Answering objections to veganism, honestly

In 2003 an animal scientist named Steven Davis did a piece of arithmetic that vegan advocates still have not fully put to bed: he estimated that a hectare of wheat or soybeans kills more field mice, snakes and ground-nesting birds a year than a hectare of pasture grazed by cattle, and argued in the Journal of Agricultural and Environmental Ethics that eating grass-fed beef therefore causes less death than eating vegetables.

Davis put the toll at about 15 wild animals killed per hectare per year in annual cropping, against roughly half that, 7.5, on grazed perennial pasture, mostly rodents and ground nesters killed by tillage, harvesting machinery and rodenticide rather than by anything eating the crop itself.

Gaverick Matheny answered him in the same journal a few months later, and found the problem was the arithmetic, not the biology. Davis had compared land area rather than food produced, and a hectare of pasture yields far less usable food than a hectare of grain. Once Matheny recalculated deaths per person fed rather than per hectare, the omnivorous diet needed several times more land, and therefore several times more field deaths, to feed the same number of people, which turned Davis's own numbers into a case for eating plants directly rather than routing them through a cow.

Feed conversion compounds the same error a second way. Most of the world's grain harvest is not eaten by people at all: 36% of global crop calories are fed to livestock rather than eaten directly, according to a 2013 analysis by Emily Cassidy and colleagues in Environmental Research Letters. Whatever field deaths a hectare of soy causes, a diet built on meat causes several hectares' worth of them, because the animal has to eat the crop before the human eats the animal.

The dispute did not end there. Philosophers Bob Fischer and Andy Lamey published a 2018 review in the same journal arguing that nobody actually has a reliable number: field death counts depend on which animals are counted, whether machinery deaths are separated from habitat-loss deaths, and whether preventing an animal's birth counts as a harm at all. Their conclusion was that the crop death estimate is far shakier than either side treats it, in either direction.

Plants do sense and respond to the world in ways that are genuinely interesting. Researchers led by Monica Gagliano reported in 2014 in Oecologia that the sensitive plant, Mimosa pudica, stopped curling its leaves after repeated harmless drops and still remembered not to react weeks later, an experiment that has since been disputed by other researchers who argue the effect is closer to motor fatigue than memory.

None of this is pain in the sense that matters morally. Pain in animals depends on nociceptors, nerve endings tuned to tissue damage, wired through a nervous system to a brain that turns the signal into a felt, unpleasant experience. Plants have none of the three: no nociceptors, no nerves, no centralised brain. A 2020 review in Protoplasma, on whether anaesthesia implies plant consciousness, found that anaesthetics do measurable things to plants, closing stomata and slowing growth, but nothing in that response requires anything to be experienced.

Taken seriously, the objection undercuts itself. If plant sensitivity were plausible enough to matter morally, the diet that destroys the least plant matter would still be the vegan one, since feeding grain to a cow before eating the cow multiplies the plant matter consumed several times over, the same feed conversion arithmetic behind the crop death argument. Worrying about plants is a reason to eat them directly, not a reason to cycle them through livestock first.

A lion cannot live on vegetables. Its gut is proportionally far shorter than a grazing animal's, it produces no salivary amylase to begin breaking down starch, and it belongs to a species built by natural selection to eat only meat. Humans are built differently: we carry an average of 6.5 copies of the starch-digesting AMY1 gene against 2 in chimpanzees, a difference geneticist George Perry and colleagues linked in a 2007 Nature Genetics paper to a long evolutionary history of eating starchy food.

That is the appeal to nature compressed into biology. Lions are obligate carnivores: their physiology permits nothing else. Humans are facultative omnivores whose gut, genes and enzymes point the other way, and even that evidence only describes what a body can process, not what a person ought to eat. Poison ivy is natural. Antibiotics are not. Nobody treats naturalness as the deciding test anywhere else in life, and pointing at a lion does not settle what belongs on a plate.

Protein is the most common objection at a dinner table and the shortest to answer properly. A varied plant diet supplies enough protein for adults without deliberately combining foods: lentils, tofu and seitan each carry roughly 9 to 25 grams of protein per 100 grams cooked, and the recommended intake for a sedentary adult is 0.8 grams per kilogram of body weight a day, a threshold most omnivores and vegans alike already clear from ordinary meals. Amino acid completeness, timing for athletes and the real gaps that do exist are covered properly on the protein evidence page.

Ask a vegan what they would do stranded on a desert island with only a pig for company, and the question being asked is about survival, not about the aisle of a supermarket that the original argument was meant to settle.

The Vegan Society's 1979 definition builds the answer into the philosophy itself: veganism seeks to exclude cruelty and exploitation as far as is possible and practicable, not absolutely and unconditionally. An island with no boat and one pig is a possible-and-practicable calculation with a single answer. A supermarket with a produce aisle is a different calculation with a different answer, and citing the first does not change the arithmetic of the second.

A single shopper who stops buying chicken does not save a single chicken. Economists studying meat markets use a figure called the cumulative elasticity factor, how much less of a product actually gets produced when demand for it drops by one unit, and estimates for chicken run between 0.06 and 0.7 depending on the study, meaning someone who stops eating ten chickens a year causes somewhere between under one and seven fewer chickens raised, not ten, according to a 2014 review by Animal Charity Evaluators of the agricultural economics literature.

The gap exists because supply chains buffer demand. Producers plan a year or more ahead, prices adjust to clear a smaller market, and part of any drop in demand is absorbed by other buyers paying a slightly lower price rather than by fewer animals being bred, an effect agricultural economists Jayson Lusk and F. Bailey Norwood describe in their published work on the economics of vegetarianism. The number is real but fractional, never zero and never one to one.

Fractional does not mean irrelevant once it is multiplied. A million people making the same change moves the same elasticity against the same market a million times over, which is the actual mechanism by which individual diet change reaches an industry: aggregate demand shifting a price, not moral example shifting a conscience.

Telling someone in a low-access neighbourhood to cook more lentils ignores that lentils require a stove, time and a nearby store that stocks them, none of which is universal. This is the strongest objection to any universalist vegan message, and it deserves to be conceded rather than argued around.

Food justice organisers increasingly use the term food apartheid rather than food desert, arguing in a 2026 paper in the Journal of Agriculture, Food Systems, and Community Development that unequal food access is the product of zoning, redlining and decades of disinvestment rather than a naturally occurring gap. That framing matters for the objection: the barrier is structural, not a preference that better recipes can talk anyone out of.

Food is inherited, not chosen fresh each generation. A dish carries a grandmother, a holiday, a region, in a way a nutrient does not, and a message that treats every traditional dish as a problem to be solved will alienate the people it most needs to reach. Veganism as a personal ethic and veganism as a universal prescription are different claims, and only the first survives contact with unequal food access and cultural weight intact.

Most of these objections do not arrive as arguments. They arrive over dinner, from someone reaching for a reason to end a conversation they did not start, and answering them like a debate tends to make the evening worse for both people.

Julia Minson and Benoit Monin documented part of the reason in a 2012 study: meat eaters who were reminded that a vegetarian might judge them rated that vegetarian more negatively afterward, purely from anticipating the judgment rather than from anything the vegetarian actually said. The researchers called the pattern do-gooder derogation, and it means a vegan who says nothing preachy can still trigger the defensive reaction just by existing as a reminder.

Winning the argument carries its own cost, documented separately as moral licensing. Anna Merritt, Daniel Effron and Benoit Monin's 2010 review found that people who feel they have already defended a choice successfully often feel freed to indulge in the behaviour shortly afterward rather than corrected by having lost the point. A dinner guest who concedes the crop deaths argument at the table is not measurably more likely to eat less meat the following week.

What tends to work instead is unglamorous: ask what someone already believes before stating anything, concede the part that is actually correct, crop deaths are real, protein needs real attention, food access is genuinely unequal, and stop once the honest concession has been made. None of that requires believing less. It requires wanting to be right less than wanting to be heard.

It is also fair to say plainly that a share of vegan advocacy is obnoxious: the uninvited photo, the lecture at someone else's barbecue, the assumption that disagreement is a moral failure. The objection to being lectured is a fair complaint about a real behaviour, not a dodge, and conceding it costs nothing because it happens to be true.

The strongest version of any of these objections deserves a straight answer rather than a smile and a change of subject, but a straight answer delivered once, calmly, at someone else's dinner table has rarely finished convincing anyone of anything on the spot. Curiosity does the work that rebuttal only pretends to.