity, which is why dietary intake still tracks with oxidative damage markers. The failed trials gave isolated high-dose beta-carotene and vitamin E capsules, which is a test of supplements, not of food, and the same movement that cites those failures rejects reductionism everywhere else. Measured directly, plant foods carry many times the antioxidant content of animal foods, and an animal is not a warehouse storing what plants made; it is a furnace spending much of it staying alive. The anti-inflammatory diet post
Fallacy: false substitution · conflating supplements with food
Do gluten and leaky gut cause autoimmune disease?
“Zonulin opens the gut and that’s where autoimmunity starts.” Coeliac disease is real, affects around one percent of people, and is diagnosed with validated tests; non-coeliac gluten sensitivity is also real and considerably rarer than the internet believes. The leap from there to gluten as a general driver of autoimmunity rests on the zonulin story, and the commercial assay used to measure zonulin has been shown not to detect zonulin at all, capturing complement C3 fragments and properdin instead, a finding published with zonulin’s own discoverer among the authors. Intestinal permeability does change in disease; whether it is cause or consequence remains unsettled. Whole grain intake, meanwhile, is consistently associated with lower mortality. The smaller brains post
Fallacy: unvalidated biomarker · single-cause reasoning
Isn’t liver nature’s multivitamin?
“Nothing on a plant-based plate matches organ meat for nutrient density.” The liver is a filtration and storage organ, which is exactly why it concentrates what you would rather not concentrate: cadmium, copper, and the fat-soluble persistent compounds the animal accumulated over its life. Its vitamin A arrives as preformed retinol, which enters your bloodstream through a different route than the beta-carotene in a carrot and is not subject to the same feedback gate, which is why hypervitaminosis A is a documented clinical entity from liver and cod liver oil and essentially unheard of from vegetables. That is why the NHS tells pregnant women to avoid liver and liver products altogether, and advises everyone else against eating them more than once a week. The regulatory picture makes the irony sharper: legal residue ceilings run far higher on the feed those animals eat than on human produce, and the persistent fraction ends up precisely in the fat and organs this argument recommends. The pesticide argument · The regulatory pattern behind preformed nutrients
Fallacy: nutrient density asserted without context
Aren’t plants full of heavy metals?
“Rice has arsenic, cacao has cadmium, plants bioaccumulate.” Rice arsenic is a genuine issue with genuine mitigations: vary your grains, choose lower-arsenic sourcing, cook rice in excess water and drain it. But bioaccumulation is an argument that runs upward through the food chain, not downward. The reason mercury is a problem in tuna and swordfish, cadmium in kidney, and dioxins overwhelmingly in animal fat and dairy (the WHO attributes more than 90 percent of human dioxin exposure to meat, dairy, fish and shellfish) is that each trophic step concentrates what the level below it absorbed. Naming contamination as a reason to eat higher on the chain inverts the mechanism being invoked. The pesticide argument
Fallacy: inverted bioaccumulation
Aren’t vegetable nitrates as bad as the nitrites in processed meat?
“Same atoms, so celery powder is no better than bacon.” Same atoms, opposite fates, and the fork in the road is chemistry rather than semantics. Nitrite in the presence of heme iron and amino acid breakdown products (that is, in meat) forms N-nitroso compounds, which are carcinogenic. Nitrate in vegetables arrives alongside vitamin C and polyphenols that block nitrosation, and is instead converted to nitric oxide, a vasodilator that lowers blood pressure, which is why beetroot juice shows up in athletic performance trials. Context is not a loophole here; it is the entire reaction pathway. Cholesterol, from the ground up
Fallacy: false equivalence
Carbs, insulin, sugar and weight
The most active front in the argument right now, and the one where a coaching funnel is usually attached to the end. The core move is to treat a macronutrient as a villain and a hormone as a lock.
Does insulin lock fat in your body?
“Carbs raise insulin, and high insulin means body fat is on lockdown.” Insulin does suppress fat release moment to moment, which is not the same as preventing fat loss over days, and the difference is where the carbohydrate-insulin model fails when it is tested rather than described. Controlled feeding studies comparing diets matched for calories and protein find no fat-loss advantage for the lower-insulin arm, and in metabolic-ward conditions the low-fat plant-based arm produced greater body fat loss than the ketogenic one despite far higher insulin. Insulin is not a lock; it is a signal in a system where total energy still governs the outcome. Dietary fat, notably, is stored with almost no processing cost at all. Carbohydrate quality, in full
Fallacy: mechanism prioritised over outcome
Is fruit a sugar bomb you should avoid?
“Modern fruit is bred to be sugar, and no diet needs it.” Sugar inside a piece of fruit behaves nothing like sugar in a glass, because the fibre, water and cell walls slow absorption enough to blunt the glucose response, which is why whole fruit intake is associated with lower type 2 diabetes risk while fruit juice goes the other way. The quantities also deserve a look: a medium apple carries roughly half the sugar of a can of soda, alongside fibre, potassium and polyphenols the soda does not contain. Compressing all of that into the word “sugar” is the composition fallacy doing the work an argument should be doing. Carbohydrate quality, in full
Fallacy: composition fallacy
Isn’t it true that there is no essential carbohydrate?
“The National Academies say the lower limit of dietary carbohydrate compatible with life is zero.” That sentence exists, and the same report continues a few lines later to note that the brain’s glucose requirement must then be met by gluconeogenesis from protein and glycerol, which is quoting the premise and stopping before the conclusion. “Essential” in nutrition science means the body cannot synthesise it, not that it is optional or unimportant: by the same standard there is no essential fibre, no essential vitamin C for a dog, and no essential vegetable. Survival is a low bar, and it is not the bar anyone is actually interested in clearing. Essential nutrients, in full
Fallacy: equivocation on “essential” · half-quotation
Do carbs cause visceral fat and diabetes?
“Visceral fat is deposited from carbs. It’s well documented.” What drives insulin resistance at the cellular level is fat accumulating inside muscle and liver cells where it does not belong, interfering with insulin signalling, and a randomised trial of a low-fat vegan diet cut liver fat by roughly a third and intramyocellular lipid by about a tenth while improving insulin sensitivity. That is the diet supposedly made of the problem draining the thing causing it. Refined carbohydrate in an energy surplus certainly contributes to fatty liver; whole intact carbohydrate in populations that eat the most of it does not. The causal arrow here is widely drawn backwards. Carbohydrate quality, in full
Fallacy: reversed causation
Do grains cause Alzheimer’s?
“Grain brain.” This is a trade-book thesis rather than a research finding, and the cohort evidence points the other way: whole grain intake is associated with lower cardiovascular and all-cause mortality, and the dietary patterns with the best cognitive outcomes in trials (Mediterranean and MIND) both include whole grains as a named component. Vascular health is among the strongest modifiable inputs to dementia risk, and grains help rather than hurt it. Carbohydrate quality, in full
Fallacy: single-cause reasoning
Does sugar feed cancer?
“Starve the tumour of sugar.” Every cell in your body runs on glucose, tumours included, and if you stop eating carbohydrate your liver manufactures it, so the fuel line cannot be cut by diet. Cancer cells also metabolise glutamine and fatty acids readily, which means restricting one substrate redirects rather than starves them. Ketogenic diets have been trialled as adjuncts in oncology, and those trials are small and short, measure glucose, weight and quality of life rather than survival, and leave survival outcomes inconclusive; the largest randomised trial in brain cancer missed its primary endpoint. Unintentional weight loss, meanwhile, is among the worse prognostic signs a cancer patient can have. The real dietary association is with obesity, which raises risk for over a dozen cancers. Carbohydrate quality, in full · Cancer and diet, in full
Fallacy: mechanism prioritised over outcome
Doesn’t keto or carnivore reverse type 2 diabetes?
“It reverses diabetes, and plant-based almost never does.” Removing carbohydrate lowers blood glucose immediately for the same reason removing water lowers a flood: it manages the symptom by withholding the substrate, which is genuinely useful and is not the same as restoring insulin sensitivity. This is why long-term very-low-carb eaters can fail an oral glucose tolerance test badly: the underlying defect is intact and merely unchallenged. Low-fat plant-based trials improve glycaemic control while increasing carbohydrate, which is the harder demonstration, and they do it by clearing lipid out of muscle and liver cells. Weight loss also improves diabetes on any diet, which is why weight-matched comparisons matter more than testimonials. The weight loss post
Fallacy: conflating symptom control with reversal
Don’t the trials show low-carb wins?
“Keto lost more weight in the ward study, DIETFITS was a tie, and the Twinkie diet proved calories are all that matter.” In the metabolic-ward comparison the scale favoured keto while the body composition ran the other way: participants lost body fat on the plant-based arm and largely fat-free mass and water on the ketogenic one, which is the finding the retellings omit. DIETFITS compared a healthy low-carb diet against a healthy low-fat one, both built on whole foods, and the tie is a result about food quality rather than a defeat for either macronutrient. The Twinkie demonstration shows that an energy deficit drives weight loss, which nobody disputes, and the professor’s LDL fell, because that is what happens to anyone who drops twenty-seven pounds in ten weeks. One person over ten weeks measures the weight loss, not the diet. The weight loss post
Fallacy: cherry-picking endpoints
Doesn’t “bonking” prove we need carbs, or the dog that got pancreatitis prove starch is bad?
“Endurance bonking is a glucose problem, and dogs don’t make amylase, so starch is unnatural.” The first argument concedes the case: bonking is glycogen depletion, which is a demonstration that high-intensity work runs on carbohydrate. The second is false on the biology: dogs produce pancreatic amylase, and one of the best-documented signatures of dog domestication is expanded copy number of the starch-digesting AMY2B gene relative to wolves, a several-fold expansion first reported in Nature in 2013. Pancreatitis in dogs is classically associated with high-fat food, not starch. A single animal anecdote from a different species is not a mechanism. The vegan cardiologist video, answered
Fallacy: false analogy · anecdote
Is vegan food just addictive carbs that damage you slowly?
“It tastes good and you feel great, and that’s how a nice-tasting poison works.” Notice the structure: feeling well is reinterpreted as evidence of harm, which means no possible observation could count against the claim. Unfalsifiable arguments are not strong arguments; they are arguments that have stopped being about evidence. Hyperpalatability is a real phenomenon and it belongs to engineered ultra-processed food, not to lentils, oats and broccoli, which are among the most satiating foods measured. If a slow poison is the hypothesis, the place to look is the long-term cohort data, and there the plant-based eaters are not the group accumulating disease.
Fallacy: unfalsifiable claim
Doesn’t the Randle cycle prove carbs block fat burning?
“It’s called the Randle Cycle, genius.” The Randle cycle is real and it does not say what it is being used to say. Described by Philip Randle in the 1960s, it explains substrate competition: when fatty-acid oxidation rises, glucose oxidation falls, and vice versa. That is a statement about which fuel a cell burns at a given moment, not a demonstration that insulin locks body fat away or that eating fruit prevents fat loss. The test is what happens to whole humans: in the metabolic-ward comparison, the higher-carbohydrate arm ran higher insulin and still lost more body fat while spontaneously eating around 689 fewer calories a day. Naming a metabolic pathway is not the same as showing an outcome, which is also the answer to its companions: “carbohydrate toxicity,” “Randle cycle deregulation,” “deuterium accumulation.” None of them arrive with a population that got sick. Carbohydrate quality, in full
Fallacy: mechanism prioritised over outcome · appeal to jargon
Doesn’t a CGM spike after fruit prove fruit is basically candy?
“Just watch someone eat a banana and watch their glucose spike like they’ve eaten candy.” A glucose excursion tells you glucose entered the bloodstream, which is what is supposed to happen when you eat carbohydrate, and in a healthy pancreas it clears within the hour. Reading a normal physiological event as pathology because a number moved is the core error of consumer glucose monitoring. The measurements also contradict the “fibre barely matters” version: whole fruit sits low on the glycemic index precisely because fibre, water and intact cell walls slow absorption: an apple around 36 and a banana around 51 against table sugar at 65 and glucose at 100. And if fruit were candy, the heaviest fruit eaters would look diabetic; instead higher whole-fruit intake tracks with lower type 2 diabetes risk, while fruit juice (the same sugar without the structure) goes the other way. That contrast only exists because the fruit is not candy. Carbohydrate quality, in full
Fallacy: surrogate marker mistaken for outcome
Isn’t dietary sugar glycating your arteries over the decades?
“It’s still glycating and inflammatory and will eventually lead to disease.” Glycation is real chemistry, and the leap from “sugars participate in glycation reactions” to “whole fruit damages your arteries” skips the entire distance between a reaction and a disease. In people without diabetes, glycation is controlled (that is what a normal HbA1c means), and atherosclerosis in non-diabetics is driven by ApoB particles carrying cholesterol into the artery wall, not by blood glucose. The population check finishes it: the Tsimané take roughly 70 percent of their calories as carbohydrate and have the lowest coronary atherosclerosis ever measured in humans, with about 85 percent showing zero coronary calcium. They are the one population that could not exist if this mechanism worked as described. Dietary advanced glycation end products, incidentally, are highest in high-protein, high-fat foods cooked with dry heat, which is to say meat far more than plants. Cholesterol, from the ground up
Fallacy: mechanism prioritised over outcome
Is anything over the 4 grams of glucose in your blood excess sugar?
“Normal blood sugar is 4 to 5 grams of glucose total, so everything above that is excess causing inflammation.” That figure is correct and it is a snapshot of the pipe, not a limit on the flow. You hold 4 to 5 grams in circulation precisely because muscle and liver pull glucose out as fast as food puts it in: the steady number is evidence the system is working, not evidence it is overwhelmed. It is like counting cars on one stretch of motorway and concluding the road can carry twenty vehicles a day. A healthy adult oxidises hundreds of grams of glucose daily while that circulating pool barely moves. Carbohydrate quality, in full
Fallacy: equivocation between a stock and a flow
Fat, cholesterol, heart disease and seed oils
The biggest scientific fight in the argument and the one where the evidence is deepest, because cholesterol’s role in heart disease has been tested at every level from cell biology to human genetics to drug trials.
Are meat and fat heart-healthy, with sugar and seed oils the real culprits?
“Sugar, gluten, grains, alcohol and industrial seed oils cause heart disease. Meat and fat don’t.” This is a true item smuggling four false ones: refined sugar and alcohol genuinely do damage, and bundling saturated fat’s exoneration into the same sentence is what the list is for. The test of the claim is what happens when you substitute one thing for another, because you always eat something instead, and the substitution evidence is consistent. The American Heart Association’s 2017 presidential advisory pooled the four best-controlled trials that replaced saturated fat with polyunsaturated vegetable oil and found 29 percent fewer coronary events (relative risk 0.71, 95 percent confidence interval 0.62 to 0.81), while trials that replaced it mainly with carbohydrate, often unspecified but typically including sugar and other refined carbohydrate, did not significantly reduce cardiovascular disease. The Nurses’ Health and Health Professionals cohorts (Li 2015) show the same split: replacing 5 percent of energy from saturated fat with polyunsaturated fat, monounsaturated fat or whole-grain carbohydrate went with 25, 15 and 9 percent lower coronary risk, while replacing it with refined starch and added sugar showed no significant association (1 percent higher). Some of the advisory’s writing group disclose industry ties, including pharmaceutical companies, walnut and canola oil groups, Unilever and, for one author, the National Cattlemen’s Beef Association. Bundling is a rhetorical technique, not a finding. Cholesterol, from the ground up
Fallacy: bundling (true claims carrying a false one)
Are seed oils inflammatory?
“Linoleic acid drives inflammation, and chronic disease rose right alongside seed oil consumption.” When linoleic acid is fed to people in controlled trials, inflammatory markers do not rise. The systematic review that established this was industry-funded, which is worth saying out loud, and the independent evidence points the same way: biomarker studies measuring tissue linoleic acid rather than asking people what they ate find higher levels associated with lower cardiovascular disease and lower type 2 diabetes risk. The timeline argument is the oldest correlation error there is: seed oil consumption rose alongside cigarettes, sugar, calories, cars and air conditioning. The mechanism sounds compelling and the outcome data refuse to follow it. Saturated fat, seed oils and heart disease
Fallacy: mechanism prioritised over outcome · post hoc
Are beef tallow and butter better than plant oils?
“Tallow is the ancestral fat. Seed oils are industrial poison.” Fats have to be compared against what replaces them, and every time that comparison is run, liquid plant oils come out ahead of solid animal fats on the marker that causally drives atherosclerosis. A 2026 clinical review in JACC: Advances placed beef tallow in its “evidence of harm” category for raising LDL cholesterol, put seed oils and seafood in “evidence of benefit”, and concluded that no evidence supports using beef tallow as a healthier alternative to plant oils. Butter is the worst common fat for LDL because it is dense in myristic and palmitic acid and carries naturally occurring trans fat; tallow and lard are less bad, largely because of stearic acid and oleic acid, and still worse than any liquid plant oil. Beef tallow versus plant oils
Fallacy: appeal to nature · false dichotomy
Is coconut oil a health food?
“It’s an MCT oil, and it raises your good cholesterol.” Coconut oil is not an MCT oil, and this is the load-bearing error: true medium-chain fats make up only around an eighth of it, while lauric acid, nearly half of it, is packaged into chylomicrons during digestion and behaves like a long-chain saturated fat. Studies of purified MCT oil therefore cannot be transferred to the jar in your cupboard. It does beat butter on LDL, which is a low bar, and it raises HDL, which has repeatedly failed to translate into fewer cardiovascular events. Coconut oil versus animal fat
Fallacy: category error
Isn’t cheese heart-neutral because of the food matrix?
“Full-fat dairy is fine; the cheese matrix cancels out the saturated fat.” The best-known trial behind this claim shows how the comparison is built. A 2015 crossover trial fed 14 overweight postmenopausal women a high-cheese diet, a meat diet matched to it for saturated fat, and a low-fat, high-carbohydrate diet for two weeks each, and found LDL cholesterol and ApoB similar on all three (Thorning 2015). The low-fat comparison diet was hardly a lean benchmark: about a quarter of its saturated fat came from coconut milk and coconut fat, another quarter from chocolate and a tenth from sweetened biscuits. The trial was funded by dairy organisations in five countries and a Danish farm-industry council, and one author was on a dairy industry advisory board. “Cheese is no worse than a coconut-and-chocolate diet” is a different claim from “cheese is heart-neutral.” The matrix point is partly real: trials cited in a large Harvard cohort analysis found that the LDL-raising effect of dairy fat in cheese was smaller than that of butter at matched fat intake, so cheese is plausibly less bad than butter. Less bad than butter is not neutral, though, and the same analysis shows what the comparison does to the answer. Across three cohorts of 43,652 men and 178,582 women, with 14,815 cardiovascular events over more than 5 million person-years, replacing 5 percent of energy from carbohydrate (other than fruit and vegetables) with dairy fat was not significantly associated with cardiovascular disease (relative risk 1.02, 95 percent confidence interval 0.98 to 1.05), but replacing 5 percent of energy from dairy fat with polyunsaturated fat went with 24 percent lower risk (0.76, 0.71 to 0.81) and with vegetable fat 10 percent lower (0.90, 0.87 to 0.93) (Chen 2016). The authors note that their data could not test the food matrix and that residual confounding is possible. The fair reading is that dairy fat looks no different from that carbohydrate and worse than unsaturated fat, which is what the substitution evidence above predicts. The dairy post · The dietary fat post
Fallacy: cherry-picked comparator
Has saturated fat been exonerated?
“Cochrane, Siri-Tarino, Chowdhury and PURE all cleared it.” Those analyses mostly asked whether eating less saturated fat helps without specifying what takes its place, and the answer depends entirely on the replacement. The Women’s Health Initiative cut total fat and replaced it with carbohydrate, alongside more vegetables, fruit and grains, and found no significant effect on coronary events or stroke; as the American Heart Association’s 2017 presidential advisory notes, its participants fell well short of the fat goal (24 percent of energy after the first year and 29 percent by year six, against 35 and 37 percent in the comparison group), and it was not a test of cutting saturated fat alone. Trials that swapped saturated fat for polyunsaturated vegetable oil found 29 percent fewer coronary events, and in the large Nurses’ Health and Health Professionals cohorts (Li 2015) replacing it with refined starch and added sugar showed no significant association with coronary risk. The Cochrane review found a 17 percent reduction in combined cardiovascular events from cutting saturated fat, across twelve trials and nearly 54,000 people, even though all-cause mortality did not move, which is a different finding from exoneration. It is also not the source of the refined-versus-whole distinction: its 2020 subgroups found carbohydrate replacement (relative risk 0.84, 0.67 to 1.06) not significantly different from polyunsaturated replacement (0.79, 0.62 to 1.00), with no data separating refined from unrefined carbohydrate. The Sydney and Minnesota reanalyses used margarines of that era, some containing trans fat, which is not a test of modern plant oils. And PURE’s own companion paper in the same issue found higher fruit, vegetable and legume intake predicted lower mortality. Saturated fat, seed oils and heart disease
Fallacy: cherry-picking · ignoring the replacement nutrient
Isn’t heart disease caused by inflammation rather than cholesterol?
“ApoB is at the scene of the crime, like a fire engine at a fire. And 75 percent of heart attack patients had normal LDL.” The fire-engine analogy fails at the one place it needs to hold: fire engines do not cause fires, and people born with genetically lower LDL have proportionally less heart disease across their whole lives, which is causation demonstrated by random allocation at conception. “Normal” in that statistic means normal for a population in which heart disease is the leading cause of death: the mean LDL in that cohort was around 105 mg/dL, far above the range seen in populations that do not develop the disease. Inflammation and apoB are not competing explanations; retained apoB particles are what the inflammation is responding to. Cholesterol, from the ground up · How plaque forms
Fallacy: false dichotomy · misreading a reference range
Does dietary cholesterol matter, and are eggs fine?
“Your liver makes most of it, and they dropped the 300 mg limit in 2015, so it was never real.” Dropping a milligram cap was a shift to food-pattern guidance, not a finding of no effect, and the same documents kept the saturated fat limits in place. The liver does downregulate synthesis when you eat cholesterol, but the compensation is partial and highly variable. Hyper-responders exist, and they have no way of knowing they are one without testing. The genuinely unsettled part is eggs specifically, where the cohort literature is mixed and honest scientists still disagree; the part that is not unsettled is that the overall package of cholesterol-and-saturated-fat-rich animal foods raises apoB. Cholesterol in your brain and hormones is manufactured locally and does not come from your breakfast. Cholesterol, from the ground up
Fallacy: half-quotation · appeal to a reversed guideline
Doesn’t high HDL or a good triglyceride ratio protect you?
“My HDL is great and my trig-to-HDL ratio is perfect, so the LDL doesn’t matter.” HDL is the clearest cautionary tale in cardiology: genetic variants that raise it do not lower heart disease risk, and every drug class developed to raise it has failed to reduce events. Triglyceride-to-HDL ratio is a useful marker of insulin resistance and is not a substitute for the particle count that drives plaque. The large-buoyant-versus-small-dense distinction is likewise real and secondary: both carry apoB, and the particle number is what predicts. A good ratio alongside an apoB of 150 is a comfortable number sitting next to the dangerous one. Cholesterol, from the ground up
Fallacy: surrogate-marker substitution
Don’t older people with high cholesterol live longer?
“The U-shaped curve shows low cholesterol kills.” Serious illness lowers cholesterol: cancer, liver disease, malnutrition, frailty and inflammation all drive it down before they kill people, so in an elderly cohort the low-cholesterol group is enriched with the already-sick. That is reverse causation, and it is exactly why the question is settled with genetics and randomised drug trials rather than snapshots of old people. In those designs, lowering LDL reduces events at every age tested, including in the elderly. Cholesterol, from the ground up
Fallacy: reverse causation
Is very high LDL harmless if you’re lean and metabolically healthy?
“Lean mass hyper-responders run LDL of 300 with perfect glucose, and KETO-CTA showed ApoB doesn’t predict plaque.” The imaging study behind this claim was funded by the Citizen Science Foundation, whose founder and president, Dave Feldman, is a coauthor and discloses Patreon membership contributions. The journal retracted the paper on March 11, 2026, at the request of the authors and the Editors, because post-publication concerns about the methodology affected the reliability of the data and the errors were too large for a corrigendum. The authors’ January 2026 preprint still concludes that ApoB failed to predict who progressed, which is close to meaningless in a cohort where everyone’s ApoB was extremely high already, like finding that height fails to predict basketball skill in a room of seven-footers. Meanwhile the participants’ plaque progressed: in that same preprint, by the pre-specified method, 21 percent had an annual rise in percent atheroma volume of 1 percent or more, the preprint’s marker of rapid progression. Feeling well and having excellent glucose numbers are not measurements of your arteries, which is what makes this whole category of claim so dangerous. Cholesterol, from the ground up · How plaque forms
Fallacy: restricted-range reasoning
Isn’t heart disease genetic rather than dietary?
“It’s Lp(a) and familial hypercholesterolemia. Diet is a sideshow.” Lipoprotein(a) is genuinely genetic, largely unresponsive to diet, and the honest asterisk on any dietary argument: everyone should know their number once. Whether lowering Lp(a) itself prevents heart attacks is still unproven: in September 2026 Novartis announced that its drug pelacarsen lowered Lp(a) but did not reduce major cardiovascular events in a phase 3 trial of about 8,300 people with existing cardiovascular disease, full results are due in November, and cardiologists are divided on whether that reflects the drug, the trial design or the hypothesis (JAMA news report). It is also additive rather than exclusive: the particles that diet does move still drive risk, and people with familial hypercholesterolemia are the clearest demonstration that lifetime apoB exposure causes disease, not evidence that food is irrelevant. Genetics loads the gun in a mechanism where diet supplies a large share of the ammunition. Cholesterol, from the ground up
Fallacy: false dichotomy
Is TMAO just a marker rather than a cause?
“Fish raises TMAO more than steak does, so the whole story falls apart.” This is the strongest live objection in the section and it deserves to be stated rather than skipped: fish does contain preformed TMAO and raises blood levels substantially without an apparent cardiovascular penalty, and whether TMAO drives disease or merely tracks with it remains genuinely disputed. What the objection does not do is rescue the diet, because TMAO was always a supporting mechanism rather than the case: the case is apoB, and it stands on genetics and drug trials that have nothing to do with gut microbes. Notably, vegans barely produce TMAO even when fed carnitine directly, because the microbial machinery is absent. Cholesterol, from the ground up
Fallacy: single counterexample treated as refutation
Isn’t cholesterol essential for your brain and hormones?
“The brain is 60 percent fat and built from cholesterol, so you have to eat it.” Your brain holds roughly a quarter of your body’s cholesterol and manufactures every molecule of it locally, because the blood-brain barrier blocks cholesterol arriving from the bloodstream. Not one molecule of the cholesterol in your brain came from an egg. The same goes for steroid hormones: your body synthesises the cholesterol it needs, which is why there is no dietary requirement for it and no deficiency syndrome anywhere in the literature. You are not built from what you eat in any literal sense, or bread would make you a loaf. Cholesterol, from the ground up
Fallacy: you-are-what-you-eat
Isn’t meat non-inflammatory according to the inflammation index?
“The dietary inflammatory index doesn’t rank meat badly, so the anti-inflammatory case for plants is marketing.” The index scores foods by their measured content of parameters chosen when it was built, which means it is an instrument with a scope rather than a verdict on reality. Several well-documented inflammatory mechanisms in animal foods (endotoxin absorbed after a fatty meal, Neu5Gc incorporation, advanced glycation end products formed in cooking) are not among its parameters, so a food can score neutrally while doing things the index never measured. Mistaking the instrument for the territory is a specific and common error. The anti-inflammatory diet post
Fallacy: instrument mistaken for the territory
Isn’t chicken fine, and isn’t the meat-cancer classification overblown?
“White meat is the healthy meat, and putting bacon in the same category as tobacco is absurd.” The IARC groups classify strength of evidence that something causes cancer, not how much cancer it causes: tobacco and processed meat sit in the same group because both are established carcinogens, at wildly different magnitudes. Mocking the grouping is mocking a misunderstanding of what the grouping means. On chicken: a controlled feeding trial found white meat raised LDL about as much as red meat did, while swapping either for plant protein lowered it substantially, so “at least it’s not beef” is a comparison against the wrong benchmark. The dietary fat post · The processed meat post
Fallacy: conflating hazard with risk
The inconvenient findings
Real data, real journals, aimed at us. These deserve the most careful treatment in the post, because waving them away is how you lose a reader who has actually read the paper.
Do vegans have a 40 percent higher risk of colorectal cancer?
“A 2026 analysis of 1.8 million people found vegetarians get more bowel cancer. That’s a very obvious problem for you.” The headline figure comes from a subgroup rather than the main result, and the pattern inside the paper is the giveaway: the signal concentrates at one anatomical subsite rather than appearing across the colon, which is not how a dietary carcinogen behaves. Vegetarian and vegan categories are also small, self-assigned and unstable over decades of follow-up, so misclassification runs high exactly where the estimate is most fragile. Set against it sits the evidence the claim is meant to overturn: hundreds of studies reviewed by IARC, a dose-response relationship, and a well-characterised mechanism involving heme iron and N-nitroso compounds. One subgroup does not outweigh a body of evidence built at every level. The colorectal cancer study, answered in full
Fallacy: subgroup cherry-picking
Do vegans break more bones?
“EPIC-Oxford found more than double the hip fracture risk, and a 2025 meta-analysis put vegans 75 percent higher.” This is the strongest health finding the other side has, and the honest answer is that it is real, narrower than the headline, and fixable. The entire EPIC-Oxford hip signal rests on 41 vegan fractures, and the pooled 2025 estimate carries a confidence interval running from 1.17 to 2.63, graded very low certainty by the reviewers’ own standard, which formally means the true effect is likely to differ substantially from the estimate. Nor is it mainly about protein, the explanation most often offered: in EPIC-Oxford, adjusting for dietary protein moved the vegan hip-fracture hazard ratio only from 2.31 to 2.21, adjusting for calcium moved it to 1.97, and vegans who met the protein guideline still had 2.71; the authors conclude that calcium and protein at most partly explain the gap, and they had no data on calcium supplements (Tong 2020). The strongest lead comes from the Adventist cohort, where vegan women taking neither calcium nor vitamin D carried a hazard ratio near 3.0, while those taking both had no excess risk at all. Diet quality matters too: in the Nurses’ Health Study, long-term plant-based diet quality was not associated with hip fracture in either direction, while in a secondary analysis of recent diet a healthful plant-based pattern went with 21 percent lower risk and an unhealthful one with 28 percent higher; that cohort scored omnivores and vegetarians alike on a plant-based index rather than studying vegans (Sotos-Prieto 2024). The finding is not an indictment of the diet; it is an instruction about calcium, vitamin D, protein, body weight and resistance training. The bone density post
Fallacy: imprecision reported as a finding
Do vegetarians have more strokes and dementia?
“EPIC-Oxford showed more haemorrhagic stroke, and the Adventist data show worse neurological outcomes in older vegetarians.” Both signals exist and neither has been cleanly attributed to the diet pattern itself rather than to specific gaps within it. The candidate explanations are all correctable: low B12 raising homocysteine, low vitamin D, and very low LDL in cohorts recruited decades ago when supplementation was uncommon and unfortified eating was the norm. The stroke finding in EPIC-Oxford sat alongside substantially lower ischaemic heart disease in the same cohort, which is the far more common event. What this argues for is supplementing properly, not eating differently. The evidence review
Fallacy: nutrient gap attributed to the whole pattern
Do vegans have smaller brains?
“An Oxford study found six times the brain shrinkage in meat-free eaters.” The study did not compare vegetarians with meat-eaters; it compared people by B12 status, and the participants with the lowest B12 lost the most brain volume. The finding is about a correctable deficiency that also occurs in omnivores, mainly the elderly and people on metformin or acid blockers. Reported as a fact about vegetarians, it is a misattribution of the exposure variable, and it happens to be an argument for the supplement the same movement calls a scam. The smaller brains post
Fallacy: misattributed exposure variable
Do vegan mothers cause brain damage in their babies?
“There are documented cases of brain atrophy in breastfed infants of vegan mothers.” Those case reports are real, they are serious, and every one of them involves an unsupplemented mother with undiagnosed B12 depletion; the same clinical picture appears in infants of omnivorous mothers with pernicious anaemia or after bariatric surgery. The cause is the deficiency, not the diet that made the deficiency more likely in someone who ignored the one instruction the diet carries. This is the single strongest reason to treat B12 denial as dangerous rather than eccentric. Supplemented pregnancy and breastfeeding do not produce these cases. The B12 post
Fallacy: cause misassigned
Do vegan pregnancies produce smaller babies?
“Vegan mothers have lower birth-weight, small-for-gestational-age babies.” The pooled evidence does show lower average birth weight in vegan pregnancies, drawn from a handful of small, heterogeneous studies whose own reviewers say they are too underpowered to support firm conclusions. Where the data allow a closer look, the gap tracks with inadequate energy, protein, iron, B12 and vitamin D rather than with the exclusion of animal products as such, and B12 supplementation normalised maternal and cord levels. Confounding by lower maternal BMI and by under-eating during a period of high requirement is difficult to remove from observational data. The practical reading is the same as everywhere else in this section: pregnancy is the time to be deliberate, tested and supplemented, not the time to improvise.
Fallacy: confounding treated as causation
Are vegan children shorter and more deficient?
“The Polish study found vegan children around three centimetres shorter with lower bone mineral content, more B12 deficiency and more anaemia.” The height and bone findings are real and the deficiency findings were concentrated in children whose supplementation was inadequate. Children are the group where the margin for casual planning is smallest, because they are building tissue on a fixed developmental timeline and their energy and amino acid needs per kilogram run well above an adult’s. That is an argument for running protein targets deliberately higher, for fortified foods, for calcium and vitamin D and B12 as non-negotiables, and for paediatric follow-up, not for abandoning the diet. Note also that the adult defence that lower bone density is largely a body-size artifact does not transfer to growing children. The bone density post
Fallacy: supplementation status confounded with diet
Are vegans more depressed?
“Meta-analyses find more depression and anxiety in vegetarians and vegans, and vegetarianism tracks with eating disorders in teenagers.” The meta-analyses disagree, and the details matter. The one that found more depression, Dobersek 2021, pooled 20 studies covering 171,802 people and found that people who ate meat reported less depression than people who avoided it, a small difference by the usual convention (Hedges’ g 0.22), and the more rigorous the study, the more consistent the link; it was funded in part by an unrestricted research grant from the Beef Checkoff, and its own authors write that the evidence “precludes causal and temporal inferences.” Askari 2020, pooling 10 studies of adults, found no association between a vegetarian diet and depression (pooled estimate 1.02, 95 percent confidence interval 0.84 to 1.25). And Iguacel 2021 found no significant difference in depression scores across nine studies; its often-quoted odds ratio of 2.14 for depression came from only two studies using unadjusted data, a result its authors call tentative (one of its authors was supported by the Alpro Foundation, named for a plant-based food brand, which the authors say was not involved in the work). Nearly all of this is snapshot data, and the direction of causation is not established: people experiencing distress change their diets. The teenage link is real as an association: a systematic review of 20 studies with 14,391 adolescents and young adults found that most reported significant correlations between vegetarianism and eating disorders, though its authors say the cross-sectional designs cannot establish cause (Sergentanis 2021). The evidence on which comes first leans one way without settling it. In a study of 93 women with an eating-disorder history and 67 controls, most of those who had been vegetarian said the eating disorder came first, which the authors say fits research suggesting vegetarian diets may be used to facilitate ongoing restriction rather than cause it, while calling for prospective studies (Bardone-Cone 2012); an earlier study cited in the same paper found about half of its anorexia patients had become vegetarian years before the disorder began. The practical answer comes from Project EAT, which found more binge eating among current vegetarians and more extreme weight-control behaviour among former vegetarians aged 15 to 23: its authors advise clinicians to ask about current and former vegetarianism when assessing risk of disordered eating (Robinson-O’Brien 2009). The site adds one more possibility, offered as an ethical hypothesis rather than a finding: some of the distress may be moral distress at what you now see clearly, which is not a psychiatric symptom. No study we have reviewed has measured it. Whatever the explanation, the practical steps are the ones a well-planned vegan diet needs anyway: a reliable source of vitamin B12, vitamin D from sunlight, fortified foods or a supplement, ALA-rich foods such as walnuts, flax and chia along with a look at whether algae-based DHA makes sense for you, meals built on whole foods rather than refined ones, people around you who share your values, and a clinician if low mood lasts.
The two trials usually cited are weaker than they sound. A workplace study at 10 offices of the insurance company GEICO, with 292 employees who were overweight or had type 2 diabetes, found that after 18 weeks depression and anxiety scores improved slightly more in the offices given weekly instruction in a vegan diet, by about 3.6 to 3.7 points on 0-to-100 scales after adjustment (Agarwal 2015). Its authors, from the Physicians Committee for Responsible Medicine, which promotes plant-based diets, call it quasi-experimental: offices rather than individuals were randomised, only the vegan offices had weekly group meetings, and they write that “the mere presence of an intervention could account for the observed effects.” In a two-week pilot trial of 39 omnivores, the group that cut out meat, fish and poultry (a vegetarian diet, not a vegan one) improved on stress and confusion, but its depression scores did not change (Beezhold and Johnston 2012). So neither trial shows that cutting out meat worsens mood, and neither shows that it treats depression. That intervention evidence is thin, and it is what there is. Worth saying plainly to anyone in it: if you are struggling, that is worth real support, and it is not a verdict on what you eat. Vegan diets and mental health, in full · The B12 post · The DHA post
Fallacy: reverse causation
Is there an unexplained neurological risk in the oldest vegans?
“Even Adventist researchers report a signal of higher neurological risk in the very oldest vegans.” This one is currently unpublished and reported secondhand, which means it cannot carry weight in either direction yet, and it deserves flagging rather than burying, because pretending inconvenient signals do not exist is how you lose credibility when they publish. The candidate explanations are the same correctable ones as elsewhere: cumulative B12 exposure over decades in a cohort recruited before supplementation was routine, vitamin D, and omega-3 status. When it publishes, it should be read carefully rather than dismissed. The DHA post
Fallacy: unpublished signal treated as a verdict
Isn’t it true that vegan diets aren’t automatically healthy?
“You admit it needs planning, so it isn’t automatically optimal.” Correct, and no claim in this post says otherwise: chips and soda are technically plants. The concession does not travel where it is being taken, though, because every diet requires planning and only this one gets treated as disqualified for needing it. The relevant comparison is between a well-constructed plant-based diet and a well-constructed omnivorous one, and that is a comparison the plant-based pattern wins on cardiometabolic outcomes. A requirement to eat thoughtfully is not a defect; it is the baseline condition of eating well on any diet.
Fallacy: concession leveraged as refutation
Aren’t calcium supplements dangerous anyway?
“So vegans either break bones or take calcium pills that cause heart attacks.” The premise is weaker than it sounds. A 2023 meta-analysis of 11 placebo-controlled trials covering more than 55,000 people found calcium supplements of about 1 gram a day, alone or with vitamin D, were not significantly associated with heart attack, coronary heart disease, stroke or death, and excluded excess risks above 0.3 to 0.5 percent a year; its authors argue that earlier meta-analyses which raised the alarm were skewed by how they selected and weighted trials (Huo 2023). Whatever concern remains attaches to large supplemental doses, not to calcium from food or fortified foods, which arrive in modest amounts spread through the day: a cup of fortified soy milk carries around 300 mg against a 1,000 mg tablet. The practical answer closes the gap with food first: fortified plant milk, calcium-set tofu, fortified juice and low-oxalate greens, keeping any supplemental dose small and taken with a meal. The dilemma only exists if you assume a pill is the only route. The bone density post
Fallacy: false dilemma
Don’t you need K2 to direct calcium into bone?
“Without K2 the calcium goes into your arteries instead.” This claim is sold by companies selling K2, which is worth knowing before weighing it. The mechanism is real in outline (vitamin K activates proteins involved in bone and vascular calcium handling), and the outcome evidence is mixed rather than absent: bone-density and fracture benefits appear mainly in trials using pharmacological doses of one form of K2 rather than supplement-shelf doses, and calcification trials show a modest slowing. What no trial has tested is the claim actually printed on the label, that K2 makes calcium supplementation safe. Vitamin K1 is abundant in leafy greens, which plant-based eaters are not short of, and gut bacteria produce some K2. Here the marketing has outrun the data. The bone density post
Fallacy: supplement-marketing mechanism
Longevity, the Blue Zones and the Adventist cohorts
An organised campaign against the population evidence, and the strongest version of it is worth taking seriously: extreme-age records really are unreliable. What follows is where that critique lands and where it stops.
Have the Blue Zones been debunked?
“Pension fraud and missing birth certificates. Japan had 230,000 dead centenarians, Greece was collecting pensions from the grave, and the best predictor of being a supercentenarian is having no birth certificate.” Newman is right about the general principle and wrong about several specifics, and the difference matters. Japan’s 230,000 figure came from the family register, a genealogical document nobody claimed was a population count, and the in-person sweep that followed found zero missing centenarians in Okinawa. The Greek figure was a pension payment roll, while Ikaria’s longevity was validated separately. Costa Rican age exaggeration was published by the Blue Zones researchers themselves. Individual supercentenarian claims deserve deep suspicion; the diet-longevity signal is measured at the population level, and it does not depend on anyone reaching 110. Saul Newman versus the Blue Zones
Fallacy: valid critique of records extended to the dietary finding
Was Loma Linda invented to give America a Blue Zone?
“An editor wanted an American zone, and CDC data show Loma Linda is average.” This attacks the geographic packaging and leaves the actual evidence untouched, which is the reverse of what the critic needs. Loma Linda’s value was never the zip code. The research underneath is a cohort of Seventh-day Adventists tracked as people across California, not as a town: Fraser and Shavlik followed 34,192 of them from 1976 to 1988 and, comparing 1980 to 1988 follow-up with 1985 California data, found life expectancy at age 30 was 7.28 years higher in men and 4.42 years higher in women than for other white Californians. That is a comparison with outsiders, so it reflects the whole Adventist lifestyle package. The gradient inside the cohort is the sharper evidence: commonly observed combinations of diet, exercise, body mass index, past smoking and hormone therapy in women accounted for differences of up to 10 years among Adventists, and being vegetarian rather than non-vegetarian was worth 1.53 years in men (0.41 to 2.65) and 1.51 in women (0.53 to 2.50) with the other factors held at medium risk. The often-quoted 9.5 years for vegetarian men is a univariate comparison with the 1985 California population, and the authors note that vegetarians usually sat further toward the healthful end on nuts, exercise, past smoking and body weight, so that figure measures a package of habits, not diet alone. The design is observational, not a trial, and the authors say psychosocial and religious factors may also be important. Dismissing the zip code does nothing to the cohort. Blue Zones: debunked or not? · What the Adventist Health Studies actually tell us
Fallacy: ecological (zip-code) fallacy
Isn’t Blue Zone longevity about lifestyle rather than plants?
“There’s a lot more to it than eating plants: community, movement, purpose.” Entirely true, and this post does not need the Blue Zones to make its case, which is worth stating plainly. Movement, social connection and low smoking rates are genuine contributors and nobody serious claims otherwise. The reason diet is separable is that it is the one variable where cohort gradients, randomised feeding trials, identical-twin studies, Mendelian randomisation and mechanism all converge on the same answer independently of any village in Sardinia. Confounding is a reason to climb to better evidence, not a reason to stop. What the longest-living people eat
Fallacy: confounding treated as refutation
Do pescatarians outlive vegans?
“The Adventist data put fish-eaters ahead of vegans. Checkmate.” The hazard ratios in question are close together with overlapping confidence intervals, which means the honest reading is that both groups did better than meat-eaters and the gap between them is not resolvable. The more interesting point is structural: veganism is an ethical position, so its adherents include people with no interest in nutrition at all, while pescatarianism is adopted almost entirely for health reasons by people already invested in it. Any residual healthy-user bias in these cohorts therefore flatters the fish-eaters, not the vegans. The evidence review
Fallacy: overreading a non-significant difference
Aren’t Adventists too healthy a group to learn anything from?
“The whole Adventist package does the work. It’s like studying the healthiest heroin users.” The headline comparison with other Californians does invite a healthy-user worry: the authors dropped the early follow-up years to guard against a healthy-volunteer effect and the estimates barely moved, but healthy habits are exactly what they say accounts for the gap. The better evidence is inside the cohort, which they chose to study because Adventists span a wide range of diets and other habits, with virtually no current smokers. It does not remove confounding: vegetarians also usually sat further toward the healthful end on nuts, exercise, past smoking and body weight, and the authors allow that their estimates may be modestly distorted by unknown confounding. That is why the cohort’s own estimate of the vegetarian effect, with the other factors held at medium risk, is the number to use (1.53 years in men, 1.51 in women) rather than the 9.5-year headline, and it remains observational, not a trial. The evidence review · What the Adventist Health Studies actually tell us
Fallacy: confounding misapplied within a cohort
Aren’t the vegan advantages just weight loss?
“Adjust for BMI and the edge disappears.” Lower body weight is one of the mechanisms by which the diet works, so adjusting it away is adjusting away part of the treatment effect, a well-known error when a covariate sits on the causal pathway rather than beside it. Where trials control the comparison tightly, the advantages appear anyway: identical twins randomised to a vegan diet showed lower LDL and fasting insulin than their siblings within weeks. If the answer to “why is this diet better” is partly “because people end up carrying less body fat on it,” that is a benefit, not an artifact. What the longest-living people eat
Fallacy: adjusting away a mediator
Don’t vegan advocates just cherry-pick their confounders?
“B12 gaps, junk food vegans, late converts: those excuses apply to everyone.” Half right, and the half that is right should be conceded. Where it fails is on misclassification: when dietary categories are assigned imperfectly, the error is generally non-differential, which pushes estimates toward the null rather than inventing an advantage. So unreliable vegan classification means the measured benefit is more likely understated than overstated. The B12 argument is also not symmetrical, because it names a specific, measurable, correctable deficiency concentrated in one group rather than a general appeal to unmeasured virtue. The evidence review
Fallacy: misdirected bias argument
Does the vegan advantage disappear outside the Adventists?
“EPIC-Oxford found no all-cause mortality benefit.” It did not, and the reason is instructive rather than damaging: the comparison group in EPIC-Oxford was health-conscious in its own right, recruited heavily through health food shops and vegetarian networks, so the meat-eaters in that cohort eat far less meat than a typical population. When your control group is already doing much of the intervention, differences shrink. Cause-specific results in the same cohort still showed substantially lower ischaemic heart disease in vegetarians. All-cause mortality is also the least sensitive endpoint in nutrition, because it includes every way of dying. The evidence review
Fallacy: selective cohort emphasis
What about long-lived meat-eaters and vegans who died young?
“Jeanne Calment ate chocolate and lived to 122. The founder of American Veganism died at 66.” Both are true and neither is evidence, which is the entire point of using populations rather than people. Individual lifespan is dominated by genetics, accident and luck, which is why a chain-smoking 95-year-old does not refute oncology. If anecdotes settled dietary questions, the collection would be trivially easy to assemble on both sides, which is exactly why nobody serious builds an argument from them.
Fallacy: anecdote
Doesn’t Hong Kong eat the most meat and live the longest?
“Longest life expectancy in the world, enormous pork consumption. Checkmate, plant people.” Life expectancy at birth is a whole-society statistic shaped by healthcare, wealth, road safety and, in Hong Kong’s case, plausibly above all smoking. When researchers analysed 263 million deaths to find out why Hong Kong tops the table, their key explanation was economic prosperity combined with successful tobacco control: reduced smoking-attributable mortality contributed to 50.5 percent of Hong Kong’s survival advantage over men in high-income countries and 34.8 percent over women. The study did not assess diet, alcohol, physical activity or obesity, so it cannot speak to pork either way. The timeline cuts against the pork argument too: people aged 80 or older accounted for half of the female life expectancy gap with other rich countries in 2016, and the authors note that most older adults in Hong Kong grew up under limited living conditions. Hong Kong’s own food survey finds adults eat about 110 grams of meat, poultry and game a day, before counting roasted and marinated items. The 664 grams (about 1.5 pounds) that circulates online comes from a 2018 University of Hong Kong emissions study and is a trade-based apparent-consumption estimate, not intake: imports plus local livestock supply, minus exports and re-exports, with live animals converted at average whole-animal weights and wastage counted as consumption. Colorectal cancer is not a clean test either: its age-adjusted rate has shown no significant trend in 41 years and has fallen since about 2010, and the rising case count mostly reflects an ageing population. A 2026 analysis does find a small net rise in men over four decades (0.5 percent a year), with elevated risk in men born after 1950, while risk in women born after 1960 is falling. Its authors list lifestyle factors including red and processed meat as possible contextual explanations rather than causal ones, so the data cannot isolate meat. And Woo and Marmot note that the rise in life expectancy has not been matched by the same rise in health span, leaving a growing dependency burden. Hong Kong, meat and longevity, in full · What the longest-living people eat
Fallacy: ecological fallacy, twice over
Isn’t longevity mostly genetic?
“Some people are just built to live long.” Twin and pedigree studies put the heritability of lifespan at roughly a fifth or less, though a 2025 analysis argues the figure rises substantially once accidental and infectious deaths are modelled out, a claim other researchers immediately disputed. Migration studies settle it most cleanly: populations who move and adopt a new diet acquire the disease profile of their destination within a generation or two, which genetics cannot explain. Genes set a range; behaviour determines where in it you land.
Fallacy: single-cause displacement
Isn’t a little meat fine?
“The data support a small amount, not elimination.” This is the most reasonable objection in the section, and it is a statement about where the data stop rather than where the benefit stops; most cohorts have too few zero-meat participants to resolve the last increment. Where gradients are measured, they improve at every step: less red meat to none, meat to fish, fish to dairy and eggs, mostly-plants to all-plants. When a curve improves monotonically at every measured step, the burden falls on whoever claims it reverses at the final one. And the health argument going quiet at the margin is not the only argument operating there. What the longest-living people eat
Fallacy: argument from where the data stop
Do studies show vegetarians are less healthy?
“The Austrian study found vegetarians reported more cancer, more anxiety and worse health.” That study is cross-sectional, meaning it photographed diet and health at the same moment and cannot tell you which came first, and people very often adopt a vegetarian diet after a diagnosis. It was also a self-report survey with a small vegetarian sample and no adjustment capable of removing that reversal. On the evidence ladder it sits several rungs below the cohort and trial data it is deployed against, which settles it without needing to attack the authors. What the longest-living people eat
Fallacy: cross-sectional design · reverse causation
Has any society ever lived to a hundred on plants?
“No population has ever reached extreme old age eating this way.” The goalposts have moved a long way here, because no population has ever eaten a modern well-planned vegan diet either: the supplement is eighty years old and fortification younger. What does exist is the California Adventist cohort, where being vegetarian rather than non-vegetarian (few of the vegetarians were vegan) was worth about 1.5 years of life expectancy with the other risk factors held at medium levels, within a single religious community. Traditional Okinawans, whose diet ran well over ninety percent plants, held the world’s highest centenarian concentration and the longest disability-free life expectancy for decades. Demanding a historical all-plant civilisation is demanding evidence the modern era has only just started generating.
Fallacy: moving the goalposts
Isn’t the vegan twin study compromised?
“The authors work for the company selling the epigenetic test, and the vegan twins lost more weight anyway.” Both caveats are legitimate and belong in any honest account of the follow-up epigenetic analysis, since surrogate markers with a commercial interest attached are not disease outcomes. They do not touch the primary trial, which measured LDL, fasting insulin and weight with genetics perfectly controlled, and found the vegan twins ahead on all three. Attack the aging-clock paper if you like; the core result stands on ordinary clinical measurements. What the longest-living people eat
Fallacy: genetic fallacy
“The science can’t be trusted”
This is where the argument retreats when the evidence goes against it, and it is almost pure cliché: a set of phrases that excuse the speaker from engaging with any finding, applied with great selectivity to one side of the table only.
Is nutrition science real science?
“There are no controls in nutrition studies. The results don’t mean anything either way.” Nutrition research has genuine limitations, and treating those limitations as a universal solvent is not skepticism but an exit. The field runs metabolic ward studies with every bite controlled, randomised feeding trials, drug trials that isolate single mechanisms, and Mendelian randomisation that uses random allocation of genes at conception to approximate a lifelong trial. The test of real skepticism is consistency: the same people who declare cohort data worthless will cite a cohort the moment it flatters them, and will accept a self-selected online survey of carnivore dieters without a murmur. A claim you can trust is one that survives at every level of evidence, and that standard is available to anyone willing to apply it in both directions. The evidence review, and the carnivore literature examined in full. The processed meat post
Fallacy: selective skepticism · isolated demand for rigour
Isn’t the evidence for vegan diets just confounded cohort studies?
“It’s all observational: healthy-user bias, food questionnaires, and trials with small, inconsistent effects.” Grant the fair part first. Cohorts are confounded, questionnaires are noisy, and for vegan-specific mortality, most individual cancers and fracture risk the evidence is either unresolved or, for fractures, unfavourable. But the central claims do not rest on cohorts. A 2022 meta-analysis of 11 randomised trials in 796 people with overweight or type 2 diabetes, each lasting at least 12 weeks, found vegan diets lowered body weight by 4.1 kg, HbA1c by 0.18 percentage points and LDL cholesterol by 0.24 mmol/L compared with control diets (Termannsen 2022). A larger 2023 meta-analysis of 30 randomised trials with 2,372 participants found vegetarian or vegan diets lowered LDL cholesterol by 0.30 mmol/L (about 10 percent) and ApoB by about 13 mg/dL (14 percent) against omnivorous diets, with no conflicts of interest declared, though heterogeneity was substantial and the ApoB result rests on fewer trials (Koch 2023). The critique has a real point there, and the same paper makes it: the weight difference was 7.4 kg against people’s habitual diets but 2.7 kg against other therapeutic diets, seven of the 11 trials came from one research group, and the authors note that differences in energy intake may explain part of the effect. On causation, a European Atherosclerosis Society consensus drawing on more than 200 studies and over 2 million people, combining Mendelian randomisation with more than 30 drug trials, concluded that LDL causes atherosclerotic cardiovascular disease, with benefit proportional to how far and for how long it is lowered (Ference 2017). Its authors disclose extensive ties to makers of cholesterol drugs, which is one reason the genetic evidence, where inheritance rather than prescription does the allocating, carries so much weight. Add the identical-twin trial, which removed genetics from the comparison, and healthy-user bias explains none of the three. A strict standard for evidence is welcome; applied evenly, it lands hardest on a carnivore literature with no randomised trials at all. The evidence review · The processed meat post
Fallacy: attacking the weakest tier of evidence · isolated demand for rigour
Don’t conflicting anecdotes cancel each other out?
“You’ll find anecdotes both ways, so they prove nothing,” said while citing the Inuit and the Maasai. The first clause is correct and the sentence refutes itself before it finishes. Populations offered as proof are anecdotes with a larger sample size and no control group, which is precisely why the field built cohorts, trials and genetic designs on top of them. If anecdote is worthless, the ethnographic exhibits go first. You cannot spend an argument’s evidence and its dismissal of evidence in the same breath.
Fallacy: self-refutation · special pleading
Isn’t the plant-based case just a narrative?
“You follow a narrative. Period.” This is an assertion wearing the clothes of an argument, and it is unanswerable by design, which is the point of saying it. Anything can be called a narrative; what distinguishes one from a finding is whether it makes predictions that could fail and then survives the test. Lower apoB predicts less atherosclerosis in genetic studies, in drug trials, in feeding studies and in populations, and any of those could have come out the other way. Name a specific claim and we can examine it; naming the category of “narrative” examines nothing.
Fallacy: poisoning the well
Isn’t nutrition research corrupted by money and agenda?
“Who’s pushing the vegan agenda? Fake soy, 3D-printed meat. Connect the dots.” Follow the money honestly and it runs the other way at every scale. The United States runs mandatory checkoff programmes that tax producers and spend the money on promotion and research: the beef checkoff alone collects a dollar per head of cattle sold, and USDA’s own report to Congress puts mandatory dairy and fluid milk assessments at $431.8 million in a single year, 2022. There is no federal checkoff at all for beans, lentils, leafy greens or whole grains. The largest crop checkoff, soybeans, brings in far more, but the board’s own materials describe animal agriculture as its number one customer for soybean meal, so even the biggest “plant” checkoff is substantially selling feed. Crop subsidies run the same way: nearly 90 percent of one recent year’s commodity payments and crop insurance subsidies went to five crops, led by corn at 38 percent, while specialty crops such as fruits, vegetables and nuts receive little or no comparable direct support. About 40 percent of that corn becomes livestock feed and nearly 45 percent becomes ethanol, not food for people either way. Compare all of that to the market size of the products this conspiracy supposedly serves.
It goes further than lobbying budgets. The Scientific Foundation for the Dietary Guidelines for Americans, 2025 to 2030, the review that replaced the government’s own advisory committee report, includes a scorecard of that committee’s recommendations, and marks its core advice, plant-first protein ordering, reducing red and processed meat, beans and legumes first, low-fat dairy, as “not implemented.” The replacement review’s own disclosure pages show why: at least four of its nine authors report National Cattlemen’s Beef Association, National Pork Board or dairy industry funding, a fifth holds an equity stake in a low-carbohydrate diet company, and a sixth discloses research support from the Atkins Foundation. A separate lawsuit filed by the Physicians Committee for Responsible Medicine in August 2026 alleges a higher number of conflicted panelists and asks a federal court to void the guidelines; that is an unadjudicated allegation from an interested party, not an established finding, and it deserves the same caveat this post applies to every other unproven claim. What is not disputed is what the government’s own documents already show.
The clearest single case study is the protein target itself. A 2015 review by Heather Leidy, funded by the Beef Checkoff and five other animal-industry sponsors, recommended 1.2 to 1.6 grams of protein per kilogram specifically as a short-term weight-loss strategy. That narrow, industry-funded number became the guidelines’ universal recommendation for every American a decade later. When a 2026 Journal of Nutrition critique examined the evidence behind it, the authors found 30 trials covering 2,042 people, every one of them already overweight or obese, measuring nothing but body weight and composition, with more than 80 percent of the added protein in those trials coming from animal foods; their conclusion was that weight-loss evidence cannot support a universal recommendation. Three of that critique’s six authors disclose their own National Cattlemen’s Beef Association ties. The objection to the industry-shaped number is not coming from vegans.
The version of this argument worth taking seriously is that funding influences research, which is true, and which is why we weight convergence across independent lines rather than trusting any single study. The version that treats every inconvenient finding as bought is unfalsifiable, and unfalsifiable is another word for useless. The full funding trail, with every figure sourced · The protein target’s origin, in full
Fallacy: genetic fallacy · unfalsifiable conspiracy · card stacking
Isn’t the online pushback against plant-based eating just regular people?
“You act like everyone who disagrees is a paid shill. We’re just regular people who tried this and it worked.” Mostly, yes, and that is worth saying plainly before anything else: the overwhelming majority of people arguing in a comment section are real, unpaid, and believe exactly what they are typing. But “mostly organic” and “documented organized effort exists” are not contradictory, and it is worth knowing what is actually on the record. The beef checkoff funds an online course called Masters of Beef Advocacy that has graduated more than 21,000 people, earmarked $572,700 for the programme in a single year, and distributes talking points to a private Facebook group of graduates; the checkoff separately describes its own Digital Command Center as monitoring more than 200 beef-related topics around the clock. A peer-reviewed analysis of 8.5 million tweets published in The Lancet found the #yes2meat countermovement against the 2019 EAT-Lancet report began organizing before the report even launched, driven by real accounts rather than bots. A 2025 leaked evaluation reported by DeSmog and The Guardian, from a PR firm called Red Flag, described a campaign to frame that same report as radical and hypocritical; the document does not name its client but shows indications of work for the Animal Agriculture Alliance. The advocacy group Changing Markets separately claims Ken Berry was among the first to post that hashtag, a claim he and others named reject, so treat that one attribution as disputed rather than established. Plant-based meat got the same treatment in 2020, when the Center for Consumer Freedom ran a Super Bowl ad calling an ingredient in plant-based burgers “a chemical laxative.” None of this proves any individual commenter you are talking to is paid, and an anonymous, unverified Reddit post claiming exactly that should be treated as exactly that: unverified. The full documented record
Fallacy: false dilemma (assuming genuine sentiment and organized amplification are mutually exclusive)
Didn’t Ancel Keys cherry-pick his data and the AHA get bought?
“Seven countries out of twenty-two, and Crisco money built the American Heart Association.” The Seven Countries Study selected nations with usable mortality data and dietary measurement, not nations with convenient results, and the twenty-two-country accusation comes from a 1957 critique built on cruder national data than the study it attacked. Whatever you conclude about Keys personally, the diet-heart hypothesis no longer rests on him in any way: it rests on Mendelian randomisation across millions of people, on four independent drug mechanisms that lower LDL and reduce events, and on controlled feeding studies. The AHA’s early funding history is real and equally irrelevant to whether ApoB causes plaque. An argument about who funded a body in 1948 is not an argument about what genetics showed in 2017. Saturated fat, seed oils and heart disease
Fallacy: genetic fallacy
Didn’t the dietary guidelines cause the obesity epidemic?
“They told everyone to eat low fat and look what happened.” The timeline runs backwards before anything else does. High-fructose corn syrup entered the food supply around 1970 and its use climbed steeply through the following two decades, meaning ultra-processed food arrived alongside and before the plant-forward advice, not decades after it; the federal Dietary Goals appeared only in 1977 and the food pyramid not until 1992. Adherence data undercut the rest of the premise. Americans never followed the guidelines: fat fell as a share of calories while total calorie intake climbed, so fat intake in grams did not actually drop, and per-capita meat and cheese consumption rose rather than fell over the same period. What the low-fat era actually produced was a food industry replacing fat with refined starch and sugar, which is a failure of implementation the guidelines did not recommend. It is also strange to cite the guidelines as discredited while celebrating the 2025 edition for floating butter and tallow. Either the process is authoritative or it is not. Cholesterol, from the ground up · The full timeline
Fallacy: post hoc · inconsistent standard
Don’t reversals and retractions show the field is unreliable?
“NutriRECS said keep eating red meat, and PREDIMED was retracted.” The NutriRECS panel’s own meta-analyses found the same risk reductions everyone else finds for eating less meat; what differed was how they graded certainty, and re-grading identical evidence is not refutation. PREDIMED was retracted and republished after a randomisation error was corrected, with the main conclusions intact, which is the system working in public, not failing. Retraction is the immune response, and pointing at it as evidence of sickness gets the diagnosis backwards. What the longest-living people eat
Fallacy: false balance · self-correction read as failure
Didn’t the Women’s Health Initiative disprove low-fat plant-based eating?
“Fifty thousand women, no reduction in heart disease, cancer or death.” The WHI tested a generically low-fat diet, not a whole-food plant-based one. The intervention aimed for 20 percent of energy from fat alongside more vegetables, fruit and grains, and gave no formal instruction on saturated fat or other specific fats. By year six the intervention group’s fat intake was only 8.2 percentage points of energy lower than the comparison group’s, and its saturated fat only 2.9 points lower; participants fell short of the fat goal, at 24 percent of energy after the first year and 29 percent at year six, against 35 and 37 percent in the comparison group. The trial found no significant effect on coronary events or stroke, and, as the American Heart Association’s 2017 presidential advisory notes, it was not a test of cutting saturated fat alone. It is a demonstration of the principle that what replaces the fat matters: in the Nurses’ Health and Health Professionals cohorts (Li 2015), replacing saturated fat with polyunsaturated fat, monounsaturated fat or whole-grain carbohydrate went with 25, 15 and 9 percent lower coronary risk, while refined starch and added sugar showed no significant association. Citing the WHI against a diet it did not test is a straw man built out of a real trial. Saturated fat, seed oils and heart disease
Fallacy: straw man
Isn’t the saturated fat research inconclusive?
“Research on whether saturated fat is good or bad is not definitive.” This is the retreat position once specific claims fail, and it works by treating any live disagreement as total uncertainty. Some questions genuinely are open (full-fat dairy, eggs, the precise effect of individual fatty acids), and they are open in a narrow band around a settled core. What is not in dispute is that saturated fat raises LDL in controlled feeding studies, that LDL causes atherosclerosis, and that replacing saturated with polyunsaturated fat lowers events. Manufacturing doubt about the core by pointing at the margins is a technique with a well-documented history in other industries. Saturated fat, seed oils and heart disease
Fallacy: argument from ignorance · manufactured doubt
Do you have to read a carnivore author’s book before you can disagree?
“You’d understand if you’d actually read her book.” A single self-published trade book by an author without a research record does not outrank systematic reviews, and the demand that critics read it before responding is a fee charged for the right to disagree. The claims in such books are the same claims answered above, which is why they can be addressed without buying it. Where a book cites primary research, we can check the primary research (that is the whole method), and where it does not, its being long is not evidence of its being right.
Fallacy: appeal to authority · shifting the burden of proof
Aren’t plant-based researchers biased too?
“Those studies were run by vegetarian advocacy groups, and vegans only say it’s healthy because they care about animals.” Ideological conflicts are real and should be disclosed: some studies finding no bone deficit in vegans do carry vegetarian-nutrition affiliations, and that belongs on the record. Two things follow. First, the conflicts that have been measured most carefully are financial, and they have a documented direction. A 2025 systematic review of red meat trials in The American Journal of Clinical Nutrition (López-Moreno and colleagues) found that every industry-linked trial reported neutral (79.3 percent) or favorable (20.7 percent) cardiovascular results and none reported unfavorable ones, while 73.3 percent of independent trials reported unfavorable results, 26.7 percent neutral and none favorable (odds ratio 3.75 for a favorable or neutral conclusion, 95 percent confidence interval 1.62 to 8.67). A 2026 review of 500 meat-and-health studies found industry ties in 78 of them and an odds ratio of 16.4 (95 percent confidence interval 7.5 to 35.8) for favorable conclusions (Teimouri and colleagues). Neither review can show intent: the first review’s authors conclude that animal-protein comparators in industry-funded studies may understate the benefit of eating less red meat, and the accompanying editorial notes that the finding does not necessarily mean researchers deliberately biased their trials. Earlier work on soft drinks, juice and milk found the same pattern (Lesser 2007). Second, motive is not evidence either way: someone can arrive at a position for ethical reasons and be correct about the data, and the way to establish that is to check the data. That is what this entire post is doing.
Fallacy: genetic fallacy · ad hominem circumstantial
Testimonials and what stands in for evidence
Having declared the scientific literature unusable, the argument needs something in its place. What it reaches for is the weakest rung on the ladder it just declared worthless.
Doesn’t “I feel great” count for something?
“Every person I’ve seen on the lion diet got healthier. I’ve never felt better.” It counts for something real: feeling well matters, and short-term improvement on an elimination diet is common and genuine, usually because ultra-processed food, alcohol and erratic eating went out with everything else. What it cannot do is tell you about your arteries, because atherosclerosis is silent for decades and the first symptom is frequently the last one. This is the exact failure mode the diet is most vulnerable to: energy and glucose numbers that look excellent alongside an apoB that is quietly doing damage. Every diet movement in history has had people who felt amazing on it. Why “it worked for me” is not evidence
Fallacy: anecdote · selection bias
What about coaches who have reversed disease in thousands of clients?
“I’ve reversed cardiovascular disease in nearly two thousand carnivore clients.” There is no control group, no independent verification, no accounting for anyone who quit or got worse, and a financial interest in the result, which is four separate reasons a number like this carries no evidential weight regardless of who says it. “Reversed” is also undefined: improved symptoms, better glucose, a scan, weight loss? A specific, true-sounding number like “nearly two thousand” does real rhetorical work here: it borrows the credibility of a statistic for a claim that has none of a statistic’s actual backing. These claims arrive attached to a funnel, usually ending in a comment prompt and a paid programme. A clinician with genuine results of that magnitude has an obvious route available, which is publishing them.
Fallacy: uncontrolled testimonial · conflict of interest · card stacking
Doesn’t a good calcium score prove the diet works?
“Calcium score of 5 at seventy, after five months carnivore.” A coronary calcium score measures accumulated calcified plaque from decades of living, so a score at seventy reports on a lifetime, not on five months, and crediting the most recent months for it is reading a total as a rate. Calcium scores also miss soft plaque, which is the type most likely to rupture, and scores frequently rise on intensive treatment as existing plaque stabilises and calcifies. A single number from a single person with no baseline is not a result. It is a starting point someone mistook for a finding. How plaque forms
Fallacy: misreading a cumulative measure
Does carnivore cure autoimmune disease, cancer and mental illness?
“It reversed MS, it shrank a tumour, low carb cures almost all known mental illness.” Autoimmune conditions relapse and remit spontaneously, which is precisely why uncontrolled case reports are so unreliable in this area and why the field requires controlled trials. Elimination diets do help some people with specific conditions, and that is worth taking seriously without inflating it into a cure for everything. The cancer claims are the most serious, because they are aimed at people making decisions under fear: there is no trial evidence that a carnivore or ketogenic diet treats cancer, and a single high-profile recovery in someone receiving conventional treatment is not evidence that the diet did it. Claiming near-universal efficacy for one intervention is the oldest signature there is. Cancer and diet, in full
Fallacy: case report treated as proof · spontaneous remission ignored
Didn’t someone put stage 4 cancer into remission with diet?
“My friend beat stage 4 cancer on carnivore.” Nobody should sneer at the person in that story; they are alive and grateful, and the problem is the inference rather than them. Ask what else was happening: if chemotherapy, surgery, radiation, hormone therapy or immunotherapy was running alongside, the diet is collecting credit the treatment earned. The people whose cancer progressed on the same protocol are not posting updates, staging in a testimonial is self-reported unless records are shown, and feeling better is a different claim from a clear scan. Every protocol also has its own remission stories, Gerson, alkaline and raw vegan among them, and since they contradict one another they cannot all be the cause. Where the choice has been measured, it is not harmless. Among 281 patients with curable, non-metastatic breast, prostate, lung or colorectal cancer who chose alternative medicine instead of conventional treatment, the risk of death was 2.5 times that of matched patients who received standard care, more than fivefold for breast cancer, though not significantly different for prostate cancer, which has a long natural history (Johnson 2018). Cancer and diet, in full
Fallacy: anecdote · survivorship bias · post hoc
What about the eighty-seven-year-old who only eats beef?
“Look how she moves. Look at her hair.” For every long-lived carnivore there is a long-lived vegan, a long-lived Mediterranean farmer and a long-lived smoker, which is why individuals cannot settle population questions in either direction. Survivorship bias does the rest of the work: the people who did badly on an extreme diet are not making videos about it. And “look how she moves” is not a measurement of anything; plenty of eighty-seven-year-olds move well on every diet humans eat.
Fallacy: anecdote · survivorship bias
Doesn’t the Harvard carnivore survey show the diet is safe?
“Two thousand carnivore dieters reported excellent health.” That study is a self-selected online survey recruited through carnivore communities, which means it sampled the people still doing it and enjoying it, and anyone who felt terrible and quit is structurally absent from the data. It collected self-reported outcomes with no control group, no verified laboratory values for most participants, and no follow-up. The true, precise-sounding numbers, “2,029 people,” “95 percent,” do exactly what a stacked statistic is supposed to do: they make a self-selected online questionnaire feel like a clinical result. The authors themselves framed it as hypothesis-generating, and among the 462 of 2,029 respondents who reported a current LDL, the median was 172 mg/dL, up from 126 before the diet (326 reported a pre-diet value and 247 reported both), the one lab value in the study and the wrong direction. It is the weakest study design in circulation, deployed by people who declare prospective cohorts of half a million unusable. The Lennerz survey, taken apart
Fallacy: self-selected survey treated as trial evidence · card stacking
What about all the doctors who recommend this?
“Chaffee, Berry, Baker, Saladino, Bikman: they’re qualified and they all say it.” Credentials establish that someone could evaluate evidence, not that they have, and a dentist, an orthopaedic surgeon and a family physician are not epidemiologists or lipidologists. The more telling detail is what has happened to the movement’s own founders: Saladino publicly quit strict carnivore after about two years, citing a drop in testosterone, heart palpitations, muscle cramps and sleep problems, and reintroduced fruit, honey and raw dairy; Mikhaila Peterson, who built the Lion Diet, announced after eight years that she is no longer carnivore, though she still eats around 95 percent meat. Meanwhile every national health body and cardiology society on earth holds the opposite position. A handful of media-facing clinicians against the assembled professional bodies of every developed country is not a balanced debate, and presenting it as one is the false-balance move in its purest form.
Fallacy: appeal to authority · false balance
Doesn’t it matter that even plant-based doctors have quit?
“Even the vegan cardiologist went back. That tells you everything.” It tells you about one clinician’s path, and conversion narratives are compelling for reasons that have nothing to do with evidence: the drama of reversal reads as credibility. People leave every diet, every profession and every position, in both directions, and there are more clinicians moving toward plant-based practice than away from it. Look at what the specific claims actually were when they are examined individually: a bonking paper, a dog with pancreatitis, an undefined clinical impression about patients above a certain BMI. This is also the place to say something the movement understands poorly: disengagement is a process, with identifiable stages, and it is triggered far more often by social friction, poor planning and exhaustion than by a diet failing physiologically. Treating each departure as a verdict on the evidence mistakes a psychological trajectory for a scientific finding.
The low-carb movement’s own flagship imaging study is worth knowing in full rather than in summary. The paper (Soto-Mota and colleagues, JACC: Advances, published in spring 2025) followed 100 people with very high LDL on long-term ketogenic diets, scanned one year apart, and was titled “Plaque Predicts Plaque, ApoB Does Not”; its conclusions describe the ApoB analysis as exploratory. Its pre-registered primary outcome was the change in noncalcified plaque volume, the soft, rupture-prone kind. A May 2025 letter to the journal said that outcome was shown only briefly in a figure, without proper description, while the authors emphasised a secondary endpoint, percent atheroma volume. The authors’ reply supplied the missing numbers, including the relative change (a median rise of 18.9 cubic millimetres, or 42.8 percent), apologised for the omission from the original publication, and called it “a sincere oversight, not intentional selective reporting.” They added that because most participants had low baseline plaque, relative changes are expected to be magnified. The version now hosted online (the July 2025 issue) reports the median change in its abstract and the 43 percent relative change in its results and figure legend.
A coauthor, James Earls, was chief medical officer of Cleerly, the company whose software analysed the scans. According to Retraction Watch, the journal says he disclosed that affiliation at submission and his equity in the company at acceptance; the published paper listed the affiliation, and its disclosure statement reported no relevant relationships for authors other than Norwitz and Feldman. Three of the authors told Retraction Watch they had not seen the data before publication, did not know Cleerly’s analysis was not double-blinded, and did not know Earls held that role when the data were collected; they say they alerted the journal immediately once they saw it. The journal retracted the paper on March 11, 2026, at the request of the authors and the Editors, because concerns raised after publication about the methodology affected the reliability of the data, and the authors and the Editors agreed the errors were too great for a corrigendum; the notice does not say who raised the concerns or what the errors were. Norwitz wrote that he was glad to retract because he considers the findings robust, while the critic Brad Stanfield called the reframing of the same data “more advocacy than science.”
In a January 2026 preprint, posted before the retraction and not yet peer reviewed, six of the authors reanalysed the scans. By the pre-specified method, QAngio, median noncalcified plaque volume still rose 5.6 cubic millimetres, a 37 percent relative increase over the one-year follow-up, which the authors again caution can be inflated by low baseline plaque; HeartFlow was run as a sensitivity analysis. The preprint concludes that ApoB and cumulative LDL exposure did not predict progression, and notes that the quantification methods were not trained to look for zero or little plaque and that there was no control group. The Reverse Transtheoretical Model, in Veganism Rewired · The vegan cardiologist video, answered · the KETO-CTA retraction, both sides, in full
Fallacy: anecdote · conversion-narrative authority
Is a vegan diet just a stepping stone to carnivore?
“Everyone ends up here eventually.” This is a story about where people go, told by someone who only meets the ones who came their way, and the traffic runs in both directions: plenty of people arrive at plant-based eating from keto and carnivore after their lipids or their gut told them to. The narrative works by treating any departure as confirmation and any arrival as noise, which is unfalsifiable by construction. What actually predicts whether someone stays is the same set of factors that predict staying with any change: preparation, social support, and whether the diet was built properly in the first place. The Reverse Transtheoretical Model, in Veganism Rewired
Fallacy: unfalsifiable narrative · selection bias
Won’t you never convince the millions who cured themselves, since experience always wins?
“Too many people have fixed their health this way for it to be wrong, and lived experience beats any study.” Start with the number, because it is never checked: no representative survey counts carnivore dieters, and the best anchor available is that around 2.2 percent of US adults were on a low-carbohydrate diet on a given day in 2017 to 2018 (CDC), of which carnivore is a fraction. The famous “2,029” is not a population, it is one study’s sample size, recruited from Facebook groups. The deeper problem is that “millions said it worked” has been true before and settled nothing. Estimates of the Atkins-era peak in 2003 and 2004 range from some 30 million Americans following Atkins (Knowledge at Wharton) to about 10 million on any low-carb diet at a time (NPD Group tracking). Then Atkins Nutritionals filed for bankruptcy in 2005, and NPD found that the share of Americans on a low-carb diet had fallen from 9.1 percent in February 2004 to 2.2 percent by mid-2005 (Washington Post, August 2, 2005). That 2.2 percent is NPD’s 2005 figure and is unrelated to the CDC’s 2.2 percent for 2017 to 2018 above. Some of those dieters no doubt lost weight and felt better for the same confounded reasons carnivore dieters do now, which is exactly why it was not evidence the diet was right. Experience tells you how you feel; it cannot tell you why, and it cannot see inside an artery. Atherosclerosis builds silently for decades and announces itself, often for the first time, as a heart attack. The whole argument, counted out with sources
Fallacy: appeal to popularity · survivorship bias
Isn’t the healthcare industry the real scam?
“$5 trillion in healthcare, 5 billion prescriptions, 400,000 deaths, and no effect on heart disease, cancer, diabetes or Alzheimer’s. That’s the real scam.” This script runs under nearly every video calling carnivore a scam, usually ending in a funnel to a free e-guide, and it is worth naming the moves rather than just the numbers. It opens with true, alarming facts, healthcare spending is real, then uses them to imply a conclusion, that carnivore therefore has evidence, that the facts never actually support: that is card stacking. It changes the subject from “where is the carnivore research” to “isn’t the medical system bad”: that is a red herring and a form of whataboutism. A healthcare system can be too expensive and an all-meat diet can still be unsupported; both can be true at once. It is also a stereotypical fad-diet tactic, reached for precisely because the diet being sold has no evidence of its own to offer instead.
The individual facts are mostly real, and worth checking rather than waving away. CMS puts US health spending at $5.3 trillion in 2024, about $15,474 per person, of which prescription drugs are roughly $467 billion, a fraction of the total rather than most of it. The “400,000 deaths” figure is not a count of prescription deaths at all; it traces to a 2013 estimate by John James of hospital deaths from preventable harm of every kind, with the study’s own authors giving a range of 210,000 to 440,000 and other researchers disputing the method. And medicine is not doing nothing: an analysis in The New England Journal of Medicine attributed about 47 percent of the decline in US coronary deaths between 1980 and 2000 to treatment, and cancer mortality has fallen 34 percent since 1991, averting an estimated 4.5 million deaths. Alzheimer’s drugs really have been disappointing, and that one is worth conceding plainly rather than spinning.
If the actual goal is getting people off medication through food, that goal has trials behind it, on the plant side specifically. In a randomized trial in type 2 diabetes, 43 percent of participants on a low-fat vegan diet reduced their diabetes medications, against 26 percent on the standard American Diabetes Association diet; Ornish’s program produced measurable regression on coronary angiograms. Carnivore has nine studies total, zero randomized trials, zero long-term cohorts, and zero outcome data of any kind. So the script defends the most invasive and violent industry on the planet, animal agriculture, for a diet with no evidence behind it, out of what amounts to an inherited ideology rather than an examined one. Where the carnivore evidence actually stands · Why testimonials aren’t evidence · Cholesterol, from the ground up
Fallacy: red herring · whataboutism · card stacking
Land, feed, efficiency and climate
The environmental argument, and the place where the best-informed objections live: several of these come from people who actually farm, and they deserve a serious answer rather than a slogan.
Isn’t 86 percent of livestock feed inedible to humans?
“Animals eat what we can’t. They don’t compete with us for food.” That figure is a tally by dry weight, which means a tonne of straw counts the same as a tonne of grain, and a weight-based accounting is structurally guaranteed to be dominated by bulky roughage. Change the unit to the currency that actually feeds people and the picture inverts: feed crops take 36 percent of global crop calories and 53 percent of global plant protein. The same paper found livestock consume about a third of global cereal production and require 2.8 kg of human-edible feed per kg of boneless ruminant meat. Even the “only” 14 percent that is human-edible works out to roughly 840 million tonnes a year, about the size of the entire global wheat harvest. The 86 percent statistic, taken apart
Fallacy: wrong unit of measurement
Isn’t grazing land too marginal to grow crops on?
“Tens of millions of acres are too wet, hilly or rocky for anything else. What would you grow on dry Montana rangeland?” This is true of some grazing land and it is the honest core of the farmer’s case. It is also not the question, because the argument has to account for the rest of the system: livestock occupy roughly 83 percent of the world’s farmland while supplying 18 percent of calories, and around 40 percent of the world’s arable land grows feed. The answer to “what would you grow on that rangeland” is, for a great deal of it, nothing. That is what rewilding means, and returning marginal land to native ecosystem is a use, not a waste. Marginal land is a defence of a fraction of the system offered as a defence of all of it. Rewilding versus regenerative grazing
Fallacy: exception generalised to the whole system
Don’t livestock just eat byproducts that would otherwise be wasted?
“Distillers grains, beet pulp, almond hulls, bakery waste: what would you do with it, landfill it?” Byproducts are genuinely fed to livestock and genuinely useful, and the accounting error is treating the byproduct as though it caused the animal rather than the other way round. Soy is the clearest case: the meal is around 80 percent of the crop’s mass and the majority of its value, so calling it a leftover from oil production reverses the economics. Demand for feed shapes what gets planted, how it is graded and where it goes. In a world with fewer livestock, far fewer of these residues would exist, and those that remained would go to anaerobic digestion, composting and soil. The 86 percent statistic, taken apart
Fallacy: reversed causation (value follows demand)
Isn’t feed grain unfit for human consumption anyway?
“It’s graded feed-quality. Mycotoxins. People can’t eat it.” Edibility here is a market classification, not a law of nature: the FAO analysis everyone quotes is careful to say “currently not eaten by humans,” which is a statement about economics. Grain is grown, harvested, stored and graded differently when the buyer is a feedlot, and much of what is graded as feed could have been grown and handled as food had the demand been there. Some genuinely is contaminated and unfit, which is a real category and a small one. The grading system is downstream of the industry, so citing it as a justification for the industry is circular. The 86 percent statistic, taken apart
Fallacy: circular reasoning
Can people eat what cattle eat on pasture?
“I’ve yet to see a vegan thrive on dry grass and clover hay.” Nobody has ever proposed this, which makes it a straw man rather than a rebuttal. The claim is not that humans should eat cattle feed; it is that the land, water and fertiliser currently producing cattle feed could produce human food directly, and the comparison has been run in exactly those terms. Two acres of good pasture raises roughly one calf a year, around half a million calories of beef; two acres of potatoes returns about thirty-two million. Even a rotation-friendly, nutritionally rounded mixed plot outproduces the steer roughly fortyfold on calories and more than tenfold on protein. Two acres: what beef produces versus what plants can
Fallacy: straw man
Don’t you need cows to farm crops?
“Without manure you’d poison every cultivated acre with synthetic fertiliser.” Most synthetic nitrogen already comes from the Haber-Bosch process rather than from animals, and a substantial share of it is applied to feed crops, so livestock are currently a major consumer of the fertiliser they are credited with replacing. Fertility without animals is not hypothetical: nitrogen-fixing legume rotations, green manures, cover cropping, compost and increasingly recovered human waste are all in commercial use. Manure also is not free fertility; it is recycled plant nutrients minus everything the animal burned, with substantial nitrogen lost as ammonia and nitrous oxide. The nutrients came from plants in the first place. Rewilding versus regenerative grazing
Fallacy: false necessity
Didn’t lab-grown meat and the alternatives fail?
“Cultivated meat flopped and the fake meat companies are collapsing.” Company valuations are a fact about markets, not about agronomy, and nothing in the land arithmetic depends on any of those companies succeeding. The plant-based case has never rested on novel products; it rests on beans, grains, potatoes and vegetables, which are the cheapest and most established foods in existence. If every cultivated meat startup vanished tomorrow, two acres of potatoes would still outproduce two acres of beef by the same margin.
Fallacy: non sequitur
Is methane from cattle just part of a natural cycle?
“Biogenic methane cycles out, and we’ve had ruminants in these numbers for fifty million years.” The historical premise is simply false: farmed animals now make up around 60 percent of mammal biomass on earth, humans most of the rest, and all wild mammals combined are down to roughly 4 percent. There was never anything like this standing stock of ruminants. The cycling argument has a legitimate kernel (methane’s atmospheric lifetime is about a decade, so a stable herd adds less additional warming than the standard metric implies), and it does not make a herd of this size neutral, because that methane is warming the planet right now while it persists. A shrinking herd would deliver rapid cooling, which is an argument for reduction, not for the status quo. Rewilding versus regenerative grazing
Fallacy: false historical premise · misapplied metric
Doesn’t regenerative grazing sequester carbon?
“Holistic grazing builds soil carbon and can go carbon-negative.” Soil carbon sequestration under well-managed grazing is real, and it is finite, reversible and small relative to the emissions of the animals doing it: soils saturate after a few decades, and the carbon leaves again if management changes or drought hits. The most-cited carbon-negative case studies rely on short measurement windows and specific starting conditions, usually degraded land where almost any improvement would register. Underneath it all sits the physics: roughly 90 percent of energy is lost at each step up a food chain, and beef returns about three percent of the calories fed to cattle. No management system repeals the second law of thermodynamics. Rewilding versus regenerative grazing
Fallacy: local effect scaled to a global claim
Isn’t crop agriculture destructive too, and isn’t food waste the real issue?
“Monoculture, tillage, soil loss, and a third of all food thrown away.” All true, and all fully compatible with everything above; nobody is defending industrial monoculture as an ideal. The relevant point is that most of that cropland is growing feed, so reducing animal agriculture is the single largest available reduction in crop agriculture’s footprint. Food waste is a genuine problem with its own solutions, and it is also the most calorie-intensive waste that occurs when a crop is routed through an animal first. Naming a second problem does not answer the first. Two acres: what beef produces versus what plants can
Fallacy: whataboutism
Don’t some species-rich grasslands depend on grazing?
“Chalk downland and species-rich meadows need livestock to survive.” This is correct, and it is one of the few genuinely good objections in the environmental section. Certain semi-natural habitats were created by centuries of grazing and lose their botanical diversity without disturbance. What it justifies is conservation grazing on a small, specific acreage, usually with hardy native breeds at low density: a rounding error against global livestock numbers, and not a defence of feedlots, soy-fed poultry or the beef industry at scale. Wild and reintroduced herbivores can perform the same function where they are permitted to. Rewilding versus regenerative grazing
Fallacy: exception generalised
The animal-harm counterattacks
The ethical counterattack: not a defence of eating animals but an attempt to show the vegan kills too, which would make the whole position hypocritical rather than wrong.
Don’t crop deaths mean vegans kill more animals?
“One grass-fed steer feeds you for months and dies once. A wheat field kills dozens.” The argument fails on the denominator: feed crops are crops, most of the world’s crop acreage grows animal feed, and it takes several kilograms of that feed to produce a kilogram of meat. Routing food through an animal multiplies the cropland your diet requires, and therefore multiplies the field deaths, before you add the slaughtered animal on top. Counted properly (field deaths plus slaughter deaths, per calorie delivered), an omnivorous diet costs roughly ten times the animal lives of a plant-based one, with chicken and eggs by far the most lethal per calorie because it takes so many small birds to fill a plate. The person making this argument is describing a cost that lands harder on their own plate. Crop deaths, counted honestly
Fallacy: wrong denominator
Don’t vegans kill billions of bees for almonds?
“You won’t eat honey but you’ll eat almonds trucked across the country on the backs of dying bees.” Migratory pollination is a real welfare problem and worth taking seriously rather than deflecting. It is also not specific to plant-based eating: almonds, apples, blueberries and dozens of other crops are pollinated this way for everyone, and alfalfa and clover grown for livestock feed are pollinated commercially too. The charge of hypocrisy misunderstands the position it is attacking, because veganism was defined from the beginning as reducing exploitation “as far as is possible and practicable,” never as a claim of zero harm. Pointing out that a vegan cannot achieve perfection is not an argument for abandoning the direction. The honey and bees post
Fallacy: tu quoque · perfectionism
Don’t almonds use enormous amounts of water?
“Almond milk is an environmental disaster.” Almonds are water-intensive per nut and that is worth knowing; the comparison that matters is against what they replace. Per litre of finished product, almond milk uses substantially less water than dairy milk, and dairy’s footprint includes the water for the feed crops, which is where most of it hides. Oat and soy milk are lower again if water is your primary concern. Selecting one metric, comparing a nut against a beverage, and never running the same calculation on the dairy aisle is how this argument survives. The honey and bees post
Fallacy: single-metric comparison with no baseline
What about fawns killed by farm machinery?
“Baby deer cut to pieces by combines.” This happens, it is horrible, and it happens overwhelmingly during forage mowing: hay and silage cut in late spring, timed exactly to when fawns are lying motionless in tall grass. That is livestock feed. Grain harvest occurs later in the year, when fawns are mobile. The image most often used to indict plant agriculture is a photograph from the animal agriculture supply chain. Crop deaths, counted honestly
Fallacy: misattributed cause
Aren’t plants sprayed with pesticides that poison wildlife?
“Pesticides kill every mouse, vole and fox that eats them, and your produce is soaked in them.” Both halves point back at the feed crop. US legal residue ceilings for glyphosate run up to around 4,000 times higher on animal feed commodities than on fruiting vegetables, and corn, soy, alfalfa and hay are where the herbicide tonnage is actually applied. Because it takes pounds of feed to make a pound of meat, eating animals multiplies the sprayed acreage your diet depends on rather than escaping it, and the persistent fraction concentrates in exactly the fat and organs this movement recommends eating. If pesticide exposure is your genuine concern, the plant-based eater is the one who has minimised it. The pesticide argument
Fallacy: inverted exposure
Don’t plants feel pain too?
“You’re killing plants, and they’re conscious too, hun.” Plants respond to damage with electrical and chemical signalling, and they have no nervous system, no nociceptors, no brain, and no evolutionary reason to evolve the capacity to suffer, since suffering exists to drive escape behaviour available only to things that can move. Even granting the premise entirely, the argument defeats itself: livestock eat many times more plant matter than humans do, so if plant suffering counts, eating animals causes vastly more of it. The tell is that nobody raising this point acts on it anywhere else in their life. It is not a claim about plants; it is a way of ending the conversation. Plant toxins and antinutrients
Fallacy: false equivalence · tu quoque
“I tried it and it didn’t work”
Everything above this line is a cliché. Everything in this section is a functional thought: a real cost, a real barrier, a real experience, and it deserves an entirely different posture. These are not arguments to be defeated; they are problems to be solved, and almost all of them have the same small set of solutions.
The physical complaints: gas, bloating, fatigue, food that tastes like cardboard, eating enormous volume, cravings, iron or B12 coming back off. Nearly all of this is the transition rather than the destination. A gut microbiome accustomed to little fibre responds to a sudden large increase exactly as described, and the population of bacteria that ferments it quietly shifts over two to six weeks, which is why ramping up gradually, cooking legumes thoroughly, rinsing canned beans and starting with smaller portions resolves most of it. Fatigue and volume problems are usually a calorie deficit wearing a costume: plant foods are less energy-dense, so the same plate holds fewer calories, and the fix is nuts, seeds, tahini, avocado, oils, dried fruit and legumes rather than more salad. Blandness is a seasoning problem and a technique problem, not a property of plants: the world’s great cuisines were built on them. Cravings fade on a timescale of weeks. Iron and B12 are measurable, correctable, and worth testing rather than guessing: supplement B12 without exception, take iron with vitamin C and away from tea or coffee, and get bloodwork if something feels off.
“I didn’t know what to eat.” This is the most common real reason a transition fails, and it is a planning problem rather than a dietary one. Most people who succeed build a rotation of eight or ten meals they can cook without thinking, and add from there. Doing this without any meal planning, without a stocked cupboard, and without a dietitian or lifestyle-medicine practitioner when the bloodwork gets complicated is attempting the hardest possible version of the change and then concluding the change is impossible.
“It’s expensive and takes forever.” Dried legumes, rice, oats, frozen vegetables, potatoes and seasonal produce are among the cheapest foods available anywhere, and a secondary analysis of a randomised trial found a low-fat vegan diet actually reduced food costs. The perception comes from comparing mock meats and specialty products at premium retailers against the cheapest possible animal foods. Time is real in the first few weeks and then falls sharply, because batch-cooked legumes and grains keep for days.
“84 percent of vegans go back.” That figure counts everyone who ever tried, including people who lasted a fortnight, and it is a statistic about how people are attempting the change rather than about whether the diet works. The strongest predictors of staying are not physiological: they are social support, having a repertoire of food you actually enjoy, ethical rather than purely health motivation, and not attempting it alone. Disengagement runs through identifiable stages, and it can be interrupted at each of them, which is why the useful response to someone struggling is a recipe and an ally rather than an argument. The Reverse Transtheoretical Model, in Veganism Rewired
“My family, my culture, the food I grew up with, and I don’t want to be the difficult one at dinner.” This is the most honest objection in the entire post and the one no amount of data addresses, because it is not a claim about health at all. Almost every cuisine on earth has a plant-based backbone underneath the meat that was historically a garnish, and the answer is usually to cook the food you grew up with rather than abandon it. You are not obliged to become an activist, to argue at the table, or to announce anything. Quietly eating differently is a complete version of this.
“A preachy vegan put me off.” Fair, and worth saying plainly: how a message is delivered affects whether it lands, and that is on advocates rather than on you. It is also worth noticing what the sentence does if you follow it: it makes your diet contingent on someone else’s manners.
“We’re all unique, and my gut genuinely can’t handle it.” Genuine individual variation exists, and existing dysbiosis makes the first weeks considerably harder for some people. The relevant question is whether that state is fixed or modifiable, and the microbiome is among the most modifiable systems in the body: the intolerance is frequently the deficiency, and the cure looks like the thing causing the discomfort. Conditions like IBS may need a structured, low-FODMAP-informed approach with professional support rather than a cold start. That is a reason for a plan, not a verdict. Every barrier, and what to do about it
“Nobody actually eats this way.” Millions of people do, across every continent and income level, and the number has grown in every developed market for a decade. Popularity was never the argument anyway (it is the appeal-to-popularity fallacy wearing a shrug), but as a factual matter the fringe is considerably larger than it looks from inside a culture that eats meat three times a day.
“The transition felt terrible on carnivore too, and that was my body healing.” One note on the mirror image of this section: the fat-adaptation and “oxalate dumping” framing is constructed so that feeling worse counts as evidence the diet is working, which makes it unfalsifiable. The plant-based transition story above is different in a specific and checkable way: it predicts that symptoms resolve, that bloodwork improves, and that both can be measured.
Identity, framing and rhetoric
The last category contains no empirical claims at all. Every item is a move designed to make the question go away without answering it, which is the definition of a thought-terminating cliché, and they are worth naming precisely because they work.
Is veganism a cult?
“It’s a hypocritical cult.” Cults are defined by features that are absent here and easy to check: a living leader with authority over members, restricted access to outside information, punishment or shunning for departure, financial extraction, and the deliberate severing of outside relationships. Veganism has no leader, no hierarchy, no membership, no fee, no scripture and no mechanism to enforce anything, and people leave constantly with nothing happening to them. What it has instead is a definition published in 1979 and a body of evidence anyone can check independently, which is close to the opposite of a closed information system. The word is doing something specific in the sentence: it converts a question about your behaviour into a diagnosis of the questioner, so that nothing further needs to be considered. Notice too that the accusation is never applied to the vastly larger group that eats animals because everyone around them does, which is what an inherited, unexamined, socially enforced position actually looks like. If the objection is really that some vegans are dogmatic, that is a complaint about people, and people are dogmatic about everything. It says nothing about whether the position is correct. Thought-terminating clichés, in Veganism Rewired
Fallacy: thought-terminating cliché · ad hominem
Is veganism a religion?
“It’s basically a faith.” Religions rest on claims that cannot be tested (the existence of the divine, the fate of the soul) and are held on faith by design. Every load-bearing claim in this post is the opposite kind: testable, falsifiable, and in fact tested, which is why this article can cite retractions, name where the evidence is weak, and concede the bone-fracture data rather than defending an article of belief. A religion does not update when the fracture data come in. The comparison also cuts in an awkward direction, since the belief that eating animals is necessary, natural and normal is genuinely transmitted through ritual, family, holiday and custom, largely without examination. Being a moral commitment does not make something a religion: opposing cruelty to dogs is a moral commitment too, and nobody files it under theology. The word is chosen because it implies a claim immune to evidence, which is exactly what the person does not want to test.
Fallacy: category error
Isn’t veganism just an ideology?
“You’re pushing an ideology.” If ideology means a coherent set of values applied consistently to decisions, then yes, and so is every other position on this question, including the one that does not know it is one. That is the asymmetry worth naming: a position that matches the surrounding culture feels like neutrality, while a position that departs from it feels like ideology, even though both are choices with consequences for someone else. Somebody who has never once examined why they eat pigs and not dogs is not operating without an ideology; they are operating inside one so widely shared it has become invisible. The word is also functioning as a hygiene label here, implying distortion without having to identify any. The reply is simple and specific: name the claim you think is distorted and we will look at the evidence for it, which is a conversation an ideology cannot survive and a position built on evidence can. Thought-terminating clichés, in Veganism Rewired
Fallacy: unexamined-default bias
Aren’t vegans preachy and self-righteous?
“That’s why I’ll never go vegan; you people are insufferable.” Sometimes accurate, and entirely beside the point of whether the claims are true: the tone of an argument’s worst messenger has no bearing on the evidence. There is a real trap here, though, and it is worth naming honestly: advocacy that is too forceful triggers reactance and entrenches people, while advocacy that is too soft never registers as a reason to change anything. That is the paradox anyone advocating for this has to navigate, and getting it wrong is a real cost that falls on the animals rather than on the advocate. The research on vegaphobia also shows that mocking vegans is one of the few forms of ridicule that remains socially frictionless, which tends to precede the moral discomfort rather than follow it. And notice the shape of the sentence: it makes a decision about animals contingent on the personality of a stranger. The Advocacy Paradox, in Veganism Rewired
Fallacy: ad hominem · tone policing
Aren’t words like “exploitation” and “violence” emotionally loaded?
“Using words like that proves you’re pushing a narrative.” Those words have ordinary definitions, and the question is whether they apply: using an animal as a means to an end without its consent is what exploitation means, and killing something that does not want to die is what violence means. The objection is not that the words are inaccurate, it is that they are uncomfortable, which is a fact about the listener rather than about the language. Notice the alternatives on offer: harvest, process, dispatch, product, unit. Every one of those is also emotionally loaded; they are just loaded in the direction of not feeling anything. Demanding neutral language in a discussion about killing is not a demand for neutrality.
Fallacy: tone policing · euphemism as neutrality
Isn’t it a personal choice?
“It’s just my personal choice. Stop nitpicking a cohort study.” Personal choice is the correct frame for questions with one participant, and it is exactly the phrase we do not accept when a choice has a second party in it, which is why nobody describes how someone treats their dog as a personal choice. That is not an accusation; it is an observation about where the category applies. As a conversational move it is also a way of exiting rather than answering: it changes the subject from whether something is defensible to whether you have standing to ask. The same applies to dismissing a specific critique of a specific study as “nitpicking,” which converts a methodological point into a personality judgement. If the study is sound, the objection should be answerable. The burden here sits where it does in every ethical question, on the party whose action affects someone else, and shifting it onto the person asking is the oldest move in the section. There is a real functional thought buried inside this one, though, and it is worth separating out: many people saying “personal choice” mean “I don’t want to be judged,” which is a request for respect rather than an argument, and it deserves respect in return. Why the burden of proof has shifted
Fallacy: shifting the burden of proof
Isn’t meat optimal even if it isn’t necessary?
“You’re attacking a claim we never made.” This is a fair correction and it is worth accepting immediately, because accepting it costs nothing and clarifies everything: the question is not whether humans can survive without meat, which is settled, but whether they do better with it. That reframing is an improvement, and it also moves the argument onto ground where the evidence is far less friendly to the person making it. “Optimal” is a measurable claim, and it requires showing that adding animal products to a well-planned plant-based diet improves outcomes: lower cardiovascular events, lower cancer incidence, lower all-cause mortality. There is no such literature. The best showings for animal products in the outcome data are null results; the gradients that do exist run the other way at every measured step. So the retreat from “necessary” to “optimal” is welcome, and it lands on a claim with less support than the one it replaced. And once eating animals stops being necessary, it starts needing a justification, which holds even if every disputed health finding is granted to the other side. Why the burden of proof has shifted
Fallacy: motte-and-bailey
Do you need to farm to have a say on animal agriculture?
“Get your feet on a farm before you talk.” Proximity is not the same as objectivity, and someone whose livelihood depends on an answer is not the neutral party in a dispute about it. That observation cuts both ways and is worth making carefully. Practical farming knowledge is genuinely valuable, and several of the strongest objections in the land section came from farmers, which is why they got full answers rather than dismissal. But the questions at issue here are answered with feed conversion data, land use accounting, emissions inventories and trophic arithmetic, none of which become visible by standing in a field. The argument also quietly assumes its conclusion: that anyone who really understood farming would endorse it.
Fallacy: ad hominem · appeal to credentials
Isn’t caring about farm animals unrealistic?
“Enough about saving animals.” This is not an argument and does not pretend to be; it is a request to stop, and it usually arrives at the point where the previous arguments have run out. Worth noticing that the real world is where all of this is happening (the numbers, the acreage, the animals), and that appealing to realism while declining to examine the reality is the move being made.
Fallacy: dismissal without argument
Isn’t vegan food just Oreos, chips and seed oils?
“Vegan doesn’t mean healthy. It means sugar and processed junk.” Correct, and nothing in this post claims otherwise: fries and soda are technically plants, and a diet of refined flour and fried food that happens to contain no animal products is a different thing from what is being described here. “Well-planned” does real work in every claim on this page, and it should. What the objection cannot do is transfer: the existence of junk-food vegans says nothing about the outcomes of whole-food plant-based eaters, who are the people the evidence is about. The comparison that matters is a well-constructed diet against a well-constructed diet, and it is available to anyone with a bag of lentils. The evidence review
Fallacy: straw man
Why these arguments get made in the first place
Most of what appears in the first half of this page is not really a factual claim. When someone says it is the circle of life, or that we are at the top of the food chain, or that plants feel pain too, they are rarely reporting a belief they arrived at by thinking. They are closing a door against discomfort. That reflex is human rather than stupid, and every one of us has some subject we handle exactly that way. If you are reading this as a meat eater and one of those phrases sounded like something you have said, that is not an accusation: it is the most ordinary thing a mind does when it meets an idea it would rather not sit with. And if you are reading it as a vegan, it explains something that has probably frustrated you for years, which is why the facts you handed someone slid straight off. You were answering a sentence. The sentence was a door.
It also explains why this page has two very different halves. A cliché exists to end thought, so it arrives short, portable and final. A position that was genuinely built after thinking arrives looking like the opposite: structured, sourced, willing to name a study. Both are doing the same job of holding a conclusion in place, which is why some sections here answer a slogan in four sentences while others work through a cohort study, a confidence interval and a conflict of interest. They are not the same kind of object, and treating them as though they were is how these conversations usually go wrong in both directions.
The other reason this page reads the way it does is that the person arguing is almost never the audience. For every commenter there are many more people reading without saying anything, and they are the movable ones. That is why nothing above is written to defeat anybody, why the findings that cut against a plant-based diet are stated at full strength rather than buried, and why the weak spots on my own side are flagged as weak. A watching stranger remembers the tone of a page long after they have forgotten who won.
Which brings me to an odd admission to make at the bottom of twenty thousand words: winning the argument is not where anyone actually moves. Correct the record once, plainly, with sources, and then stop, because the part that changes people happens somewhere else entirely. It looks like listening for the value a person already holds, the compassion or fairness or honesty they are protecting without ever naming it, and meeting them inside it instead of trying to drag them out of it. It looks like leading with the food rather than the label. It looks like leaving someone warmer than you found them. None of that is on this page, because none of it is an argument.
So if you came here to disagree, nothing here asked you to give anything up. If one claim looked worth checking, follow the source and check it for yourself. That is the entire ambition of a page like this one, and it is enough.
And if you came here as a vegan who is tired of meeting these same objections: the method for the other part, the part that actually moves people, is laid out at veganlinked.com/advocacy and worked through in full in Veganism Rewired.
A living document. This post is rigorously researched and actively maintained: I revise it as the evidence develops. I do much of the research and most of the writing and proofing myself, with the assistance of top-tier AI tools, which means that despite careful checking an error can occasionally slip through or be introduced during an update. If you spot something that looks wrong, please tell me in the comments below: I read them, I take accuracy seriously, and I’ll assess and correct genuine errors promptly. Sourcing and citations are provided so you can verify claims for yourself.
Last updated: November 1, 2026 at 8:34 PM ET
