01 / BEDROCK
The starting points
Five simple ideas almost everyone already accepts
These five ideas are the floor. They are not controversial and not political — most people already act as if they are true, dozens of times a day.
They matter because nearly every disagreement about homosexuality turns out, when you dig, to be a disagreement about one of these five. If two people agree on all five, most of the surface arguments settle themselves. If they don’t agree, then no amount of evidence will help — and it is worth finding that out early rather than after three hours of shouting.
Starting point A1 Suffering is bad. People doing well is good.
The simplest moral idea there is. Nobody argues with it.
Why it matters: If a rule causes real suffering to real people, the rule needs a reason good enough to be worth that suffering. "I don't like it" is not good enough.
Starting point A2 You need a reason to stop someone living their life.
Freedom is the default. Interference is the exception, and the exception has to be argued for. To stop someone doing something, you have to point to a harm — to an actual person. Not to your discomfort. Not to a vague feeling that society is going downhill.
Why it matters: This is the most powerful idea on the page, because it flips who has to prove what. The question is never why should gay people be allowed to do this? The question is what harm does it cause, and to whom? — and that blank has to be filled in by the person who wants to interfere.
Example Your neighbour paints his front door a colour you find hideous. You have to look at it every morning. You genuinely hate it.
You still have no right to stop him. Your annoyance is real — but it isn’t a harm, it’s a preference. Now suppose instead he leaves his bath running and floods your flat. That is a harm, and now you can stop him.
Everyone already knows the difference between these two cases. The whole question is which one homosexuality resembles.
Example · who has to prove what In a criminal trial, the accused doesn’t have to prove they’re innocent. The prosecution has to prove they’re guilty. If the prosecution has nothing, the accused walks free — not because innocence was demonstrated, but because the burden was never met.
A2 says freedom works the same way. Gay people don’t have to justify themselves. The person who wants to restrict them has to produce the harm.
Starting point A3 Treat similar cases the same way.
If you apply a rule to one group, you have to apply it to every similar group — or explain what makes them different.
Why it matters: This is where most arguments against homosexuality quietly die. Nearly every one of them, applied consistently, also condemns infertile couples, elderly couples, couples who don't want children, and adoptive parents. Almost nobody accepts that. Watching them refuse is how the argument ends.
Example Imagine a rule: “People with poor eyesight may not drive, because they’re a danger on the road.” Fine. But then someone points out that people who drive while exhausted, or while texting, are statistically more dangerous — and asks whether they should be banned too.
You now have exactly two honest options. Either say yes, ban them too, or explain what makes the cases genuinely different. What you may not do is apply the rule only to the group you already disliked.
Starting point A4 Facts are settled by evidence, not by how you feel.
“It feels wrong to me” tells you something true — about the speaker. It tells you nothing about the world.
Why it matters: It splits the debate into two piles. Factual questions — is it chosen? is it an illness? are children harmed? — can be settled, and mostly have been. Value questions can't be settled by evidence, but they can be tested for consistency (A3). Mixing the two piles up is the most common error, and both sides do it.
Example Someone can be genuinely disgusted by the thought of eating snails, insects, or raw fish. That disgust is real and it isn’t a character flaw.
But it tells you nothing about whether snails are nutritious. That’s a separate question, answered by measuring, not by feeling. Disgust is a fact about the diner, not about the dinner.
Starting point A5 How things are doesn't tell you how they should be.
No fact about nature, or statistics, or tradition, gets you to a moral conclusion by itself. You always need an extra moral step — and that step is usually the hidden one.
Why it matters: This kills it's unnatural, evolution didn't intend it, and it's always been this way in one stroke. But it cuts both ways. "Animals do it too" and "they were born that way" are also just facts, and they also prove nothing on their own.
Example Cholera is natural. Antibiotics are not. Nobody concludes that we should have more cholera and fewer antibiotics.
Wisdom teeth are natural. Removing them is not. Glasses are unnatural — your eyes were “supposed” to work without them.
“Natural” and “good” are two different words, and once you notice they’ve been quietly swapped, the argument stops working.
Go deeper — why this is called the is/ought gap
The philosopher David Hume noticed in the 1700s that writers would spend pages describing how the world is, and then suddenly, with no warning, start writing about what people ought to do — as if the second followed from the first. It doesn’t. Somewhere in between, a moral premise got smuggled in.
The practical version: whenever you hear a fact followed by “…therefore it’s wrong,” stop and ask what the fact was doing there. A fact alone never carries a “therefore.” Something else is doing the lifting, and that something else is what you should be arguing about.
The pattern repeats endlessly: homosexuality is [some fact], therefore it is wrong. A step is always missing. Finding it is the entire job.
02 / EVIDENCE
What the evidence actually shows
Which questions are closed — and which are not
The point of this layer isn’t to pile up numbers. It’s to work out which questions are finished, so nobody wastes time arguing about them.
One warning first. Overstating the evidence is the main way a good case gets destroyed. If you claim more than a study shows, and someone who knows the study corrects you, everything else you said gets thrown out with it. So every piece of evidence below carries its weaknesses on the front, in red.
Start here The one review to read
Six researchers — including a historian — were asked to summarise what the science actually says, precisely because the political debate kept quoting it badly. It is still the best place to begin.
Bailey, Vasey, Diamond, Breedlove, Vilain & Epprecht (2016). Psychological Science in the Public Interest 17(2), 45–101.
Weakness — It is a review, not new data. Reviews reflect the judgement of their authors — though these six do not all agree with each other, which is part of why it is trustworthy.
2.1 Is being gay a choice?
No.
Three completely separate lines of evidence point the same way. None of them is a “gay gene” — and anyone who claims one exists is about to be corrected by a better-informed opponent.
Evidence · genes Many genes, small effects, no prediction
Genes are involved, but no single gene decides anything, and you could not predict a person’s orientation from their DNA even if you had all of it.
Ganna et al. (2019). Science 365(6456), eaat7693.
Weakness — The study measured same-sex behaviour, not attraction. Other researchers published a formal objection saying this lumps very different people together (Hamer, Mustanski, Sell et al., Science, 2021). Use it for “complicated and partly genetic”. Never for “determined by DNA”.
Example · what 'polygenic' means Think about height. There is no “tall gene.” Height is influenced by hundreds of genes, each nudging things a fraction of a centimetre, plus nutrition, plus illness, plus chance.
And yet nobody thinks height is a lifestyle choice. Nobody thinks a short person could just try harder. “No single gene controls it” and “you chose it” are completely different claims, and the first does not lead to the second.
Orientation works like height, not like eye colour.
Evidence · before birth The older-brother effect
This one is strange, and it is the single most useful fact on the page.
Every older biological brother a man has raises the chance he is gay by roughly a third. It has been found again and again, across different countries and decades, in more than ten thousand people.
Then came the test that settled it. Researchers compared men who grew up with biological older brothers against men who grew up with step- or adopted older brothers.
- Biological older brothers raised the odds — even if the man had never met them.
- Step-brothers he shared a bedroom with for eighteen years had no effect at all.
The number of sons his mother had carried is what mattered. Not who was in the house.
Blanchard & Bogaert (1996). American Journal of Psychiatry 153(1), 27–31. · Bogaert (2006). PNAS 103(28), 10771–10774. · Bogaert et al. (2018). PNAS 115(2), 302–306. · Balthazart (2018). PNAS 115(2), 234–236.
Weakness — Found in men; there is no equivalent established effect in women. It also explains only some gay men, not all of them.
Why this matters so much Upbringing cannot reach backwards in time. A brother you never met, in a house you never lived in, cannot have “influenced” you — but he still shifts the odds. Something happened before you were born.
Go deeper — how could a brother you never met affect you?
The leading explanation involves the mother’s immune system.
A male foetus produces certain proteins that a female body has never encountered. The mother’s immune system may treat them as foreign — the way it treats anything unfamiliar — and produce antibodies against them.
The useful comparison is an allergy. Many allergies get stronger with repeated exposure: the first bee sting does little, the fifth can be dangerous. The immune system “learns.”
The idea is that with each male pregnancy, the mother’s response strengthens slightly. Later sons develop in a womb where that response is a little stronger — and this appears to affect development in the brain regions involved in attraction. Researchers identified a specific protein (NLGN4Y) and found mothers of gay sons had higher antibody levels against it.
The honest caveat: this mechanism is well-supported but not proven. What is established is the effect itself — biological older brothers matter, social ones don’t. The “why” is still being worked out.
Evidence · an accidental experiment The strongest test nobody meant to run
Between roughly 1960 and 2000, doctors sometimes performed surgery on baby boys born with severe genital malformations, or injured in surgical accidents, and had them raised as girls from infancy. Families were told to raise them as daughters. Many never knew.
This is the most complete “nurture” experiment ever performed on a human being — total, lifelong, from birth.
They grew up attracted to women anyway.
The reviewers conclude that this shows how extraordinarily hard it is to push male sexual orientation off course by social means — even when you try with everything you have.
Reviewed in Bailey et al. (2016). Psychological Science in the Public Interest 17(2), 45–101.
Weakness — Small numbers. And these were real people, badly harmed by a medical practice since abandoned. The evidence is powerful; it should be handled with care, not brandished.
Put together If a child can be raised as a girl from birth and still grow up attracted to women — then “he was raised wrong” is not an explanation of anything.
2.2 Is it a mental illness?
No — and the reason people think so proves the opposite.
Homosexuality was removed from the psychiatric manual (the DSM) in 1973, and from the World Health Organization’s list in 1990.
The next question always comes immediately, and it is a fair one: “Then why are gay people more depressed? Why the suicide rates?”
They are. That’s true, it’s been found repeatedly, and denying it is both dishonest and a fast way to lose credibility. So don’t deny it. Explain it.
Evidence · the explanation Minority stress
Being constantly rejected, hidden, insulted, and expecting to be rejected again tomorrow, is exhausting in a way that damages people. Not because of who they are — but because of what is done to them.
Meyer (2003). Psychological Bulletin 129(5), 674–697.
Weakness — It is a framework built on meta-analysis, not a single controlled experiment — social causes are hard to isolate. Its strength is that it makes a testable prediction, and the prediction holds (see 2.4).
Example Children who are bullied at school have far higher rates of anxiety and depression than children who aren’t.
Nobody concludes that being the sort of child who gets bullied is a mental disorder. Everybody understands immediately that the bullying is doing it.
The same logic, applied to the same numbers, gives the same answer here. The only difference is that people are more willing to apply it in the bullying case.
Example · the test that decides it Suppose two doctors disagree about a patient’s cough. Doctor A says the patient has a lung disease. Doctor B says the patient works in a smoke-filled room.
There is an obvious way to find out: take the patient out of the room. If the cough clears, Doctor B was right.
That experiment has effectively been run — see the next section.
2.3 Can it be changed?
No. And the study that claimed otherwise was taken back by the man who wrote it.
Evidence · does it work? The author who withdrew his own paper
For years, the single study most often waved around as proof that conversion therapy works was one published in 2003 by the psychiatrist Robert Spitzer.
In 2012, Spitzer wrote to the journal himself. He said the criticisms had been right all along, and apologised to gay people who had spent years in reparative therapy because they believed he had proved it worked.
Spitzer (2012). Archives of Sexual Behavior 41, 757.
Weakness — Be exact: the journal did not formally retract the paper. Spitzer published a letter withdrawing his own conclusions. Saying “it was retracted” is wrong, and being corrected on it hands the other side a free win.
Evidence · who does it, and what follows Conversion therapy in practice
In a large representative US survey, about 7% of gay, lesbian and bisexual adults had been through some attempt to change their orientation. Of those, 81% received it from a religious leader — not from a doctor or a therapist.
Those who had been through it showed roughly twice the odds of having seriously considered suicide at some point in their lives.
Blosnich, Henderson, Coulter, Goldbach & Meyer (2020). American Journal of Public Health 110(7), 1024–1030.
Weakness — The suicide association is contested — a real, unresolved dispute between researchers. See Layer 05. It should never be presented as settled.
What survives even the dispute Even setting the suicide statistics entirely aside: it is a treatment that doesn’t work, for a condition that isn’t a disease, delivered mostly by people with no medical training. That is already enough. The case does not need the contested number.
2.4 Does acceptance actually change anything?
Yes. This is the most important evidence on the page.
Evidence · families What rejection at home does
Gay, lesbian and bisexual young adults who said their families rejected them strongly as teenagers, compared with those whose families didn’t:
Ryan, Huebner, Diaz & Sanchez (2009). Pediatrics 123(1), 346–352.
Weakness — 224 people, asked to remember their teenage years afterwards. Memory is unreliable, and this is not a controlled experiment. The direction is well-supported; the exact multiplier should be quoted as a link, not a law.
Evidence · the law When marriage laws changed, fewer teenagers tried to kill themselves
American states legalised same-sex marriage at different times. That accident of history created something close to a natural experiment.
In states that legalised it, teenage suicide attempts fell compared with states that hadn’t — and the drop was concentrated among gay, lesbian and bisexual students, exactly as you would predict if stigma were the cause.
Raifman, Moscoe, Austin & McConnell (2017). JAMA Pediatrics 171(4), 350–356.
Weakness — Shows a strong link, not proof of cause. States that legalised early were different from states that didn't in other ways too. Other researchers published exactly this objection. Concede it cheerfully — it costs nothing and makes everything else more believable.
Go deeper — what 'difference-in-differences' means, and why it's clever
Suppose you want to know whether a new bypass reduced traffic in a town. You can’t just look before and after — traffic changes for a hundred reasons: petrol prices, weather, the economy.
So you find a similar town that didn’t get a bypass, and measure both, before and after.
- Town A (got the bypass): traffic fell 30%.
- Town B (no bypass): traffic fell 10%.
The 10% is whatever was happening everywhere anyway. The extra 20% is your best estimate of what the bypass did. You subtracted one difference from another — hence the name.
That’s exactly what the marriage study did: states that legalised versus states that didn’t, before versus after.
Why it isn’t proof: it only works if the two towns were really on the same path to begin with. If Town A was already changing faster for some other reason, the method quietly gives you the wrong answer. That’s the objection critics raised — and it’s fair.
If being gay were the illness, acceptance would not be the cure.
2.5 Are children of gay parents worse off?
No.
There is a real objection to the older research here, and it should be admitted straight away rather than resisted: most early studies were small, and used volunteers.
Example · why volunteers are a problem Imagine studying whether a diet works by advertising in a health-food shop for volunteers.
The people who answer are the ones who already think the diet is going well. Nobody who quit in week two and gained five kilos is going to fill in your form. Your results will look great — and they will be worthless.
Early gay-parenting studies had a version of this problem. This is a fair criticism. Concede it. Then answer it.
Evidence · everyone, not volunteers The whole country
Researchers in the Netherlands used government registry data covering the entire population. Nobody volunteered. Nobody could opt out. Every child in the country was in the data, whether their family was thriving or falling apart.
Children raised by same-sex parents did not do worse at school. They did slightly better. The authors suggest a mundane reason: same-sex couples cannot have children by accident. Every one of those children was planned, wanted, and usually arrived into a household with more resources.
Mazrekaj, De Witte & Cabus (2020). American Sociological Review 85(5), 830–856. · Wider review: Manning, Fettro & Lamidi (2014), Population Research and Policy Review 33(4), 485–502.
Weakness — One country, and school results are only part of “doing well” in life. But the volunteer problem is completely gone, and that was the entire objection.
There is one study that says the opposite, and it gets cited constantly. It is dealt with in full at Layer 05 — and anyone who cites the evidence above without first reading that section is going to get ambushed.
2.6 Is it “against nature”?
Not in the way the phrase suggests.
Same-sex behaviour has been documented across an enormous range of animals — insects, birds, mammals, primates. There is a whole scientific literature on it.
Careful This fact does far less work than people think
The researchers themselves warn against calling animals “gay.” The behaviours are wildly varied, and often serve social purposes with nothing to do with human attraction.
So this fact has exactly one job: it shows that “it doesn’t happen in nature” is factually false.
It does not show homosexuality is good. That would be the naturalistic fallacy (A5) — the very same mistake, pointed the other way. Using it that way is an invitation to be knocked down, and deservedly.
Bailey & Zuk (2009). Trends in Ecology & Evolution 24(8), 439–446. · Monk et al. (2019). Nature Ecology & Evolution 3, 1622–1631.
The right use of this fact It’s a shield, not a sword. It stops a false claim. It doesn’t prove anything on its own — and pretending it does will cost more than it wins.
04 / ARGUMENTS
The arguments
Each one at its strongest — then answered
Every entry below follows the same shape: the claim, as people actually say it; its strongest version, the best case its defenders would make; the hidden premise; the answer; and the consistency test, turning the rule back on cases the person already accepts.
The strongest versions are here on purpose. A page that only beats up weak arguments is useless the first time someone makes a strong one.
Argument 4.1 “It's unnatural.”
Strongest version Not the version about animals — that one is simply false. The serious version goes like this:
Body parts have purposes. Eyes are for seeing. Lungs are for breathing. The sexual organs are for reproduction. Deliberately using a body part in a way that cannot possibly achieve its purpose is a misuse of it.
This is a real philosophical position, argued by serious people, and it deserves a real answer rather than mockery.
Finnis (1994). “Law, Morality, and ‘Sexual Orientation’.” Notre Dame Law Review 69, 1049–1076.
The hidden premiseUsing a body part against its biological purpose is morally wrong.
Once that step is out in the open, it becomes very hard to hold.
Example Chewing gum uses the digestive system purely for pleasure, with no nutrition at all. Nobody thinks it’s immoral.
Whistling isn’t what lips are for. Handstands aren’t what arms are for. Earrings, hair dye, painkillers, glasses — all of these override or ignore what the body “intended.”
People don’t actually believe the hidden premise. They just hadn’t noticed they were leaning on it.
Consistency test “An eighty-year-old couple gets married. They will never have children. Is their marriage wrong? Is their sex life wrong?”
Almost nobody says yes.
And the instant they say no, they have conceded the whole argument. They’ve just agreed that sex and marriage have real purposes beyond making babies — love, companionship, comfort, commitment, building a life with someone.
Every single one of those is available to a same-sex couple.
Don’t gloat. Just let it sit.
Go deeper — the serious philosophical reply
Defenders of the natural-law view saw the elderly-couple problem coming, and tried to escape it. Their move is roughly: an infertile heterosexual couple still performs the kind of act that is naturally ordered towards reproduction, even if it can’t work in their particular case — like a broken clock that is still a clock.
The counter-argument is that this doesn’t hold up. If what matters is the kind of act rather than the actual result, the concept is being stretched to include exactly the cases they want and exclude exactly the cases they don’t — which is not a principle, it’s a preference wearing a principle’s coat.
Stephen Macedo made this argument in detail, and it is the standard reply in the philosophical literature.
Macedo (1995). “Homosexuality and the Conservative Mind.” Georgetown Law Journal 84, 261–300.
Argument 4.2 “It's a choice. It's a lifestyle.”
Strongest version There is something right here, and admitting it immediately is what buys credibility for everything after.
Behaviour is chosen. Whether to act on an attraction is a decision. So is whether to come out, whether to call yourself gay, whether to live openly. Those really are choices.
The distinction between attraction, behaviour, and identity is real, and it is the other side’s best foothold in this argument.
Concede it. Then draw the line where it belongs: attraction itself is not chosen.
Example Ask anyone, gently: “When did you decide to be attracted to your husband?”
The answer is always some version of “I didn’t decide, it just was.”
That is the point. Nobody sits down and picks. You can choose what to do about an attraction. You cannot choose to have it.
And then the older-brother evidence (2.1), which no theory of “choice” or “upbringing” can survive: a brother you never met, in a house you never entered, still shifts the odds.
Important warning "Born this way" is a trap — don't build on it
It’s tempting to make innateness the foundation of the whole case. Don’t. Two reasons:
1. It puts the moral argument at the mercy of future science. Ask honestly: if researchers announced tomorrow that orientation was 100% chosen, would discrimination suddenly become acceptable? Obviously not. Which proves the science was never doing the moral work.
2. It asks for pity instead of equality. “They can’t help it, poor things” is a weak place to stand. Religion is chosen — and gets full legal protection. Political belief is chosen — and gets full legal protection. Being unable to help something is not why people deserve respect.
The real foundation is A2: it harms nobody. Innateness is only useful as a reply to “they could just stop.” It is a shield. It is not the case itself.
Argument 4.3 “It's a mental illness — look at the depression and the suicide rates.”
The hidden premiseIf a group has more mental illness, then whatever defines that group must itself be a disorder.
The fact is true, and should be agreed with, not denied. The reasoning is what breaks.
Example · run the rule on other groups Apply that hidden premise consistently and look what happens:
- Bullied children have more anxiety → so being a bullied child is a disorder?
- Recently widowed people have far more depression → so widowhood is a disorder?
- Refugees have very high rates of PTSD → so being a refugee is a disorder?
Nobody accepts any of these. In every case we understand instantly that the world is doing something to these people, and the distress is the evidence of it.
Consistency test “If being gay were the illness — why is acceptance the cure?”
This is the most efficient sentence on the page. It takes the other person’s own observation and turns it into evidence against their own conclusion.
Because the numbers move. When marriage laws changed, teenage suicide attempts fell (Raifman 2017). When families accept their child instead of rejecting them, suicide attempts drop by a factor of eight (Ryan 2009).
Diseases don’t work like that. Wounds do.
Argument 4.4 “A child needs a mother and a father.”
Strongest version Two halves — one weak, one genuinely strong.
The weak half: mothers and fathers each give a child something the other cannot.
The strong half: for decades, the studies showing “no difference” were small, used volunteers, and were run by researchers who wanted a particular answer. This is a fair criticism. Anyone who denies it is not arguing in good faith.
So agree with it — and then answer it with data the criticism cannot touch: the entire Dutch population (2.5). No volunteers. No opting out. No deficit.
The hidden premiseWhat matters for a child is the sex of the parents.
What the research actually finds matters is something else entirely: stability and resources. Children do badly when family life is chaotic — separations, custody battles, new partners arriving and leaving. That is what does the damage. Whether the parents are two women, two men, or a man and a woman turns out not to be the variable.
Example Imagine someone claims that red cars are dangerous, and produces data: drivers of red cars crash more often.
Then someone checks, and finds red cars are disproportionately bought by young men, who drive faster. Control for the driver’s age, and the colour effect vanishes entirely.
The colour was never doing anything. It just happened to travel alongside the thing that was.
“Same-sex parents” was the red car. Family instability was the young driver. Keep this example in mind — it is exactly what went wrong in the Regnerus study at Layer 05.
Consistency test “Should single parents be banned from raising children? Widowers? The evidence of disadvantage there is far stronger than anything ever produced against gay parents.”
The answer is always no.
Then ask, calmly: on what principle? There is no answer that doesn’t also clear same-sex parents. That’s not a rhetorical trick — it’s the point.
Argument 4.5 “The Bible says it's a sin.”
Arguing about scripture is usually a mistake. You are fighting on someone else’s ground, in their vocabulary, about their sacred text — and the moment it feels like an attack on their faith, they stop listening and nothing after that lands.
There are two better routes, and the second is much stronger.
Route 1 — everyone already picks and chooses
The same body of religious law also forbids mixed fabrics, shellfish, lending money at interest, and divorce.
The point of raising this is not to mock. Mocking loses. The point is to establish something quiet and important: the person is already interpreting. They are already deciding which rules bind them today and which belonged to another time.
Once that is admitted, the question changes shape. It is no longer “what does the text say?” It becomes “why this rule and not those?” — a question about the person, not about the book.
Route 2 — belief and law are different things
This route doesn’t touch their faith at all. It separates two things that get glued together:
- “I believe this is a sin.”
- “Therefore it should be forbidden, or the person may be treated badly.”
The second does not follow from the first — and most people, when they think about it, don’t actually believe it does.
Example Many Catholics believe missing Mass on Sunday is a sin. Not one of them is campaigning to make it illegal.
Many observant Jews believe eating pork is forbidden. They are not trying to ban bacon for everyone else.
Plenty of Christians believe divorce is wrong — and still don’t want divorced people fired from their jobs.
People already know how to hold a religious conviction without imposing it. They do it every day. This argument just asks them to do the thing they already do.
Consistency test “Should the law — or how we treat a member of our own family — be based on a religious rule that the person themselves doesn't share?”
Never attack the faith. Attack only the leap from private belief to public rule. Go after the faith and you turn someone who might have moved into someone who never will.
Argument 4.6 “Where does it end? Paedophilia? Bestiality? Incest?”
This is the argument most likely to make someone furious — and it is genuinely the easiest one here, because the answer is a single word.
Consent.
A child cannot consent. An animal cannot consent. Two adults can.
And this is not a convenient line invented for this argument. It is the same line we already use everywhere else, and nobody finds it controversial anywhere else.
Example · consent is already doing this job A surgeon cuts open your chest with a knife. That’s surgery, and we pay them for it.
A stranger in an alley does the identical physical act. That’s attempted murder.
Same act. Same knife. The only difference is consent — and it turns one into a job and the other into a life sentence.
Or: taking someone’s car with permission is borrowing. Without permission it’s theft.
The move Their proposed line — man and woman — does not separate rape from marriage. It does not separate assault from surgery. It doesn’t do the job.
The line drawn here — consenting adults — separates all of them cleanly, and always has.
Be honest here Adult incest is a genuinely different case, and bluffing about it will cost the whole exchange.
There, real objections exist that can be named: risk of genetic harm to any children, and the fact that families contain power imbalances that make free consent doubtful. Those are actual harms, pointed at actual people (A2).
The gay case has no such harm to point at. That’s the difference — and saying it plainly is far stronger than pretending the cases are identical.
Argument 4.7 “They recruit. They'll influence children.”
This is false, and it should be said plainly. There is no evidence that being around gay people — gay parents, gay teachers, gay neighbours — changes a child’s orientation. If it could, orientation would be far easier to shift than every piece of evidence in Layer 02 shows it to be. Children of same-sex parents are not more likely to be gay.
Notice the contradiction The same person very often makes both arguments — 4.2 and 4.7 — sometimes in the same conversation.
- If chosen, then nobody can be recruited. A choice is made by the person choosing.
- If catchable, then it was never a choice in the first place.
They cannot both be true. Making someone pick one is more effective than refuting either — and it’s a fair question, not a gotcha. They simply hadn’t noticed.
Example Millions of gay people were raised entirely by straight parents, taught by straight teachers, surrounded by straight couples on every television programme, in a world that told them constantly and from every direction to be straight.
It didn’t take. If a lifetime of overwhelming heterosexual influence can’t make someone straight, one gay teacher is not going to make anyone gay.
Argument 4.8 “I've nothing against them — but why the parades? Why so visible?”
The hidden premiseTheir being visible costs me something.
Ask what the cost is. Pressed honestly, the answer is nearly always discomfort. And discomfort is not harm (A2) — see the front-door example in Layer 01.
Example Some people are uncomfortable seeing an interracial couple holding hands. Some are uncomfortable seeing a woman in a hijab, or a man in a kippah, or someone heavily tattooed.
We do not ask any of those people to be less visible. We concluded, correctly, that the discomfort was the observer’s problem to manage, not the other person’s problem to solve.
And visibility isn’t decoration. It’s the thing that breaks the isolation — and isolation is what’s killing the teenagers in the numbers at 2.4. A gay fifteen-year-old in a small town who has never once seen an adult like himself is the population in the Ryan study.
Consistency test “Be married — but never mention your spouse. Never hold hands outside. Never put a photo on your desk. Would that be acceptable?”
The request to “tone it down” is, structurally, a request to go back in the closet. Making it symmetrical is what reveals that — usually to the person making it, who genuinely hadn’t seen it.
Argument 4.9 “Fine — but not marriage. And not adoption.”
The hidden premiseThere is something marriage is FOR that a same-sex couple cannot provide.
Make them name it. Every available answer has already been closed:
- “Children.” → The eighty-year-old couple (4.1). They can’t have children either. Their marriage still counts.
- “Kids need a mum and a dad.” → The whole Dutch population says otherwise (4.4).
- “Tradition.” → A5. Tradition is a fact about the past, not an argument. Slavery was traditional. So was denying women the vote.
- “It changes what my marriage means.” → Ask how. Concretely. Nobody has ever been able to finish that sentence.
Example When divorce was legalised, nobody’s existing marriage evaporated. When interracial marriage was legalised in 1967, no white couple’s marriage meant less the following morning.
Rights are not a fixed quantity. Extending them to someone else does not remove them from you. This feels obvious in every case except the one currently being argued about — which is itself a clue.
And the evidence closes it: when marriage laws changed, teenage suicide attempts fell (Raifman 2017). The law is not just a symbol. It reaches into people’s lives, and it changes how long some of them live.
Argument 4.10 “Tolerance now means I have to approve. I won't be forced to celebrate.”
Strongest version This one deserves genuine respect, because it is partly right — and because it’s the argument where the whole conversation is most often actually won or lost.
Nobody can be forced to approve of anything. Approval isn’t the sort of thing that can be extracted. And demanding it tends to produce resentment rather than agreement.
So don’t demand it. Nobody is asking for approval. What is being asked for is much smaller: that people not be sacked, beaten, refused a flat, or have their relationships legally erased.
Example Most people think at least one of the following is a bit distasteful: heavy drinking, extreme sports, huge tattoos, gambling, casual sex, extravagant wealth, competitive eating.
Almost none of them want any of it banned, and none of them want the people who do it fired from their jobs.
Everyone already knows how to disapprove of something without persecuting it. That’s the entire ask.
Nobody has to celebrate. They only have to not harm.
Note Offer this exit rather than blocking it. Someone who moves from “it’s wrong and should be stopped” to “it’s not for me, but it’s not my business” has travelled an enormous distance.
Sneering at that will send them straight back. People concede far more when they’re allowed to keep their dignity while doing it — and someone who has been humiliated into silence has not been persuaded of anything.
05 / WEAK GROUND
Where this page is weakest
The objections, stated honestly
This layer is not politeness. It’s necessary.
Any argument collection that hides the strongest objections to its own evidence is propaganda, and it will fall apart the first time it meets someone who has done the reading. Everything below will be used against the layers above — so it is better learned here than discovered mid-argument.
Weak point · the contested study The Regnerus study
In 2012 the sociologist Mark Regnerus published findings claiming that the grown children of parents who had been in a same-sex relationship did worse on all sorts of measures. It is still cited constantly, and anyone arguing this topic will meet it.
What actually went wrong
His “same-sex parents” group was not what it sounds like. It was mostly people whose parent had at some point been in a same-sex relationship — typically after a straight marriage collapsed. Very few had actually been raised by a stable gay couple.
He then compared them against intact, never-divorced families.
Example · what this is like Imagine studying whether left-handedness makes children unhappy.
You build your “left-handed” group mostly out of children whose parents went through a bitter divorce. You build your “right-handed” group out of children from stable, intact homes. You find the left-handers are less happy.
You have discovered absolutely nothing about handedness. You have rediscovered that divorce is hard on children — and mislabelled it.
This is the red-car problem from 4.4, and it is the entire flaw.
What happened when others checked his data
- Cheng & Powell found large numbers of miscategorised people. Fixed, the differences mostly vanished. (Social Science Research 52, 615–626, 2015.)
- Rosenfeld re-ran the analysis accounting for family upheaval — which Regnerus had simply left out — and most of the negative findings stopped being statistically meaningful. (Sociological Science 2, 478–501, 2015.)
- The journal published a critical audit of its own review process for the paper. (Sherkat, Social Science Research, Nov 2012.)
Do not overstate this The paper was not retracted. It was not fraud. It was a serious methodological failure, and its author has published defences of it. Saying otherwise to someone who knows the details hands them the argument.
The accurate sentence: “It compared children from broken homes with children from stable homes, and blamed the difference on the parents being gay. When you account for family stability — in his own data — the effect disappears.”
Weak point · a live dispute Conversion therapy and suicide
The finding in 2.3 — that people who went through conversion therapy have higher suicide rates — has been directly challenged, and the challenge is not stupid.
The study measured suicidal thoughts across a whole lifetime. The critic, Sullins, pointed out that some of that suicidality may have happened before the therapy, not after. And if you only count what happened afterwards, he argues, the association reverses.
The original authors published a reply disputing this. The dispute is not resolved.
Sullins (2022), Archives of Sexual Behavior. · Blosnich et al. (2023), Archives of Sexual Behavior 52(3), 885–888.
Example · why the objection has force Suppose you found that people who visit A&E die sooner than people who don’t, and concluded that hospitals kill people.
The problem is obvious: people go to hospital because they’re already ill. The illness came first. You have the order backwards.
Sullins is making that kind of argument — that distressed people are more likely to be sent to conversion therapy in the first place, so the distress may have come before, not after.
What survives anyway Drop the suicide claim entirely and nothing important is lost. It remains a treatment that doesn’t work, for a condition that isn’t a disease, given mostly by people with no medical training — and its most famous supporting study was withdrawn by its own author. That is more than enough.
Weak point · method The marriage-law studies show a link, not proof
The Raifman finding (2.4) compares states, not identical worlds. States that legalised same-sex marriage early were already different from states that didn’t — more liberal, more urban, possibly already improving. Some of the drop in suicide attempts may belong to those other differences.
Critics said exactly this in print, and they were not wrong to.
Concede it happily. It costs nothing, and it makes every other claim on this page more believable. Someone who admits the limits of their own evidence is much harder to dismiss than someone who doesn’t.
Weak point · genuinely open The questions evidence cannot settle
These are not scientific questions, and pretending they are is dishonest:
- Religious freedom versus anti-discrimination law. The baker who won’t make the cake. The church hall. The religious adoption agency. Reasonable, decent people disagree here — and this is the strongest ground the other side has left. It should not be waved away as obvious.
- Whether the state should be endorsing any kind of relationship at all, rather than just registering contracts between adults.
- How to weigh one person’s conscience against another person’s equal treatment in the genuinely hard cases.
Knowing where you don’t have a knockdown answer is the difference between understanding a position and reciting one.