Justified True Belief

(●) A hasty generalization concludes about a whole group from too few cases, or from cases picked to fit. + Everybody generalizes; it is how we get through the day. The fallacy is doing it on evidence that could never carry the weight, and then talking as though it had. (1) The shape of the move is a verdict on everyone, drawn from a few. - A woman moves to a new town, and her first two neighbours never wave back. By Friday she has decided the town is unfriendly. Two houses have been made to speak for eight thousand people she has never met. - Reasoning from the cases you have seen to the ones you have not is induction, and nearly everything we know came that way: every apple so far has fallen, so the next one will. The cases you looked at are the sample; the group you conclude about is the population. - Logicians file the move among the fallacies of weak induction: reasons that really do bear on the claim, and far too few of them. - Hurley's textbook puts two questions to any sample. Is it big enough? And is it representative, meaning does it resemble the population in the ways that matter? Two cold neighbours fail the first. A survey of the town's grumpiest street would fail the second. (2) It tempts because the mind finishes a pattern early. - Nobody has time to check every case, so we learn from a few, and the few we remember are the ones that stung. A rude waiter stays with you. Forty polite ones do not. - A pollster ringing numbers at random has drawn his sample: every voter had the same chance of being asked. Memory never rings at random. What it hands us is a sample chosen by whatever hurt. - So from the inside it feels like evidence. "I've seen it with my own eyes" is honestly said, and the eyes did see it. They just did not see much. (3) Every cousin in the family is a sample chosen instead of drawn. - Aristotle catalogued the ancestor around 350 BC in the Sophistical Refutations: the slide from what is true "in a certain respect" (secundum quid, in the medieval Latin) to what is true flat out. Unfriendly on her street becomes unfriendly, full stop. Hurley's textbook keeps the older name converse accident, the reverse of Copi's fallacy of accident, which forces a general rule onto a case it never fit. - Anecdotal evidence is a sample of one. "My grandfather smoked forty a day and lived to ninety-four." - Cherry-picking is a sample chosen to fit. The estate agent shows the three sunny photographs and none of the damp wall. - Quote mining is cherry-picking done to a text. Darwin wrote that to suppose the eye was formed by natural selection "seems, I freely confess, absurd in the highest possible degree", and in the next sentence began explaining how it had happened anyway. - The Texas sharpshooter fires at the barn first, then paints the target round the tightest cluster of holes. Therefore, a hasty generalization is not a lie about the cases. It is a truth about a few of them, stretched over everyone it never met.

(●) Simple example: "Two rude waiters in one evening. Everyone in this city is rude." + You have been that tourist, tired at the end of a long day, saying it as an observation rather than an argument. (1) What is claimed and what is offered. - The claim is about a city of two million. The support is two men, in two restaurants, on one evening, both near the end of a shift. (2) Why the support does no work. - Ask the two questions. Big enough? Two out of two million is not a sample, it is a coincidence. Representative? Both worked the district where tourists eat, at the hour when the kitchen is slammed. A hundred waiters from that street would only tell you about that street. - Try the swap. Two friendly waiters the next evening would prove the city warm by the same reasoning, and a rule that proves opposite conclusions on alternate nights is not a rule. (3) What the sentence was really doing. - It was turning a bad evening into a verdict, which is restful, because a verdict needs no more looking. "I had a rough night" says exactly what the evidence supports.

(●) A critic says: "Religion causes wars." + Whoever says this has Belfast and the Crusades in mind, and is not making any of it up. The history is real. The generalization built on it is another matter. (1) What is being argued. - The claim is that religion is a leading cause of war, and often that the world would be more peaceful without it. The support is a list: the Crusades, the Thirty Years' War, Northern Ireland, the Middle East. - The published ancestor is Christopher Hitchens's God Is Not Great (2007), whose chapter "Religion Kills" walks through Belfast, Beirut, Bombay, Belgrade, Bethlehem and Baghdad. Every one of those cities saw real blood, some of it shed in God's name. (2) Where the fallacy sits. - The slide runs from "here are wars with religion in them" to "religion causes wars", and the list that carries it was chosen rather than drawn. Draw the sample instead and you meet the two World Wars, Napoleon and Genghis Khan, none of them fought over doctrine. - Drawn samples exist, with limits. Phillips and Axelrod's Encyclopedia of Wars (2005) indexes about 121 of its wars under religious wars. Hold the familiar figure loosely: 123 of 1,763, under seven percent, is one reader's tally rather than a number the authors printed; it counts wars rather than deaths; and the Thirty Years' War lands in either pile depending on who sorts it. - William Cavanaugh's The Myth of Religious Violence (2009) goes further, arguing that the line between religious and political motives is itself unstable. If he is right, that cuts against the critic's category and against a tidy seven percent alike. - Try the mirror. "Atheism causes mass murder: Stalin, Mao, Pol Pot" is the same move from a Christian's chair, three chosen cases doing the work of a census. - Cherry-picking reaches the text too. "I have not come to bring peace, but a sword" (Matthew 10:34) gets quoted as a call to arms. The next verse says what the sword is: a man set against his father, the cost of discipleship inside a family. The same Gospel has him tell Peter, "Put your sword back into its place. For all who take the sword will perish by the sword" (Matthew 26:52). (3) The sound version. - Religion has caused some wars and a great deal of violence, and Christians have shed blood in Christ's name that he never asked for. That charge is real, and it is faced case by case: the Crusades, the wars of religion, and hardest of all the wars God commanded in the Old Testament. - What none of it can settle is whether Christianity is true. A false religion can be peaceful and a true one can have violent followers. (4) How to answer it. - "Some of them, yes, and I won't defend those. But if we're going to count, let's count all the wars, not just the ones that come to mind, and then ask separately whether any of it tells us whether God is there." See also: Common Objections: Did God Command Genocide?

(●) A Christian says: "I prayed and got the job, so prayer works." + The same move from our side of the table, and a warmer one, because it comes out of real gratitude. It is also the easiest sentence for a thoughtful skeptic to take apart. (1) What is being argued. - The claim is that prayer produces results, and behind it, that God is there and answers. The support is one prayer and one job offer, in that order. - The strongest form is not silly. A believer who has prayed for thirty years has a record rather than a case, and a record of the right kind can be evidence. The sentence offers one case and asks it to do a record's work. (2) Where the fallacy sits. - The sample is one, chosen after the fact. A Christian who applied for six jobs, prayed over all six and got the fourth remembers the fourth. The other five went into the drawer marked "not yet", and the target was painted round the hole that hit. - The skeptic has a drawn sample, and it deserves an answer rather than a wave. In a controlled trial, one group gets the treatment and a matched group does not, so any difference can be laid at the treatment's door. - The largest such trial of prayer was the STEP study of 2006. It followed 1,802 heart-bypass patients, some of whom received intercessory prayer, which is prayer offered for another person, here by strangers at a distance. It found no benefit, and the patients told they were certainly being prayed for did slightly worse. One answered prayer cannot stand against that. - What the trial does not show is that God does not answer prayer. It tested a mechanism, as though prayer were a tablet given to one group and withheld from another. Christian teaching about prayer never claimed a mechanism, and God is not obliged to turn up for a trial designed to catch him. (3) The sound version. - Scripture itself refuses the argument. Paul asked three times to be rid of his thorn and was told, "My grace is sufficient for you, for my power is made perfect in weakness" (2 Corinthians 12:9). In Gethsemane Jesus prayed, "nevertheless, not as I will, but as you will" (Matthew 26:39). Prayer, on its own account, is asking a Father, and a Father's no is an answer too. A claim of that shape cannot be tested by counting yeses. - The honest thing to say is "I asked, and I believe God gave", a report of a relationship, which stands or falls with the case for the God it assumes. That case rests on grounds owing nothing to anyone's job offer: whether the universe had a beginning, whether right and wrong are real. (4) How to answer it, when a fellow Christian says it. - "I'm glad about the job, and I'd thank God for it too. But if the job proves prayer works, someone will ask about the prayers that got a no. Let's start with why God is real." See also: Natural Theology: The Kalam Argument Natural Theology: The Moral Argument

(●) A charge that misses: "'No one dies for a lie' takes a handful of martyr stories and turns them into a law of human nature." + The charge is aimed at a real weakness in the slogan, and part of it lands. The rest misses, because the argument the slogan stands in for is not about martyrs. (1) What the charge claims. - Millions have died for false beliefs: kamikaze pilots, the followers at Jonestown, soldiers on both sides of every war. A rule that sincere death proves truth is refuted by a single case, and history offers thousands. On this view the disciples' deaths show that they were brave, and nothing more. (2) Why the charge misses. - The apologist's argument was never that dying proves a belief true. Gary Habermas and Michael Licona, in The Case for the Resurrection of Jesus (2004), draw the line the slogan blurs. A modern martyr dies for what she believes on someone else's word. The disciples claimed to have seen the risen Jesus themselves, so they were in a position to know whether it was true. That is a far smaller class than "people who died for a belief", and the kamikaze pilot is not in it. - Induction from a sample that is big enough and representative is not hasty at all; it is how science works. A drug trial with 1,802 patients is not a hasty generalization, and neither is a conclusion drawn from a dozen witnesses who told the same story and paid for it separately. - One case can refute a universal claim but never establish one. A single black swan on an Australian river in 1697 ended "all swans are white", and one kamikaze pilot ends "no one dies for a lie" said as a law. That much of the charge lands, and the slogan should give way to the narrower claim. - The test to carry away: was the sample drawn or chosen, and how big would it have to be? Hurley's two failures, too small and unrepresentative, are the whole checklist. - Hold up the mirror as well. "Christians who leave the faith never really believed" is a generalization about ex-believers drawn from the few a man happens to know, and it fails the same test. (3) What the charge would need to be true. - It would need the witnesses' sincerity to rest on martyrs in general. It does not. It rests on what the earliest sources say about these men: they proclaimed the appearances within months, in the city where the tomb was, and kept it up under threat. - Where the charge partly lands, grant it. Sincerity shows that the witnesses believed what they said, not that they were right. And the evidence that most of the Twelve were executed is thin, which is why Habermas and Licona lean on the well-attested deaths of Peter, Paul and James the brother of Jesus rather than the full roll. (4) How to answer it. - "You're right that dying for a belief proves nothing; people die for false beliefs every day. The claim is narrower: these men said they had seen him, and they died rather than say they hadn't. That tells us what they believed; whether they were right is the next question." See also: CE / Resurrection: Minimal Facts Argument (Habermas) Christian Evidences: Transformation and Conduct of the Key Witnesses