(●)Testimony is someone telling us that something is so.
+Marta has been sitting in a hospital waiting room for four hours when the surgeon comes through the double doors, still in scrubs, and tells her that the operation went well and her husband will be fine. Marta has never seen the inside of an operating room, and she could not read his chart if someone handed it to her. She believes the surgeon at once, and she is right to. Then she phones her sister, who believes Marta, and by evening a dozen relatives in three states know that he came through, though not one of them saw a thing.
That ordinary chain of telling and believing is what philosophers study under the name testimony.
- Testimony, in this sense, is any case of someone telling you that something is so, and to believe on testimony is to believe it because you were told. It need not happen in a courtroom. A surgeon's report counts, and so do an encyclopedia, a road map, a textbook, a news story, the label on a medicine bottle, and a letter from a friend.
- Almost everything a person knows runs through it. You know your own birthday, who your parents are, that the earth is round, that Australia exists, and that Lincoln was shot, because someone told you.
- Augustine noticed this sixteen centuries ago. Thinking over how much he believed without having seen it, he recalled "with how unshaken an assurance I believed of what parents I was born, which I could not know, had I not believed upon hearsay" (Confessions, Book 6).
So the question is not whether to rely on testimony. Everyone does, and no one could live otherwise. The real questions are what makes it reasonable to believe what we are told, and when, and how much.
For Christians those questions are close to home, because after the first generation everyone who has believed the gospel has received it as testimony. Luke wrote his Gospel from what "those who from the beginning were eyewitnesses and ministers of the word have delivered" (Luke 1:2), and John wrote, "that which we have seen and heard we proclaim also to you" (1 John 1:3).
(●)Philosophers divide over whether testimony needs supporting reasons or is trusted by default.
+Ask why Marta was right to believe the surgeon, and you will hear two classic answers, with a family of newer ones gathered around them.
The first answer is reductionism. It holds that testimony justifies a belief (makes it reasonable to hold) only when the hearer has other reasons, from her own perception, memory and reasoning, to think the report is reliable. On this view, Marta's trust is reasonable because of what she has seen and remembered for herself about reports like the surgeon's. The name comes from the idea that the justification testimony gives can be "reduced" to those other sources.
- David Hume put it this way in 1748: "our assurance in any argument of this kind is derived from no other principle than our observation of the veracity of human testimony, and of the usual conformity of facts to the reports of witnesses." Veracity means truthfulness: we trust witnesses because we have watched them tell the truth, and watched the facts bear them out, many times before.
- Reductionists differ over how much the hearer needs. Global reductionism asks for reasons to trust testimony as a whole: a track record of reports checked against the facts.
- Local reductionism, defended by Elizabeth Fricker in the 1990s, asks for less. It wants only reasons to trust this kind of report, from this kind of speaker, on this occasion: what this pharmacist says about this prescription, for instance.
The second answer is anti-reductionism. It holds that testimony is a basic source of justification, like perception or memory. A basic source gives reasons on its own, without borrowing them from anything else, the way seeing a red apple makes it reasonable to believe the apple is red without anyone first proving that your eyes work. On this view, a hearer may believe what she is told unless she has some reason not to. Marta needed the surgeon's word and nothing more, so long as nothing gave her cause to doubt it.
- Thomas Reid, the eighteenth-century Scottish philosopher and minister, argued in 1764 that God built two matching instincts into us. One is "a propensity to speak truth." The other, "implanted in us by the Supreme Being, is a disposition to confide in the veracity of others, and to believe what they tell us."
- He called them the principle of veracity, our bent toward telling the truth, and the principle of credulity, our bent toward believing it. (Credulity here means a readiness to believe, not gullibility.)
- More than two centuries later, the Australian philosopher C. A. J. Coady (1992) argued that reductionism cannot work, because the track record it asks for is itself known almost entirely through testimony.
- Tyler Burge (1993) proposed an acceptance principle: a person is entitled to accept what is presented to him as true, if he can understand it, unless he has stronger reasons not to. To be entitled, in Burge's sense, is to have a right to the belief even if you could not give a reason for it. He rested that entitlement on the tie between speech a hearer can understand and the reasoning mind behind it.
Newer views look for a middle way.
- Jennifer Lackey (2008) argues that both classic answers leave something out. The speaker's report must actually be reliable, she holds, and the hearer must also have some positive reason to accept it. She argues as well that testimony can give a hearer knowledge the speaker herself does not have (D6).
- The assurance view of Richard Moran and Edward Hinchman (2005) holds that telling someone something is not handing her evidence. It is giving her your word, an invitation to trust that the speaker takes responsibility for (D5).
Christians stand on every side of these questions. Reid was a minister, and Reformed philosophers, from the Protestant tradition that runs back to John Calvin, follow his lead; Nicholas Wolterstorff and Alvin Plantinga are among them. Richard Swinburne, for his part, defends a principle of testimony: what others report experiencing is probably as they report it, unless there is reason to doubt. It sits beside his own principle of credulity, which says the same about how things seem to us.
(●)Probabilistic evidentialism treats testimony as evidence whose strength can be measured.
+Evidentialism is the view that a belief is justified by the evidence the believer has. Timothy and Lydia McGrew hold a probabilistic version of it: a belief is justified to the degree that the believer's evidence makes it probable. Applied to testimony, this evidential view says that what we are told is not a special source of justification with rules of its own. It is evidence, like a fingerprint or a footprint, and it is weighed the same way.
The weighing turns on one question, asked from where the hearer stands: given everything she knows, how much likelier is this report if what it says is true than if it is false? For Marta, the question is how much likelier the surgeon is to come through those doors and say the operation went well if it did than if it did not.
- That comparison is called the likelihood ratio, or the Bayes factor, after the eighteenth-century minister and mathematician Thomas Bayes. A report that is ten times likelier if true than if false has a likelihood ratio of ten.
- The rule that follows is a form of Bayes's theorem, a proved result in the mathematics of probability, and it is both exact and simple to use. Start with the odds of the claim before the report came in, which means its chances of being true set against its chances of being false. Multiply those odds by the likelihood ratio, and what you get is the odds after the report.
- Suppose a claim starts out unlikely, at odds of one to four: one chance of being true for every four of being false. A report with a likelihood ratio of ten multiplies that to ten to four, which is five to two in the claim's favor.
The idea is old, and Christian thinkers did much to sharpen it in the long argument over Hume's essay on miracles.
- In 1767 the Welsh minister Richard Price answered Hume with a lottery. Before the draw, the chance of any particular sequence of numbers is so small that it "exceeds all conception." Yet we believe a newspaper's report of the result at once, because the paper is far likelier to print the numbers that were drawn than any others.
- The mathematician Charles Babbage pressed the point further in 1837. He showed that if witnesses are truly independent, meaning that nothing ties their reports together except what actually happened (P3), and each is likelier to report an event if it happened than if it did not, then enough of them can outweigh any starting improbability, however small.
- Timothy McGrew and Robert Larmer set the reasoning out in the entry on miracles in the Stanford Encyclopedia of Philosophy, a standard reference work for philosophers. When reports are independent, both if the claim is true and if it is false, their likelihood ratios multiply. They add a warning: the assumption of independence "should not be made casually" (D13).
- In 2009 the McGrews applied the same method to the testimony for the resurrection of Jesus.
This view sits between the classic answers. It does not ask, with global reductionism, for a census of every report checked against the facts. Nor does it grant testimony weight before the hearer has any reason to give it. What it asks is what the hearer knows about this report: whether the speaker was in a position to know, whether he had a reason to mislead, whether the report fits the rest of the evidence, and whether others report the same thing independently.
The case can be set out as a deductive argument, one whose premises work like a locked chain: if each link holds, the conclusion cannot slip out. Each premise need only be more plausible than its denial, not certain beyond all doubt. In outline, much testimony is strong evidence on its own (P1, P2, C1), and independent witnesses who agree multiply that strength (P3, C2).
(P1)A report supports what it says to the degree it is likelier if true than if false.
+This premise holds what people tell us to the same test we use for every other kind of evidence.
(1) Evidence is something likelier if the claim is true.
- Muddy boots by the back door are evidence that your son went out in the rain, because they are far likelier to be there if he did than if he did not.
- A report is an event in the world too. Someone said something, or wrote it down, and we can ask what best accounts for its being said, just as we ask what best accounts for the boots.
- So a report counts as evidence for what it says in exactly the way the boots do: when it is likelier to be given if the claim is true than if it is false.
(2) The strength of a report is how much likelier it is.
- Suppose the school nurse calls to say your daughter has a fever. That call is far likelier if she has one than if she does not, because nurses do not phone parents about fevers that are not there.
- Now suppose a classmate mentions that your daughter "looked sick." That report is likelier if she is ill, too, but only a little, since children say such things for all sorts of reasons.
- Both reports point the same way. The nurse's carries far more weight because the gap between "likely if true" and "likely if false" is so much wider.
- Likelier, here, always means likelier given everything the hearer knows. So two hearers who know exactly the same things should give a report the same weight, and a parent who knows that the classmate exaggerates gives his remark even less, and is right to.
(3) The starting probability matters as well.
- The rule is exact: the odds after a report are the odds before it multiplied by the report's likelihood ratio. So where the odds started matters just as much as how strong the report is.
- A surprising claim begins with long odds against it, and it needs a report strong enough to overcome them.
- But strong reports can overcome very long odds. Before a lottery draw, the chance that a particular number will come up may be one in millions. Yet we believe the newspaper's report of the number the next morning, because the paper is far likelier to print the number that was drawn than any other.
(4) Careful people already weigh testimony this way.
- Courts ask whether a witness was in a position to know, whether she had a reason to lie, and whether her account fits the other evidence. Each of those questions asks, in effect, how likely her testimony would be if it were false.
- Scripture gives the same counsel: "The one who states his case first seems right, until the other comes and examines him" (Proverbs 18:17).
Therefore, a report supports what it says to the degree it is likelier if true than if false.
(P2)Much testimony is far likelier if true than if false.
+A report can be false only if the speaker is lying or mistaken. For much of what people tell us, a hearer has good reason to think neither is likely.
(1) People usually say what they believe.
- Lying takes effort, and it needs a motive. Most of what people tell each other gives them no motive at all: nobody gains by getting the time of day wrong, or last night's score, or what is for dinner.
- Picture a bus driver calling out the next stop. He has no reason to name the wrong one, and every reason to name the right one.
(2) We can often tell when a speaker is in a position to know.
- The surgeon knows how the operation went, and a man who was standing on the corner knows what color the car was. A man who only heard about it from his cousin knows much less.
- Suppose you ask whether the roads are icy this morning. A coworker who just drove in on them can tell you. A friend three states away who glanced at a forecast cannot, however honest she is.
(3) Much testimony can be checked, and is.
- A published train timetable, a bank statement and a dictionary are easy to check and costly to get wrong, so they seldom print what is not so.
- And when you have checked what you were told, it has usually held up, even when it told you something you had no other reason to expect. The bridge really was closed, as the sign said. The store really did have the item the website listed.
(4) A hearer learns much of this firsthand.
- No one can check testimony as a whole, and weighing a report does not require it. What a hearer needs is background knowledge about when reports tend to be true, and much of that she gathers for herself.
- She has watched people for years. She knows from the inside what it takes to lie, because she knows her own temptations, and she knows how rarely ordinary talk gives anyone a reason to.
- A girl learns, long before she could explain it, which of her brothers exaggerates and which one can be trusted with a message.
Therefore, much testimony is far likelier if true than if false.
(C1)Therefore, much testimony strongly supports what it reports.
+This follows from (P1) and (P2) by modus ponens, the most basic valid form of reasoning: if A, then B; A; therefore B. Here, if a report is far likelier if true than if false, it strongly supports what it says (P1). Much testimony is far likelier if true than if false (P2). So much testimony strongly supports what it says.
(1) "Much" does not mean "all."
- Testimony is neither suspect by default nor trusted by default. It is weighed. A rumor passed along by someone with a grudge may carry almost no weight, and a sworn report from a careful eyewitness a great deal.
- Scripture says it in a single line: "The simple believes everything, but the prudent gives thought to his steps" (Proverbs 14:15).
(2) Anti-reductionism gets something right.
- Anti-reductionists are right that we do, and should, believe most of what we are told. The disagreement is over why. We should, not because testimony is trusted by default, but because what we already know makes most ordinary reports far likelier if true than if false (P1, P2).
- Once that likelihood explains why we are right to believe most of what we are told, a special principle of default trust has no work left to do.
- Take the gate number printed on a boarding pass. You walk to that gate without asking anyone to confirm it, and you are right to, because the airline is far likelier to print the gate your flight leaves from than any other.
(3) Testimony carries knowledge.
- What we know of history, of distant places, and of most of science comes to us this way, and it is real knowledge. A student who learns from a good atlas that Lima is the capital of Peru knows it.
- Strong support for a claim that was not unlikely to begin with leaves it highly probable (P1). When such a belief is also true, and the reasoning behind it rests on nothing false, it is knowledge.
(P3)Independent reports that each support a claim multiply its support.
+Two witnesses who have not compared notes are worth far more than one witness heard twice.
(1) Independent means linked only through the truth of the matter.
- Two reports are independent when the only thing tying them together is whether the claim is true. Once that is settled, either way, knowing what one witness said tells you nothing more about what the other will say.
- Picture a crash at an intersection. A woman waiting to cross and a man pumping gas on the far corner each saw the pickup run the red light, and they have never met. Compare a single witness who saw it and told her husband, who then repeats it to the police. That is two voices, but only one sighting.
- Independence is most often lost on the side of falsehood, that is, in how witnesses could come to agree if the claim were false. Witnesses who talked it over, copied one another, or got the story from a common source can agree on something false with no help from the facts. So can witnesses who never met but share one cause of error, like two neighbors who both hear a car backfire and report a gunshot.
(2) Independent support multiplies.
- If each of two independent reports is ten times likelier if the claim is true, the two together are a hundred times likelier. Three such reports are a thousand times likelier.
- Roll two dice. The chance that one shows a six is one in six, but the chance that both do is one in thirty-six, because each die has to come up six on its own.
- Independent witnesses work the same way. For them to agree on a false story, each must go wrong on its own, and in the same way, so the chance of their agreeing is the product of their separate chances. The same holds for their agreeing on the truth, and so the ratios multiply.
(3) Each witness must carry some weight of its own.
- A report that is just as likely if false as if true has a likelihood ratio of one. Multiplying by one leaves the odds exactly where they were, so a hundred such reports add nothing.
- Imagine a weather vane rusted solid, pointing east. It points east whether the wind blows from the east or not, so a dozen such vanes on a dozen barns tell you nothing about the wind.
(4) Scripture's law of witnesses rests on the same insight.
- "A single witness shall not suffice against a person for any crime or for any wrong in connection with any offense that he has committed. Only on the evidence of two witnesses or of three witnesses shall a charge be established" (Deuteronomy 19:15).
- The law demands more than one witness because separate witnesses who agree are far harder to explain away. It also takes for granted that each of them actually saw what he reports.
Therefore, independent reports that each support a claim multiply its support.
(C2)Therefore, independent witnesses together can make probable what no single witness could.
+This follows from two rules: independent reports that each support a claim multiply its support (P3), and the odds after the evidence are the odds before it multiplied by that support (P1).
(1) The arithmetic is striking.
- Suppose a claim starts at one chance in a thousand, and three independent witnesses each make it twenty times likelier. Twenty times twenty times twenty is eight thousand, and so the claim ends at about eight to one in its favor.
- No one of those witnesses could have carried it alone. One of them would leave the claim at about fifty to one against, and two would still leave it less likely than not. It takes all three together.
- This pattern turns up whenever two things are true together: each witness carries only modest weight, and the claim starts with long odds against it. In that situation one witness cannot carry the claim, while enough independent ones can.
(2) Independence must be shown, not assumed.
- Witnesses can collude, copy, draw on a shared rumor, or be fooled by the same thing, and when they do, their agreement counts for much less (D13). Independence comes in degrees, and each case has to be examined.
- One way to see it is through an undesigned coincidence, where one account explains a puzzle in another without either seeming to intend it. The feeding of the five thousand gives a clear case.
• In John's account, Jesus asks Philip, of all the disciples, where to buy bread (John 6:5). John never says why.
• Elsewhere John mentions that Philip was from Bethsaida (John 1:44).
• Luke, who never mentions Philip in the story, places the miracle near Bethsaida (Luke 9:10). Put the pieces together and the puzzle is gone: Jesus asked the man from the nearby town.
- Fits like that are hard to plan and easy to miss, which is why they are evidence of separate witnesses telling what happened.
(3) Christian testimony asks to be weighed this way.
- The New Testament does not ask to be believed on its own say-so. It presents the testimony of named witnesses and invites inspection. Of the crucifixion, John writes of the one who saw it: "his testimony is true, and he knows that he is telling the truth" (John 19:35).
- How strong that testimony is depends on two things: how much weight each witness carries, and how independent the witnesses are (P3). Those are historical questions, and they are argued on their own terms.
See also:
Philosophy / Epistemology: EvidentialismPhilosophy / Epistemology: The Principle of CredulityChristian Evidences / New Testament Criticism: Undesigned Coincidences as Evidence of Gospel ReliabilityChristian Evidences / Resurrection of Jesus: Maximal Data Method (McGrew)
On the epistemology of testimony:
Burge, Tyler. "Content Preservation." Philosophical Review 102 (1993): 457-488.
Coady, C. A. J. Testimony: A Philosophical Study. Oxford: Clarendon Press, 1992.
Fricker, Elizabeth. "Against Gullibility." In Knowing from Words, edited by B. K. Matilal and A. Chakrabarti, 125-161. Dordrecht: Kluwer, 1994.
Fricker, Elizabeth. "Telling and Trusting: Reductionism and Anti-Reductionism in the Epistemology of Testimony." Mind 104 (1995): 393-411.
Goldman, Alvin I. "Experts: Which Ones Should You Trust?" Philosophy and Phenomenological Research 63 (2001): 85-110.
Hardwig, John. "Epistemic Dependence." Journal of Philosophy 82 (1985): 335-349.
Hinchman, Edward S. "Telling as Inviting to Trust." Philosophy and Phenomenological Research 70 (2005): 562-587.
Lackey, Jennifer. Learning from Words: Testimony as a Source of Knowledge. Oxford: Oxford University Press, 2008.
Moran, Richard. "Getting Told and Being Believed." Philosophers' Imprint 5 (2005): 1-29.
Plantinga, Alvin. Warrant and Proper Function. New York: Oxford University Press, 1993.
Swinburne, Richard. The Existence of God. 2nd ed. Oxford: Clarendon Press, 2004.
Wolterstorff, Nicholas. Thomas Reid and the Story of Epistemology. Cambridge: Cambridge University Press, 2001.
On testimony, probability and miracles:
Ahmed, Arif. "Hume and the Independent Witnesses." Mind 124 (2015): 1013-1044.
Babbage, Charles. The Ninth Bridgewater Treatise: A Fragment. London: John Murray, 1837.
Campbell, George. A Dissertation on Miracles. Edinburgh, 1762.
Douglas, John. The Criterion: or, Miracles Examined. London: A. Millar, 1754.
Earman, John. Hume's Abject Failure: The Argument Against Miracles. Oxford: Oxford University Press, 2000.
Kruskal, William. "Miracles and Statistics: The Casual Assumption of Independence." Journal of the American Statistical Association 83 (1988): 929-940.
Lewis, C. S. Miracles: A Preliminary Study. London: Geoffrey Bles, 1947. Chapter 13.
McGrew, Lydia. Hidden in Plain View: Undesigned Coincidences in the Gospels and Acts. Chillicothe, OH: DeWard, 2017.
McGrew, Timothy, and Lydia McGrew. "The Argument from Miracles: A Cumulative Case for the Resurrection of Jesus of Nazareth." In The Blackwell Companion to Natural Theology, edited by William Lane Craig and J. P. Moreland, 593-662. Oxford: Wiley-Blackwell, 2009.
McGrew, Timothy, and Lydia McGrew. Internalism and Epistemology: The Architecture of Reason. London: Routledge, 2007.
McGrew, Timothy, and Robert Larmer. "Miracles." In The Stanford Encyclopedia of Philosophy, edited by Edward N. Zalta. First published 2010; revised 2024.
Paley, William. A View of the Evidences of Christianity. London, 1794. Part 1, proposition 2.
Price, Richard. Four Dissertations. London: A. Millar and T. Cadell, 1767. Dissertation 4.
On the New Testament witnesses:
Habermas, Gary R., and Michael R. Licona. The Case for the Resurrection of Jesus. Grand Rapids: Kregel, 2004.
Lüdemann, Gerd. The Resurrection of Jesus: History, Experience, Theology. Translated by John Bowden. Minneapolis: Fortress Press, 1994.
On the psychology of testimony:
Bond, Charles F., Jr., and Bella M. DePaulo. "Accuracy of Deception Judgments." Personality and Social Psychology Review 10 (2006): 214-234.
Innocence Project. "DNA Exonerations in the United States (1989-2020)." innocenceproject.org.
Koenig, Melissa A., Fabrice Clément, and Paul L. Harris. "Trust in Testimony: Children's Use of True and False Statements." Psychological Science 15 (2004): 694-698.
Loftus, Elizabeth F. Eyewitness Testimony. Cambridge, MA: Harvard University Press, 1979.
Loftus, Elizabeth F., and Jacqueline E. Pickrell. "The Formation of False Memories." Psychiatric Annals 25 (1995): 720-725.
Neisser, Ulric, and Nicole Harsch. "Phantom Flashbulbs: False Recollections of Hearing the News about Challenger." In Affect and Accuracy in Recall: Studies of "Flashbulb" Memories, edited by Eugene Winograd and Ulric Neisser, 9-31. Cambridge: Cambridge University Press, 1992.
Historical sources:
Augustine. Confessions. Translated by E. B. Pusey. Book 6, chapter 5.
Blunt, J. J. Undesigned Coincidences in the Writings Both of the Old and New Testament. London: John Murray, 1847.
Hume, David. Philosophical Essays Concerning Human Understanding. London: A. Millar, 1748. Later titled An Enquiry Concerning Human Understanding. Section 10.
Paley, William. Horae Paulinae. London, 1790.
Reid, Thomas. An Inquiry into the Human Mind, on the Principles of Common Sense. Edinburgh, 1764. Chapter 6, sections 20 and 24.
Biblical citations:
Deuteronomy 19:15; Proverbs 14:15; Proverbs 18:17; Luke 1:2; Luke 9:10; John 1:44; John 6:5; John 19:35; John 20:30-31; 1 Corinthians 15:3-8; 1 John 1:3; 1 John 4:1; 1 John 5:9. Scripture quotations are from the English Standard Version.
+C. A. J. Coady showed that reductionism cannot work. The track record we would need to trust testimony is itself known mostly through testimony, so testimony must be a basic source.
1. Global reductionism does fail.
Coady's point is a strong one, and it should be granted plainly. Ask yourself how you know that newspapers are usually right, or that encyclopedias are reliable, and most of your answer will come from what other people have told you. No one could check testimony as a whole against her own observations.
2. The evidential view never asked for a census of all testimony.
Weighing a report requires reasons to think that this kind of report, from this kind of source, on this kind of subject, tends to be right. And much of the background behind those reasons is gathered firsthand: what people are like, when they have motives to mislead, what a person standing on the corner can see (P2).
- A juror does not need statistics on all witnesses everywhere to see that a witness who stood ten feet away in daylight was in a better position than one who stood across a dark parking lot.
3. Testimony and observation check each other.
The sharpest reply along Coady's lines is that even these local reasons were mostly learned from others. Many were. But that is no circle, so long as testimony is never left to vouch for itself:
- The first weight comes from what a hearer sees for herself.
- Each report she accepts after that is weighed, when it comes in, against what she already has (D2).
- Where her own observations and many independent reports keep agreeing, each then adds weight to the others (P3).
That is how a traveler comes to trust a guidebook: the first three things it told her turned out to be right.
See also:
Philosophy / Epistemology: The Principle of Credulity
+Small children believe what they are told long before they can weigh any evidence, and they still come to know their own names, their parents, and the words of their language. So testimony cannot depend on the hearer's evidence.
1. Children's trust is real and good.
Reid noticed it in 1764: the disposition to believe what we are told "is unlimited in children, until they meet with instances of deceit and falsehood." And a good thing, too. A child who doubted everything her parents said would never learn to speak.
2. Children are gathering evidence from the start.
A small child is told that the stove is hot, that the dog is called Max, and that Grandma is coming on Saturday, and then she sees all three come true. Her trust is rewarded hundreds of times a week. Each time what she is told turns out to be so, she has a little more evidence, even before she could describe it.
- Children use that evidence early. In experiments, preschoolers already prefer to learn a new word from a speaker who has named familiar objects correctly over one who has named them wrongly.
- Justification comes in degrees. A child's belief can be reasonable to the degree her small store of evidence supports it, and that store grows fast.
3. The question about adults is different.
How a child first comes to trust her mother is a hard question for every theory of testimony, the evidential view included, and nobody should pretend otherwise. The question for a thinking adult is a different one: what makes it reasonable for her, now, to accept a report? That question is answered by what she has to go on.
See also:
Philosophy / Epistemology: Externalism
+Thomas Reid taught that God implanted in us a disposition to believe what we are told. Natural trust is how the Creator designed us to learn, and demanding evidence first refuses his gift.
1. Reid's observation about human nature is right.
We do trust by instinct, and Reid was right to see that as a gift. A world in which no one believed anyone would be a world without teaching, friendship, or history.
2. Reid himself saw the instinct corrected by experience.
In the very passage where he describes the disposition, Reid notes that it is unlimited in children only "until they meet with instances of deceit and falsehood." The instinct is meant to be trained by what we learn about who tells the truth.
A disposition to believe, in other words, is a starting habit rather than a reason. What a reason does is tell us when to keep the habit and when to check it.
3. Every source earns its weight the same way.
Reid's sharpest argument is one of parity, the demand that like cases be treated alike. Reason and the senses come from the same Maker, he argued, so a person who trusts one without proof has no ground for distrusting the other. He drew the same comparison between the senses and testimony.
That charge is fair against anyone who takes his own eyes on trust and then demands credentials from every speaker. It does not touch the evidential view, though, because the evidential view asks the same of every source. Perception and memory are weighed by what the believer is aware of, and testimony is weighed right alongside them.
- Think of a referee who calls fouls on one team and waves the other through. The cure is one rule for both teams, not a game with no fouls at all.
4. Scripture commends trust that keeps its eyes open."The simple believes everything, but the prudent gives thought to his steps" (Proverbs 14:15). And John tells believers, "Beloved, do not believe every spirit, but test the spirits to see whether they are from God, for many false prophets have gone out into the world" (1 John 4:1). God's gift of trust comes with the command to use judgment.
See also:
Philosophy / Epistemology: The Principle of Credulity
+Tyler Burge's acceptance principle: anyone may accept what is presented to him as true, if he understands it, unless he has stronger reasons not to. The entitlement comes from reason itself, not from evidence about speakers.
1. The principle captures something true.
Burge argued in 1993 that understandable speech comes from a being with reason, and that reason aims at truth. So, he concluded, a hearer has a default entitlement to believe, without needing to check the speaker first.
Part of that should be granted: no one has to investigate a particular speaker before believing anything he says. What remains to ask is whether the default rests on reasons the hearer has.
- A tourist who asks a stranger the way to the station does not run a background check before turning left.
2. The best reason for the default is that such speech is usually true.
Why should a hearer accept what a reasoning being presents as true? The best answer is a likelihood. Reasoning beings usually aim to say what is so, and so a reasoning being is far likelier to present something as true if it is true than if it is false. That likelihood is what makes such speech usually true, though it is not itself a tally of how often reports turn out right. A hearer who sees it has a reason, and the evidential view gladly counts it as part of her background knowledge.
Burge meant more than that, though. His entitlement holds even for a hearer who could never give the reason. Two terms from the theory of knowledge show what is at stake:
- Internalism holds that whether a belief is justified depends only on the believer's inner life: what she is aware of, remembers, and can reflect on. Externalism, its rival, holds that a belief can be justified by facts the believer has no awareness of, such as whether her faculties are in fact reliable.
- Picture two people whose inner lives are exactly alike, down to the last memory and impression, who hear the same stranger say the same thing. Internalism says they are equally reasonable in believing him. Externalism allows that one could be more reasonable than the other, because of facts neither of them can see.
That leaves Burge's principle at a fork:
- If the hearer can see the link between reason and truth, the entitlement is a reason she has, and it is weighed like any other.
- If she cannot, the entitlement rests on a fact outside her mind, and two hearers with the same inner life could differ in how reasonable their beliefs are. That is externalism, and it fails where externalism fails.
3. A default is not a blank check.
Once the principle's support is seen to be a likelihood, it can be weighed like any other. The weight an ordinary report carries then comes from the hearer's own background reasons, from what he knows of how reasoning beings speak, and that weight gives way when there is reason to doubt. What a general reason like that cannot do, by itself, is tell a hearer whether he is listening to a teacher or a con man.
- Picture a friendly stranger at a bus station who tells you the last bus has left, and that he can drive you for forty dollars. He is presenting something as true. The principle's own "unless" clause is where the real work gets done.
See also:
Philosophy / Epistemology: ExternalismPhilosophy / Epistemology: Internalism
+The assurance view: telling someone something is not handing her evidence. It is giving her your word, and treating a friend's word as mere data to be weighed insults the friend.
1. Assurance is real and matters morally.
Richard Moran and Edward Hinchman are right that telling is a personal act. A speaker who says "Trust me" takes on responsibility, and a hearer who believes her is relying on her. To treat a friend's promise like a reading on a gauge would miss something important about friendship.
2. An assurance gives reason because of who gives it.
A promise from an honest friend gives you reason to believe her, while the same words from a stranger running a con give you none. What makes the difference is what you know about the person: her honesty, her knowledge, her record. Assurance carries weight for you exactly to the degree that, on what you know of her, she would seldom give it if it were false.
- Consider a personal check. The signature makes it a promise, and a shopkeeper takes one from a customer she has known for years and turns down the same check from a stranger passing through. What she knows of the signer decides what the promise is worth to her.
3. A free act can still be evidence.
The strongest form of the assurance view, Moran's, starts from the fact that a statement is a free act. Smoke is tied to fire by nature, and it rises whether anyone means it to or not. A sentence has no tie like that. The only tie it has is the one its speaker gives it, by standing behind it. So the reason a telling gives, Moran argues, comes from the speaker's taking responsibility for it, not from its being evidence.
But free acts are evidence all the time. Suppose a careful friend chooses to tell you the ice on the pond is too thin to walk on. Her choice is far likelier if the ice is thin, because that is the kind of reason she picks her words for. That a statement is chosen is exactly why it matters who chose it, and what you know of her.
4. Faith in God is trust grounded in who he is.
Christian faith is personal trust, and it is not blind. Believers trust God's word because of who he has shown himself to be: in creation, in Israel's history, and above all in raising Jesus from the dead. Trust and evidence are not rivals. The evidence is what makes the trust wise.
See also:
Philosophy / Epistemology: Evidentialism
+Jennifer Lackey's creationist teacher disbelieves evolution yet teaches it faithfully, and her students come to know it. Testimony works through the statement, not the speaker's belief, and a hearer's reasons are not enough unless the statement is in fact reliable.
1. Lackey's case is a good one.
Lackey describes a fourth-grade teacher who, for religious reasons, does not accept evolution, but who teaches it carefully and accurately because it is her job. Her students come to know the theory from her, though she does not believe it herself. Many philosophers agree that the students know. If they do, the case refutes the older idea that testimony can only pass along knowledge the speaker already has.
2. The students' evidence never ran through the teacher's beliefs.
A report weighs for a hearer according to how much likelier it is, on what she knows, if true than if false (P1). Think about what the students actually know: that teachers teach the set curriculum, that the curriculum follows the textbooks, and that the textbooks report what biologists have found. On that evidence the lesson is far likelier if biologists have found it than if they have not, and the teacher's private doubts are no part of it.
- Suppose the students do assume she believes it. The assumption carries none of the weight. Told of her doubts, they would rightly go on believing, because what they were relying on was the curriculum and what stands behind it.
- A court interpreter may personally doubt a witness's story and still translate it word for word. The jury learns what the witness said from the interpreter.
3. Lackey's demand for positive reasons is the evidential view's own.
Lackey holds that a hearer needs positive reasons to accept a report, and on that she and the evidential view agree. They part over her second condition: that the statement must also be reliable in fact, whether or not the hearer has any way to tell. That condition is real. But it bears on whether the hearer knows, not on whether she is reasonable.
4. Reliability matters because it decides whether the hearer's premises are true.
Take a hearer with excellent reasons to trust a speaker who is, in fact, a practiced liar, and who happens to tell the truth this once. She is reasonable, but she does not know, because her belief is true only by luck. Philosophers call a case like this, where good reasoning lands on the truth by luck, a Gettier case.
Look at what her reasoning ran through: a falsehood, namely that this speaker is honest and would not have said it unless it were so. The internalist answer to such cases explains why that matters. A belief whose reasoning rests on a false premise lacks justification proper, the strict justification that knowledge requires.
- Justification proper adds only one condition to her being reasonable: that the premises her reasoning rests on are true.
- How reasonable she is stays fixed by her inner life, as internalism says. Picture two hearers with exactly the same evidence, one listening to an honest speaker and one to the practiced liar. They are equally reasonable, even though only the first of them knows.
- So the speaker's actual reliability is not a second source of justification beside the hearer's evidence. What it decides is whether the premises her evidence rests on are true.
See also:
Philosophy / Epistemology: The Gettier Problem
+David Hume showed that no testimony can ever establish a miracle, because the uniform experience against a miracle will always outweigh the word of any witness.
1. Hume's maxim sets a standard in principle.
Hume wrote that "no testimony is sufficient to establish a miracle, unless the testimony be of such a kind, that its falsehood would be more miraculous, than the fact, which it endeavours to establish." That is a standard. Whether any actual testimony has met it is a separate question, and Hume went on to argue that none has (D14).
2. Uniform experience cannot be assumed.
Hume rested the maxim on a uniform experience against miracles. Read as a claim about what people have actually observed, that begs the question, which means it assumes what it sets out to prove. Experience is uniform against miracles only if every report of one is false, and that is the very thing in dispute. The reports are part of human experience too, and they have to be weighed, not struck from the record in advance.
- Picture a census taker who announces that nobody in town keeps chickens, after leaving off his list every household that said it did.
3. Read in terms of probability, the maxim is fair.
A claim with long odds against it needs testimony that is very much likelier if true than if false, and every evidentialist accepts that.
- George Campbell, a Scottish minister, answered Hume in 1762. Richard Price, in a dissertation published five years later, put the answer in terms of probability: how improbable the event was beforehand is only half the question, and the other half is how improbable it is that this testimony would be given if the event had not happened.
- Many writers on probability since, not all of them believers, have read the maxim the same way.
Whether any testimony meets the maxim, then, is not settled by the maxim. It depends on the testimony.
4. Improbable events are established by testimony every day.
Price's lottery shows that a report can establish an event whose improbability beforehand "exceeds all conception," when the report is strong enough.
- It does not show that a single miracle report is as strong as a newspaper's. Think about how a paper could get the result wrong. A misprint could land on any of millions of wrong numbers, so it is most unlikely to land by mistake on the one particular number printed. A false report that some wonder happened is not scattered across millions of possibilities in that way.
- What the lottery does show is narrower, and it is enough: a low starting probability never settles the matter by itself.
Where the lottery comparison does come back is when several witnesses agree in detail. Then each false report has to land on the same place, the same people and the same words, and independent errors seldom land together (P3). Babbage showed that enough truly independent witnesses can, in principle, outweigh any starting improbability (C2). Real witnesses are seldom as independent as his arithmetic assumes, though, and something they share can put the same story in all their mouths (D13).
5. The real question is historical.
Whether the testimony to the resurrection is strong enough is a question about that testimony: how many witnesses there were, how independent they were, what position they were in to know, and what they had to gain or lose. It cannot be answered by a rule laid down before the evidence is examined.
See also:
Philosophy / Epistemology: EvidentialismChristian Evidences / Resurrection of Jesus: Maximal Data Method (McGrew)
+Eyewitnesses are unreliable. Mistaken eyewitness identification played a part in about 69 percent of the first 375 wrongful convictions in the United States overturned by DNA evidence.
1. The research is sobering.
The figure comes from the Innocence Project, a legal organization that works to free the wrongly convicted, and it is real: in roughly seven of every ten of those cases, a witness picked the wrong person.
The strongest form of the objection reaches past lineups. Psychologists such as Elizabeth Loftus have shown that memory of events, not only of faces, is rebuilt each time it is recalled, and that it can be bent by stress, by time, and by leading questions. No honest account of testimony can ignore it.
2. The research shows when witnesses fail.
The errors cluster in particular conditions: picking a stranger out of a lineup after a glimpse of a few seconds, under stress, with a weapon drawing the eye, or after suggestive questioning. A hearer who knows those conditions held should give the report less weight, because such a report is not much likelier if true than if false.
- Recognizing a stranger glimpsed in a dark parking lot is a different act from recognizing a friend you have known for years. A police lineup tests the first kind. Nobody needs a lineup to pick out his own brother.
3. Details bend more than the core.
Distortion falls hardest on details: a speed, a color, who stood where.
- Psychologists asked students, the morning after the space shuttle Challenger was lost in 1986, how they had heard the news, and then asked them again two and a half years later. Many gave confident new accounts that did not match what they had written down. But what they misremembered was how they heard, not that the shuttle was lost.
- False memories of whole events have been produced in experiments, by sustained suggestion, and a hearer can ask whether a witness was exposed to it.
4. The research itself is weighed testimony.
What we know of these studies we know from reports by researchers, court records, and the Innocence Project. We trust those reports because of what we know of the care taken in them, and that is the evidential view at work.
+Studies show people can barely tell liars from truth-tellers, scoring about 54 percent where guessing would score 50. So no hearer can really weigh whether a speaker is sincere.
1. Demeanor is a poor guide.
Charles Bond and Bella DePaulo's 2006 review of over two hundred studies found that people sort lies from truths correctly only about 54 percent of the time, barely better than flipping a coin. Nervous glances and fidgeting say little about honesty.
2. Most evidence about testimony is not demeanor.
A careful hearer asks other questions, and none of them depends on reading a face (P2, P3):
- Was the speaker in a position to know?
- Did he have a motive to mislead?
- Can the report be checked?
- Does it fit the other evidence?
- Do others report the same thing independently?
A detective who cannot tell when a suspect is lying still checks his alibi against the security camera.
+The Gospel witnesses are not independent. Matthew and Luke copied Mark, and the early church shaped the stories it passed on, so their agreement shows dependence, not separate testimony.
1. Some dependence is real.
Most New Testament scholars hold that Matthew and Luke used Mark's Gospel as a source. Most also hold, though a significant minority doubt it, that the two shared a second, lost collection of Jesus' sayings that scholars call Q (from the German Quelle, "source"). Where Matthew and Luke follow Mark or Q, their agreement counts as one source, not two (P3).
2. A shared oral tradition is a common source too.
The strongest form of the objection reaches further back. Before anything was written, the stories were told and retold in Christian communities. Form critics, scholars who study how the Gospel stories took shape while they were passed on aloud, argued that those communities shaped them to fit their preaching and worship. Agreement that traces only to a shared community tradition counts once, not once per Gospel, and that should be granted.
But a shared tradition is only as good as its tie to the people who were there, and while the stories circulated, those people were still alive and known to the churches.
- Paul, writing some twenty to twenty-five years after the crucifixion, passes on a list of witnesses that he says he had received. Many historians, skeptical ones among them, date his receiving it to within a few years of the events.
- The list names Peter and James, and Paul adds that most of the more than five hundred who saw the risen Jesus were still living (1 Corinthians 15:6).
- Think of a family story told every Thanksgiving. While Grandpa, who was there, still sits at the table, the story drifts far less than it will after he is gone.
Whether the tradition stayed tied to such witnesses or drifted from them is a historical question, argued on the evidence.
3. Independence comes in degrees.
Dependence in one place does not mean dependence everywhere.
- Matthew and Luke each include material found nowhere else.
- John's Gospel tells much of the story in its own way.
- Paul's list of the resurrection appearances (1 Corinthians 15:3-8) comes from a separate early tradition.
4. Undesigned coincidences show independence where it exists.
When one account explains a puzzle in another without seeming to notice, copying is an unlikely explanation, because a copier would have smoothed the puzzle over (C2). William Paley in 1790 and J. J. Blunt in 1847 collected such fits, and Lydia McGrew extended the case in 2017. Scholars disagree about how much these coincidences show, and that is the right place for the argument to happen: in the details.
See also:
Christian Evidences / New Testament Criticism: Undesigned Coincidences as Evidence of Gospel ReliabilityChristian Evidences / New Testament Criticism: Convergence of Independent New Testament TraditionsChristian Evidences / Resurrection of Jesus: Contra Legend Hypothesis
+No one can put a number on a witness. Saying a report is "twenty times likelier if true" is pretend precision, so measuring testimony by probability is a sham.
1. Exact numbers are rarely available.
That is true, and anyone who claims to know the likelihood ratio of a first-century witness to three decimal places is fooling himself.
2. Comparisons do not need exact numbers.
"Far likelier," "about as likely," and "at least ten times likelier" are judgments people make well, and they are enough to reason with. A doctor who knows a test is "much more often positive in people who have the disease" can reason soundly without the exact figure.
3. The numbers teach the structure.
Even rough numbers show things that matter: that a weak report cannot overcome long odds, that a strong one can, that independent reports multiply, and that a witness with no weight of his own adds nothing (P1, P3). Those lessons hold whatever the exact figures turn out to be.
See also:
Philosophy / Epistemology: Evidentialism
+Faith rests on God's testimony, not on human witnesses. Scripture says, "If we receive the testimony of men, the testimony of God is greater," so weighing apostles like a jury puts human judgment above God's word.
1. God's testimony is greater.
John writes, "If we receive the testimony of men, the testimony of God is greater, for this is the testimony of God that he has borne concerning his Son" (1 John 5:9). Christians agree that God's word outweighs any human word. Many Christians also hold that the Holy Spirit bears an inner witness that gives a believer assurance.
2. The verse assumes human testimony is rightly received.
John's argument runs from the lesser to the greater. It takes for granted that "we receive the testimony of men," and receive it rightly, and then argues that God's testimony deserves more. Far from setting human testimony aside, it builds on it.
3. John wrote testimony so that readers would believe."Now Jesus did many other signs in the presence of the disciples, which are not written in this book; but these are written so that you may believe that Jesus is the Christ, the Son of God, and that by believing you may have life in his name" (John 20:30-31). The apostles' testimony is how God's testimony about his Son reaches those who were not there. Whatever the Spirit's inner witness does for the believer, it does not make that testimony idle. And it is the testimony that a believer can lay before anyone who asks.
+Real witnesses are rarely independent in the sense Babbage's multiplication needs. Witnesses who never met can still share one cause of error, and no number of them can push past the ceiling that shared cause sets.
1. A shared cause of error puts a ceiling on agreement.
The statistician William Kruskal warned in 1988 against assuming independence casually, and the philosopher Arif Ahmed argued in 2015 that Babbage's witnesses, though they never colluded, are not independent in the sense his arithmetic needs. The point should be granted in full.
Multiplication requires that, if the claim is false, each witness goes wrong on his own (P3). Now suppose there is one thing which, if it happened, would lead all of them astray together, into the very story they tell. Then, if the claim is false, their agreement can never be less likely than the chance that this one thing happened. Their combined weight has a ceiling, and adding witnesses cannot raise it.
- Think of a magic show. A thousand people watch a woman sawn in half, and none of them compared notes. Their agreement is worth about what one viewer's report is worth, because the trick that fooled one fooled them all.
2. Uncertainty can tie witnesses together as well.
There is a quieter way for witnesses who never met to lose their independence.
- Suppose you are reading glowing reviews of a gadget on a website you know nothing about. If the gadget is junk, each glowing review is a hint that fakes get through on that site, and that makes the next one easier to explain.
- Reports of wonders work the same way. Suppose a hearer does not know how readily people in some setting report wonders that never happened, and suppose that this time nothing happened. Then one such report is a sign that people there report them readily, and that makes a second report less surprising.
3. The ceiling is set case by case.
How high the ceiling sits depends on how likely a shared error is, and that is a question about particular witnesses: what could have misled all of them at once, and how probable it was.
- A magic show is built to fool a whole room from one angle. Witnesses who saw from different places, at different times, under different conditions, with different hopes and fears, are much harder to fool all at once, and the ceiling over their agreement sits much higher.
- So the objection does not show that agreement among witnesses is weak. What it shows is that its strength cannot be read off a headcount.
4. The evidential view already demands that independence be shown.
The probabilistic case for testimony has never rested on counting heads. Independence must be shown, not assumed (C2), and it comes in degrees: partial dependence lowers the multiplier without reducing it to one. The encyclopedia entry that sets out the multiplication rule warns, in the same breath, against assuming independence casually.
5. Applied to the resurrection, the shared cause has to be named.
Here the objection turns into a set of definite proposals: a group hallucination, a legend grown up in one community, a single source copied by the rest. Each is a historical hypothesis, and each is weighed on two counts:
- How probable it is to begin with
- How well it would produce the testimony we actually have, with its timing, its variety and its witnesses
The ceiling on the testimony's combined weight is set by those hypotheses, taken together. That is where the debate belongs.
See also:
Philosophy / Epistemology: CoherentismChristian Evidences / Resurrection of Jesus: Contra Hallucination HypothesisChristian Evidences / Resurrection of Jesus: Contra Legend Hypothesis
+Every religion has witnesses to its miracles, and Hume pointed to well-attested healings at a Paris tomb. If agreeing witnesses can establish the resurrection, they establish rival miracles too, and since the religions contradict one another, the method proves too much.
1. Rival miracle reports are a fair test of any method.
Hume pressed this objection hard. He argued that the miracles of rival religions cancel one another out: each is offered as proof of its own religion, so each counts as evidence against the rest.
His sharpest example of strong testimony was recent in his day. In the years after 1727, crowds gathered at the grave of the Abbé Pâris in the Paris cemetery of Saint-Médard. Pâris had been a deacon and a Jansenist, a member of a strict reform movement within the Catholic Church, and the crowds there reported healings attested by many witnesses, some of them people of standing. Hume asked where one could find so many circumstances agreeing to corroborate a single fact. A method that could not set such reports aside, when they deserve it, would be broken.
2. One standard can give different verdicts.
The evidential view applies a single test to every report: how much likelier it is if true than if false, given everything the hearer knows. One test does not mean one verdict.
Hume's own observations count here, and they should be granted:
- Wonder is pleasant to believe and to repeat.
- Religious zeal can make a man say more than he saw.
- A report that serves a cherished cause deserves a closer look.
Each of these raises the chance that a miracle report would be given even if nothing happened, and the test applies them to Christian reports as fully as to any other. How much each raises it depends on the case. Zeal swells a report that confirms what its witnesses already believed far more easily than one that overturns what they expected and costs them dearly.
- Imagine a court that convicts one defendant on the word of two witnesses and acquits another on the word of two others. It is not being inconsistent if the first pair watched the crime in daylight from across the street and the second pair are brothers who heard about it the next day.
3. Each case is weighed on its witnesses.
For every reported miracle the questions are the same:
- Were the witnesses in a position to know?
- Could one cause, such as the fervor of a crowd at a shrine, have produced all their reports (D13)?
- Is there a natural explanation of what they saw?
- Did they report independently?
- What did saying it cost them?
Many reports fail these tests badly, some Christian ones among them, and the evidential view sets them aside for that reason. The objection succeeds only if some rival miracle, offered as a sign confirming a teaching that contradicts Christianity, has evidence as strong as the resurrection's on the same tests. That is a comparison of evidence, made case by case, and it cannot be settled by noting that rival reports exist.
4. Hume's own example fails the tests.
The Saint-Médard healings were not rival miracles at all, since the Jansenists were Catholic Christians. What Hume offered them as was testimony as strong as any, which his readers would reject all the same. Critics in his own century examined the reports closely and found them much weaker than he made them sound:
- Crowds of the sick came to the tomb for several years, until the king had the cemetery closed in 1732, and the reported cures were few beside them. Among so many who come hoping, some will recover whatever the cause. It is like a lottery stand that posts photographs of its winners and none of the thousands who lost.
- Many of the cures came slowly, after repeated visits or days of prayer, or only in part, and some were of complaints that time and strong feeling can relieve.
- The reports were gathered and published by one party in a bitter church quarrel, as proof that its side was right.
Hume himself did not answer the witnesses by weighing them. All he had to set against so many, he wrote, was that what they reported was impossible. The evidential view grants no such premise (D7), and it still sets these reports aside, on its own tests.
See also:
Philosophy / Logic: Common Fallacies: Special PleadingChristian Evidences / Resurrection of Jesus: Maximal Data Method (McGrew)
+Much of what we believe comes from experts whose reasons we could never follow. John Hardwig argued that a layperson rationally believes such things without having the evidence for them, so justification cannot depend on the hearer's evidence.
1. Hardwig is right about our dependence.
A juror cannot rerun the DNA lab's analysis, and most of us could not check a single result in a physics journal. Hardwig argued in 1985 that in such cases the rational thing is to defer, and that a layperson who insisted on weighing the expert's evidence herself would be less rational, not more. That should be granted plainly. No one could know much of modern science, medicine or history any other way.
2. The layperson weighs the report, not the expert's evidence.
What she cannot follow is the expert's reasoning. What she can weigh is the expert's report, and the evidential view's question is about that report: how much likelier is it, on what she knows, if true than if false? On that she has a good deal to go on:
- The expert's training and record
- Whether he gains anything by the answer
- Whether his field's predictions keep coming true
- Whether other experts, working separately, say the same
Hardwig himself allowed that the layperson has a good reason of her own: that the expert has good reasons. He declined to call it evidence for the claim, but it is exactly the kind of reason the evidential view weighs.
- Take a homeowner whose electrician says the wiring behind her walls is unsafe. She cannot read a circuit, and she knows he would earn the repair job, so she calls a second electrician from another company, who finds the same fault. She still could not explain the problem, but she now has strong evidence of it, because two independent electricians would seldom agree on a fault that was not there.
3. Her justification is her own.
Two homeowners whose evidence is exactly the same are equally reasonable in what they believe, whatever is in the electricians' notes. That is exactly what internalism says, so here internalism is holding, not failing.
Her justification also rests on different evidence from the expert's, and it is usually lower: she is rightly less sure than the electrician who traced the fault. Whether she knows depends, as always, on whether what her reasoning rests on is true (D6).
4. Where a hearer can weigh nothing, the report weighs little.
The hard case is testimony about a matter the hearer cannot evaluate at all, from a source she knows nothing about. The evidential view says such a report carries little weight for her, and that is the right verdict.
- Imagine an unsigned message, forwarded through people you have never met, announcing that a common medicine causes a rare disease. Believing it on that alone would not be trust. It would be luck, good or bad.
See also:
Philosophy / Epistemology: InternalismPhilosophy / Epistemology: Externalism