(●)Alvin Plantinga holds that warrant, the quality that turns true belief into knowledge, comes from faculties working as they were designed to work.
+Every Saturday morning Ellen bakes bread, and she weighs the flour on the kitchen scale her mother gave her. The scale is old but sound. Its spring does what it was built to do, in the kind of place it was built for: a kitchen counter, on solid ground, under ordinary gravity. So when it reads five hundred grams, Ellen believes there are five hundred grams of flour in the bowl, and she is right.
Now put the same scale in the galley of a ship rolling through heavy seas. Nothing is broken, and the spring works perfectly. But the scale was never designed for a floor that heaves, and its readings stop being trustworthy. Or leave the kitchen as it is and swap in a cheap scale with a cracked spring. This time the place is fine, and it is the scale that is not working the way it should.
Alvin Plantinga, one of the most important Christian philosophers of the twentieth century, built a theory of knowledge on exactly that picture.
- Warrant is Plantinga's name for the quality that, when there is enough of it, turns a true belief into knowledge. Picture Martha glancing at the hall clock, which stopped hours ago and by sheer luck shows the right time. Her belief about the time is true, but it lacks warrant, and so she does not know what time it is.
- Plantinga keeps the word justification for something narrower: doing one's intellectual duty, believing responsibly. He grants that justification in this sense is an inner matter. What he denies is that it is what makes knowledge, and he argues that the tradition went wrong by treating it as if it were.
- His own answer is called proper functionalism. On this view a belief has warrant when it is produced by cognitive faculties (the mind's belief-forming powers, like perception, memory, and reasoning) that are working properly, the way they were designed to work.
In Warrant and Proper Function (1993), Plantinga spells out four conditions. A belief has warrant for a person when:
- the faculties that produce it are functioning properly, with no malfunction or damage;
- they are working in a cognitive environment (the surroundings a faculty operates in) sufficiently like the one they were designed for, as Ellen's scale is on the kitchen counter and not on the ship;
- the part of the design plan that governs them is aimed at truth, not at comfort or survival or something else;
• A design plan is the set of specifications for how a thing is supposed to work. A heart's design plan calls for pumping blood. Plantinga notes that some of our faculties may be aimed at other goods, like the optimism that helps a sick person recover, and those do not produce warrant.
- and there is a high objective probability that a belief formed that way, in that environment, is true.
• Objective probability is a fact about the world, not about how confident anyone feels. A fair die has a one-in-six chance of landing on six whether or not the person rolling it believes so.
Warrant also comes in degrees. The more firmly a person holds a belief that meets these conditions, the more warrant it has, and enough warrant, added to true belief, makes knowledge.
(●)Plantinga built proper functionalism to fix what he saw as the failures of both internalism and reliabilism.
+Plantinga did not arrive at proper functionalism by accident. In Warrant: The Current Debate (1993) he surveyed the leading theories of his day and argued that each one fails, and proper functionalism was built to succeed where they did not.
- Against internalism.Internalism says that what justifies a belief lies entirely within the believer's own mind: their experiences, memories, and reasoning. Plantinga traced it to the idea of duty in René Descartes and John Locke, seventeenth-century philosophers who taught that we must govern our believing responsibly. His complaint was that doing one's duty in believing is not enough for knowledge. A person can be perfectly dutiful and still hold a belief that is a product of brain damage.
- Against reliabilism.Reliabilism says a belief is justified when it is produced by a reliable process, one that turns out true beliefs most of the time, the way a good thermometer turns out accurate readings. Plantinga answered it with a famous counterexample. Imagine a brain lesion that, among its many effects, causes its victim to believe that he has a brain lesion. The process is perfectly reliable, since everyone with that lesion believes they have one and is right, and yet no one would say the belief is knowledge. What is missing, Plantinga argued, is proper function. The lesion is a malfunction, however reliable it happens to be.
- A refinement for luck. Some true beliefs are right only by luck, and in later work (most fully in Warranted Christian Belief, 2000) Plantinga added a refinement about the immediate surroundings to handle them. Our faculties may be working in the right kind of world overall, which he calls the maxi-environment, while a particular belief is formed in a deceptive local situation, the mini-environment.
• Picture a county where the locals have put up dozens of fake barn fronts, painted boards with nothing behind them, and a traveler driving through happens to look at the one real barn. He believes he sees a barn, and he is right, but only by luck.
• Cases like this are called Gettier cases, after the philosopher Edmund Gettier, who showed in 1963 that a belief can be true and well supported and still fall short of knowledge by luck. The puzzle of how to answer them is known as the Gettier problem.
The theory also carries a theological payoff, and that is a large part of its appeal to Christians.
- In the final chapters of Warrant and Proper Function, Plantinga argues that naturalism, the view that nature is all there is, with no God or anything beyond the physical world, cannot make sense of proper function at all. Talk of faculties working "as designed" fits most naturally with a designer, in the way talk of a watch keeping time "as it should" points naturally to a watchmaker.
- His evolutionary argument against naturalism goes a step further. A naturalist who believes our faculties came from unguided evolution, it argues, has a reason to doubt that those faculties are reliable, that is, that they mostly produce true beliefs.
- In Warranted Christian Belief (2000), Plantinga applies the theory to faith itself. He calls his account the Aquinas/Calvin model, after two theologians who both taught that human beings have a natural awareness of God. On this model God has given human beings a sensus divinitatis, Latin for "sense of the divine," a faculty that produces belief in God in the right circumstances. A person looking up at the night sky, or feeling the weight of guilt, may simply find herself believing that God is there, without reasoning her way to it. And the Holy Spirit works in believers to produce belief in the gospel, an inward prompting Plantinga calls, borrowing Aquinas's phrase, the internal instigation of the Holy Spirit.
- If that is how things are, belief in God is warranted without any argument, because it comes from faculties working as God designed them.
- From this Plantinga draws a striking conclusion: whether Christian belief is warranted depends on whether it is true. If Christianity is true, then Christian belief very likely has warrant. So a critic cannot show that Christian belief lacks warrant without first showing that it is false.
Many Christian philosophers, Michael Bergmann, Tyler McNabb, and Kelly James Clark among them, have defended and developed this picture. It is a serious and influential view, and the Christian intellectual world owes Plantinga a great deal.
(●)Proper functionalism, like every externalism, ends in epistemic nihilism.
+Suppose one Saturday Ellen's loaf comes out flat, and she starts to wonder about the scale. That question she can settle. She sets a sealed one-kilogram bag of sugar on the scale and reads the dial. But suppose the question turns on Ellen herself: "Is my belief that the scale is accurate in good order?" On Plantinga's theory, the answer depends on whether her faculties are working properly, in the right environment, under a design plan aimed at truth. None of those is something she can set on the counter and read.
That is what makes proper functionalism a form of externalism: the view that what gives a belief its positive standing can lie outside the believer's mind, in facts they may have no way of seeing. Whether my faculties are working properly, whether my environment is the one they were designed for, and whether the relevant part of the design plan is aimed at truth are not things I can find by looking inward. Plantinga says so plainly, and it is the point of the theory.
In Internalism and Epistemology: The Architecture of Reason (2007), the Christian philosophers Timothy and Lydia McGrew argue that this is exactly where externalism fails. (Epistemic simply means having to do with knowledge and justification.) Their central thesis is that externalism leads past skepticism, the view that we know little or nothing, all the way to epistemic nihilism, the denial of the very possibility of justification. By justification the McGrews do not mean Plantinga's dutiful believing. They mean the standing a belief has when it rests on evidence the person has in view and can use, evidence that makes the belief probable.
- Their main weapon is the metaregress argument, set out in a chapter titled "Internalism, Externalism, and the Metaregress." A regress is a chain of questions in which every answer raises the same question again, like a child asking "but why?" after every answer. Meta means "about," and a metaregress climbs through beliefs about beliefs: is this belief warranted, is my belief that it is warranted itself warranted, and so on up.
- They aim it at one feature in particular. Epistemic principles are the principles that say what counts as good evidence for what: for instance, that if nine of the ten tickets in a hat are blue, a ticket drawn at random is probably blue. The McGrews argue that the metaregress is fatal to any view on which those principles are contingent.
• A contingent truth is one that happens to be true but could have been false, the way "Paris is the capital of France" could have gone otherwise. Philosophers say it is true in some possible worlds (complete ways things could have been) and false in others.
- On proper functionalism, epistemic principles are contingent, because what makes a belief warranted depends on which design plan our faculties happen to have and which environment we happen to be in.
- Their own internalism escapes, they argue, because its epistemic principles are analytic and knowable a priori: true by the meaning of their terms, and knowable by reflection alone. "Every bachelor is unmarried" is like that. Nobody needs to survey bachelors to know it.
Tyler McNabb has replied to the McGrews on Plantinga's side (2018), defending proper functionalism at the metalevel, the level of our beliefs about whether our beliefs are warranted. In its strongest form, a defense at that level is a serious reply, and (D1) answers it at full strength.
The argument is deductive. Its 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. Step by step, the metaregress runs from (P1) to (C2).
(P1)On proper functionalism, warrant depends on facts beyond our reach.
+Plantinga's four conditions are facts about our equipment and our world. Call a fact beyond our reach when there is no way to get at it from where we stand: we cannot find it by looking inward, and we cannot reason our way to it by standards we can see to be sound.
(1) Each of Plantinga's four conditions is external.
- Whether a faculty is functioning properly is a fact about the faculty, not about how things seem. A person whose memory is quietly failing feels just as confident as a person whose memory is sound.
- Whether the environment matches the design plan is a fact about the world. Ellen's scale cannot tell whether it sits on a kitchen counter or on a ship.
- The last two conditions are no different. Whether the design plan is aimed at truth is a fact about how the faculty was made, and whether beliefs formed that way are objectively probable is a fact about how it performs. Neither shows up in experience.
- Inner states do have a part in Plantinga's account: the experiences that prompt a belief, and how firmly it is held. But whether those inner states add up to warrant is settled by the four conditions.
(2) Plantinga accepts this as the heart of the theory.
- He describes warrant as an externalist notion. In his view the internalist tradition went wrong by trying to keep knowledge within reach of reflection, meaning what a person can find out just by thinking it over.
- That is why he says the question whether Christian belief is warranted cannot be settled apart from the question whether it is true. The warrant depends on facts about God and the human mind that the believer cannot check by looking inward.
(3) Two believers with the same inner life can differ in warrant.
- Picture two people with exactly the same experiences, memories, and reasoning. One has faculties that work as designed. The other's faculties were shaped by a malfunction that happens to give the same results.
- On proper functionalism, one has warrant and the other does not, and neither could ever tell which one they are.
- Think of two kitchen scales, both showing five hundred grams. One is sound, and the other has a cracked spring that happens to stop on the right number. You could study the dials all day and never learn which is which.
(4) Even the standards of good reasoning are contingent.
- On proper functionalism, what counts as good evidence for what is itself part of the design plan. On Plantinga's account, creatures designed differently could properly respond to the very same experience with different beliefs.
- That closes the second road as well. We might hope to reason our way to the four conditions, but whether our reasoning is good reasoning is one more fact about which design plan we have, and reflection does not reveal it.
- Think of a tape measure whose markings are right only if the factory printed them to the right standard. You can measure anything with it, but measuring again with the same tape will never tell you whether the factory got it right.
Therefore, on proper functionalism, warrant depends on facts we can neither see by reflection nor reach by reasoning we can check.
(P2)If warrant depends on facts beyond our reach, we can have no reason to think any of our beliefs warranted.
+This is the McGrews' metaregress, and it turns on what it means to have a reason. A reason here is something we have in view and can weigh, such as an experience, a memory, or an argument, whose bearing on the question we can see for ourselves.
(1) Having a reason is more than having a well-made belief.
- A belief can be produced the right way and still have nothing behind it that we can weigh. A belief like that is not yet a reason we have.
- Think of a student who copies the right answer from a classmate she has no grounds to trust. She has the right answer, but she does not have a reason for it.
- The proper functionalist uses "reason" more broadly, for any belief formed the right way. This is more than a quarrel about words, because only a reason in the narrower sense can answer someone who asks, "Should I trust this belief?" The two senses meet head on in (D1), and the argument stands or falls there.
(2) Asking the question produces a new belief.
- Suppose you wonder whether your belief that the bread needs five hundred grams of flour is warranted. To answer, you have to form a second belief: "My faculties were working properly, in the right environment, when I formed that belief."
- On proper functionalism, that second belief is warranted only if the faculties that produced it were working properly in the right environment. And that, once again, is a fact you cannot see.
(3) The question climbs without end.
- Ask whether the second belief is warranted, and you need a third belief, about the faculties that produced the second. Then you need a fourth, about the third.
- At no level do you reach anything you can check. Each level simply hands the question up to the next.
- The climb could stop in one place: at a principle you can know by reflection alone, the way you know that two plus two is four. But if the facts that make for warrant are beyond our reach, so are the standards for reasoning about them (P1), and that one exit is closed.
- Think of a certificate saying a scale has been inspected. It is only as good as the inspector, whose license is only as good as the agency that issued it, whose authority rests on yet another office. If every office is behind a locked door, the chain never ends in anything you can read.
(4) Checking faculties with themselves proves nothing.
- The obvious move is to test your faculties by their track record. The trouble is that the only way to check the track record is to use the very faculties being tested. Philosophers call this epistemic circularity: a faculty vouching for itself.
- Imagine a man who checks whether his watch is accurate by looking at his watch. Whatever the answer, the check could not have come out any other way, and a test that cannot fail tells you nothing.
- The McGrews argue, in a chapter titled "What's Wrong with Epistemic Circularity," that this circularity is vicious, meaning it leaves the check with no power to support the belief at all. Some externalists hold that it is often harmless, and (D9) answers that reply on its own.
Therefore, if warrant depends on facts beyond our reach, we can have no reason to think any of our beliefs warranted.
(C1)Therefore, if proper functionalism is true, we can have no reason to think any of our beliefs warranted.
+This follows from (P1) and (P2) by a valid form called a hypothetical syllogism: if A then B; if B then C; so if A then C. If it rains, the game is canceled; if the game is canceled, we go to the movies; so if it rains, we go to the movies.
(1) This is what the McGrews call epistemic nihilism.
- On proper functionalism, a person might have warranted beliefs, even a great many of them. What she could never have is a reason to think so, about any belief at all, including her belief in proper functionalism itself.
- The McGrews argue that this, followed through, is the denial of the very possibility of justification in their sense: a standing that rests on evidence a person has in view and can use. That is why they call the result nihilism and not just skepticism.
- Think of a bank account that may hold a fortune in your name, at a bank that will never show you a statement or let you draw a cent. You might be rich. But you would never have any reason to think so, and the money would never be yours to use.
(2) The point is not that warrant requires knowing you have it.
- C1 does not say that a person must know they have warrant in order to have it. What it says is that, on proper functionalism, no one could ever have a reason to think so.
- A man may own a winning lottery ticket without knowing it. But if a theory says that nobody could ever check any ticket, something has gone wrong with the theory.
Therefore, proper functionalism leaves every believer without any reason to think that a single one of their beliefs is warranted.
(P3)We can have reason to think some of our beliefs warranted.
+This premise is the ordinary conviction that we can sometimes tell that we know. "Reason" means here what it meant in (P2): something we have in view and can weigh.
(1) Some cases are as clear as anything can be.
- You know that two plus two is four, and you can see that it is true, not merely believe it. When you stub your toe, you know you are in pain, and you have every reason to think that belief is in good order.
- To deny that anyone can ever have a reason to think any belief warranted is to deny it even here.
(2) P3 is far weaker than internalism.
- Internalism is the claim that justification depends only on what is in the believer's own mind. P3 says only that sometimes we can see that a belief is in good order, and it leaves room for knowledge to have outside conditions too, such as truth and freedom from luck.
- So P3 does not assume the McGrews' theory. A view of knowledge with outside conditions can accept it, provided some standards of good reasoning can be seen to be sound by reflection. The only view P3 rules out is one on which nothing we can see ever bears on whether a belief is warranted.
- Think of a doctor reading your blood test. She cannot see every cell in your body, but she can see what the test shows. That gives her reasons about some things without access to everything.
(3) Plantinga's own work assumes it.
- Plantinga wrote a trilogy of books on warrant arguing that proper functionalism is true. He offered reasons, answered objections, and expected his readers to weigh them and see whether they were good.
- Now think about a reader who is persuaded. On proper functionalism she could have a warranted belief that the theory is true, but she could never have a reason to think so, because whether her being persuaded was proper function or malfunction would be one more fact beyond her reach.
- It is like a referee who rules that no call in the game can ever be checked on replay, and then asks the players to review the tape of that very ruling.
(4) Christian witness assumes it.
- Peter tells believers to be "always being prepared to make a defense to anyone who asks you for a reason for the hope that is in you" (1 Peter 3:15). Giving a reason assumes the giver can tell that they have one.
- In the synagogue at Thessalonica, Paul reasoned from the Scriptures on three Sabbath days (Acts 17:2), "explaining and proving that it was necessary for the Christ to suffer and to rise from the dead" (Acts 17:3). In Berea his hearers went further, "examining the Scriptures daily to see if these things were so" (Acts 17:11).
- None of these texts settles a theory of knowledge, and proper functionalists read them comfortably. What they do show is that Scripture treats reasons as things a person can hold up, examine, and see to be good.
Therefore, we can have reason to think some of our beliefs warranted.
(C2)Therefore, proper functionalism is false.
+This follows from (C1) and (P3) by modus tollens: if A, then B; not B; therefore, not A. If proper functionalism were true, we could have no reason to think any of our beliefs warranted (C1). But we can (P3). So proper functionalism is not true.
(1) What the conclusion does not deny.
- It does not deny that our faculties were designed by God, or that they work properly most of the time. Christians believe both.
- Nor does it deny that proper function matters to knowledge. A belief produced by a malfunctioning brain is in poor shape on any theory.
- It denies only that proper function is what warrant consists in. What warrant does consist in is a further question, and the syllogism leaves it open. Its reasoning does narrow the field, though: by the same steps (P1, P2), any account on which the standards of good reasoning are contingent meets the same regress.
(2) Other objections point the same way.
- In a 1993 symposium on Plantinga's work, Richard Feldman and Ernest Sosa each argued that proper function is not the right account of warrant, and Plantinga replied in the same issue.
- Sosa also pressed a case adapted from Donald Davidson's Swampman (1987). Lightning strikes a dead tree in a swamp, and by a freak accident the molecules rearrange into an exact duplicate of a human being. Swampman appears to have no design plan, and yet he seems to know that he is standing in a swamp. Plantinga answered the case in 1993, doubting that such a creature is really possible and suggesting that, if it were, we would still credit its organs with functions, and so with a design plan. Whether his answer succeeds is still debated.
(3) The McGrews' own answer is internalism.
- The McGrews answer the open question this way: justification is a matter of what a person's evidence makes probable. On that view the principles of good evidence are knowable a priori, so the regress stops where a person sees what their evidence supports.
- Think of a juror. She cannot see into the defendant's past, but she can see the fingerprints, the receipt, and the witness in front of her, and she can tell what they make probable. Whether she is justified does not wait on facts locked away from her.
- Internalism is a further thesis with its own argument, and (C2) does not depend on it. A reader could reject proper functionalism and still be unpersuaded by the McGrews' alternative.
Therefore, if (P1), (P2), and (P3) hold, proper functionalism is false.
See also:
Philosophy / Epistemology: InternalismPhilosophy / Epistemology: Classical FoundationalismPhilosophy / Epistemology: Dwindling ProbabilitiesPhilosophical Theology / Prolegomena: Meta-Apologetics: Reformed
Plantinga's view:
Plantinga, Alvin. Warrant: The Current Debate. New York: Oxford University Press, 1993.
Plantinga, Alvin. Warrant and Proper Function. New York: Oxford University Press, 1993.
Plantinga, Alvin. "Why We Need Proper Function." Noûs 27 (1993): 66-82.
Plantinga, Alvin. Warranted Christian Belief. New York: Oxford University Press, 2000.
Defenders of proper functionalism:
Bergmann, Michael. "Epistemic Circularity: Malignant and Benign." Philosophy and Phenomenological Research 69 (2004): 709-727.
Bergmann, Michael. Justification without Awareness: A Defense of Epistemic Externalism. Oxford: Oxford University Press, 2006.
Clark, Kelly James. Return to Reason: A Critique of Enlightenment Evidentialism and a Defense of Reason and Belief in God. Grand Rapids, MI: Eerdmans, 1990.
McNabb, Tyler Dalton. "Proper Functionalism and the Metalevel: A Friendly Reply to Timothy and Lydia McGrew." Quaestiones Disputatae 8 (2018): 155-164.
The McGrews' critique and the internalist alternative:
Fumerton, Richard. Metaepistemology and Skepticism. Lanham, MD: Rowman & Littlefield, 1995.
McGrew, Timothy J. The Foundations of Knowledge. Lanham, MD: Rowman & Littlefield, 1995.
McGrew, Timothy, and Lydia McGrew. "Internalism and the Collapse of the Gettier Problem." Journal of Philosophical Research 23 (1998): 239-256.
McGrew, Timothy, and Lydia McGrew. Internalism and Epistemology: The Architecture of Reason. London: Routledge, 2007.
Other critics of proper functionalism:
Davidson, Donald. "Knowing One's Own Mind." Proceedings and Addresses of the American Philosophical Association 60 (1987): 441-458.
Feldman, Richard. "Proper Functionalism." Noûs 27 (1993): 34-50.
Sosa, Ernest. "Proper Functionalism and Virtue Epistemology." Noûs 27 (1993): 51-65.
On level distinctions and epistemic circularity:
Alston, William P. "Level-Confusions in Epistemology." Midwest Studies in Philosophy 5 (1980): 135-150.
Alston, William P. "Epistemic Circularity." Philosophy and Phenomenological Research 47 (1986): 1-30.
Vogel, Jonathan. "Reliabilism Leveled." Journal of Philosophy 97 (2000): 602-623.
Historical and background sources:
Descartes, René. Meditations on First Philosophy. 1641.
Gettier, Edmund L. "Is Justified True Belief Knowledge?" Analysis 23 (1963): 121-123.
Locke, John. An Essay Concerning Human Understanding. 1690.
Biblical citations:
Acts 17:2-3; Acts 17:11; 1 Peter 3:15. Scripture quotations are from the English Standard Version.
+The metaregress confuses levels. A belief can be warranted without a further belief that it is, and beliefs about warrant can themselves be warranted by faculties working properly, which is how proper functionalism is defended at the metalevel.
1. The level distinction is real.
A small child knows her mother's face. She has no belief at all about whether her face-recognition is reliable, and she does not need one. William Alston warned against level confusion: mixing up the conditions for being justified with the conditions for being justified in believing that one is justified. This should be granted, and the metaregress does not deny it. It does not say that a first-level belief (the child's belief that this is her mother) needs a warranted second-level belief (a belief about that belief) in order to be warranted (C1).
2. The strongest reply charges the argument with assuming internalism.
Tyler McNabb has replied to the McGrews at this level on Plantinga's behalf (2018). The strongest form a metalevel defense can take goes further. Second-level beliefs, such as "my memory is working properly," can themselves be warranted on proper functionalism, produced by faculties aimed at truth and working as designed. A warranted belief of that kind, the reply says, simply is a reason. To demand more, a reason the believer can inspect, is to assume internalism and then complain that externalism does not deliver it. On this reply, (P2) and (P3) use "reason" in a sense the proper functionalist never had to accept.
- This deserves to be granted as far as it goes. If "a reason" meant only "a warranted belief," (C1) would be false, because proper functionalism can supply warranted beliefs about warrant at every level.
3. A warranted belief is not a reason the believer can use.
The metaregress asks the question a person puts to herself: "Should I take this belief to be in good order?" A warranted second-level belief cannot answer that for her. A malfunctioning mind would produce the same belief with the same confidence (P1), so even if the answer is there, she cannot tell it from its counterfeit.
- Moving the question up a level moves the hidden fact up a level, and it stays just as hidden.
- Think of the inspection certificate again (P2). Hiring a second inspector to vouch for the first helps only if you can read the second inspector's license, and on proper functionalism that license is behind a locked door too.
4. Asking for a usable reason does not assume internalism.
Internalism says justification depends only on what is in the mind. (P3) says something much weaker: that sometimes we can see that a belief is in good order. That is compatible with outside conditions on knowledge, truth among them. So the demand does not beg the question, that is, quietly assume the very point in dispute. It asks whether proper functionalism leaves room for the ordinary reflective question ever to be answered, and the metaregress argues that it does not.
- Here the two sides reach bedrock. A proper functionalist can accept (C1) and deny (P3) in the McGrews' sense, holding that warranted belief at every level is all anyone ever had. That is a consistent position.
- But a position can be consistent and still cost too much. The McGrews' charge is that this one is nihilism under another name. Justification in their sense, a standing that rests on evidence a person has in view and can use, is gone at every level, and no one can answer "Should I believe this?" from where they stand.
- The cost reaches the proper functionalist's own case. He offers the McGrews arguments and expects them to see that the arguments are good. Yet on the position he has taken, no one could ever have a reason, in that sense, to think a belief formed by weighing his arguments is warranted. It is the referee's problem again (P3).
- Proper functionalists deny that anything worth having is lost. The McGrews hold that what is lost is the very thing argument is for. That is the choice each reader has to make, with the cost in full view.
5. Stopping the regress needs a principle known a priori.
The internalist can end the regress with principles of evidence that are known by reflection alone. The proper functionalist cannot, because on that theory even the standards of good reasoning are contingent (P1). So the level distinction, sound as it is, does not supply the missing stopping point.
+Internalism faces the same regress. To know that your evidence supports a belief, you need a further belief that it does, and that belief needs support too, so the internalist is no better off.
1. This is a natural and fair reply.
Its strongest form is Michael Bergmann's dilemma for internalism. A dilemma is a choice between two options that each lead somewhere bad, and Bergmann's runs like this:
- Either the believer must see that her evidence supports her belief, and then that seeing is a further belief needing support of its own;
- or she need not, and then, from her own point of view, the belief might as well be a guess.
If the internalist has to climb the same ladder, the metaregress is no argument against externalism in particular. The McGrews answer that the metaregress does not touch their internalism, and their answer turns on what kind of principle stops the climb.
2. Analytic principles stop the regress.
On the McGrews' view, the principles that say what counts as good evidence are analytic and knowable a priori. Suppose nine of the ten tickets in a hat are blue. A person can see that a ticket drawn at random is probably blue in the same way she sees that two plus two is four, and seeing it is not one more belief waiting on an outside guarantee.
- Whether every principle of good reasoning is analytic is the most disputed part of the view. The hardest case is induction, reasoning from past cases to new ones (the sun has risen every morning so far, so it will rise tomorrow), and the McGrews give it a separate defense. If some principle turned out to be contingent after all, the internalist would face the regress at that point too.
3. Reflection can correct a misleading insight.
The sharpest form of the reply says the internalist's "seeing" could itself be a malfunction, and history supplies an example. The logician Gottlob Frege built his system on a basic law he took to be a truth of logic, and in 1902 Bertrand Russell showed that it led to a contradiction.
- Grant it: a person can seem to see what is not so.
- But look at how the error came to light. Russell found it by reasoning that Frege, and anyone else, could follow and check. Reflection corrected reflection. On proper functionalism, by contrast, whether a faculty is working is a fact no further reflection can reach.
- Whether the internalist's stopping point is as secure as the McGrews claim is the live question between the two views.
4. The foundations are within reach.Classical foundationalism, Timothy McGrew's view of how knowledge is built, says that knowledge rests like a house on a foundation of beliefs that cannot be mistaken, with everything else resting on that foundation by good reasoning. The foundations are beliefs like "this is how it looks," which contain the experience they are about and so cannot be false.
- You can be wrong that the apple on the table is red, since the light may be strange. You cannot be wrong that it looks red to you right now.
- So the regress stops at beliefs a person can be certain of, connected by principles a person can see to be true.
See also:
Philosophy / Epistemology: InternalismPhilosophy / Epistemology: Classical Foundationalism
+Proper functionalism gets the hard cases right. The brain lesion that causes a true belief is reliable but not warranted, and only proper function explains why.
1. The brain lesion case is a strong objection to reliabilism.
Plantinga's lesion, which causes its victim to believe he has a lesion, is reliable and yet yields no knowledge. It is a real problem for reliabilism, and Plantinga deserves credit for it.
2. Internalism handles the case without proper function.
Think about what the victim actually has to go on. The belief pops into his head with no evidence behind it, and he has no reason at all to think he has a lesion. On internalism, that is exactly why he is unjustified, whether or not the lesion is a malfunction. So the case is a counterexample to reliability without reasons. It is no argument for proper function.
3. Proper functionalism has hard cases of its own.
Swampman, who appears to have no design plan but seems to know where he is, and the objections Feldman and Sosa pressed (C2) suggest that proper function is neither clearly necessary nor clearly sufficient for warrant.
- A condition is necessary when you cannot have the thing without it: no fuel, no fire. It is sufficient when having it guarantees the thing: being a square guarantees having four sides. Swampman challenges the first, since he seems to have warrant with no design plan at all.
+The deceived victim of an evil demon is justified but not warranted. Proper functionalism handles the case by keeping both notions: she has done her duty, but her beliefs are not knowledge.
1. Plantinga's split is honest.
The case goes back to Descartes. Imagine a powerful demon who feeds a woman false experiences, so that her whole life seems exactly as yours does: the coffee, the commute, the faces of her children. She reasons as carefully as anyone, yet nearly everything she believes about the world is false.
Plantinga grants that the demon's victim is justified in his sense of the word: dutiful, blameless, responsible. What he denies is that her beliefs are warranted. That is a consistent position.
2. The split concedes an inner standing that does real work.
Now add a second victim of the same demon, one who believes whatever she likes. Both are blameless for being deceived. Yet the first reasons well and the second badly, and that is a real difference in how well they believe. Nothing outside their minds separates them, since the same demon deceives them both. What does separate them is more than duty. It is what their evidence supports.
3. That inner standing is the one a person can use.
When a believer asks "Should I believe this?", only the inner standing can answer her from where she stands. Warrant, on Plantinga's account, can never be consulted (C1). So the standing that guides belief is the one internalism describes.
See also:
Philosophy / Epistemology: Internalism
+Plantinga argues that naturalism cannot account for proper function, and that this supports belief in God. Rejecting proper functionalism throws away an argument for God.
1. The most striking form of the argument does depend on it.
If warrant is proper function, and only design makes sense of proper function, then a naturalist cannot make sense of knowledge itself. That conclusion stands or falls with proper functionalism, and rejecting the analysis gives it up. This should be granted plainly.
2. The argument for design does not depend on the analysis of warrant.
Plantinga's case that the idea of things working "as designed" points to a designer can stand as an argument for God whether or not warrant consists in proper function. A heart that pumps blood, an eye that sees, and a memory that recalls are all working as they should, and the naturalist owes an account of that "should" either way. Christians can keep the insight that our faculties were designed while denying that being well designed is what makes a belief one a person is entitled to hold.
3. The evolutionary argument against naturalism survives too.
Plantinga's argument that the naturalist has a reason to doubt the reliability of his own faculties can be put in terms of probability and evidence, and it has been discussed in those terms at length. Plantinga himself frames its key step that way: given naturalism and unguided evolution, the chance that our faculties are reliable is low, or impossible to estimate. The argument does not need proper functionalism to be the right account of warrant.
4. Christians need not tie a sound argument to a failing analysis.
If proper functionalism falls to the metaregress, it is better for the design argument to stand on its own than to share the fate of the analysis. When two boats are roped together and one is taking on water, the sensible thing is to cut the rope.
+Without proper functionalism, ordinary Christians who cannot give arguments lose their warrant. Plantinga's model shows how belief in God can be warranted without argument.
1. Plantinga's motive is a good one.
Plantinga wanted to show that the ordinary believer, who has never read a book of apologetics, is not irrational. That is a worthy aim, and the McGrews share it.
2. Internalism does not require arguments.
Internalism requires reasons present to the mind, and an argument is only one kind of reason. Experience of God's presence, the testimony of people one has good reason to trust, the order of the world, and the facts of history all count. Most believers have a good deal of evidence, even if they have never written it down.
3. Plantinga's model ties warrant to truth in a way the believer cannot check.
Plantinga holds that if Christian belief is true, it is very likely warranted. But then the believer who asks "Is my faith warranted?" must first answer "Is my faith true?"
Critics have pressed the same point from another side, under a memorable name. Plantinga himself used the Great Pumpkin as his example of a belief that is not properly basic, meaning rightly held without being argued for from other beliefs. (The name comes from Linus in the Peanuts comic strip, who waits every Halloween for the Great Pumpkin.) Critics turned the example around: a devotee of another religion could tell the same story about his own faculties and claim that they, working as designed, produced his belief. This is the "Son of Great Pumpkin" objection.
- Plantinga answers that the parity fails. Anyone can tell such a story about his own faculties, but telling it does not make it true, and the objection offers no criticism of Christian belief that would still hold if Christian belief is true. If Christianity is true and the other faith false, only the Christian's belief comes from faculties working as designed. On his theory that is right.
- But it returns the believer to where she started. She cannot tell which of the two she is without first settling which faith is true, and settling that takes reasons she can weigh.
Far from shutting the ordinary believer out, evidence that can be examined and shared is her way out of this trap.
See also:
Philosophy / Epistemology: InternalismPhilosophical Theology / Prolegomena: Meta-Apologetics: Reformed
+Plantinga's mini-environment condition solves the Gettier problem, and internalism cannot. Proper functionalism explains why the traveler in fake barn country does not know he sees a barn.
1. The mini-environment condition is a real addition.
A Gettier case is a belief that is true and well supported but still not knowledge, because it is right only by luck. Plantinga's distinction between the broad environment and the local situation handles many of these lucky-belief cases, and it is one of the more careful treatments in the literature.
2. It works by bringing in a fact the believer cannot see.
The fake barn traveler does not know because of a hidden fact about the barns around him. Internalists have long granted that knowledge has an outside element; truth itself, after all, is a fact about the world. But an outside condition for knowledge does not show that the believer's justification is outside.
3. The McGrews offer their own account of Gettier cases.
In "Internalism and the Collapse of the Gettier Problem" (1998), the McGrews distinguish two things. One is what they call legitimation, the rationality a Gettier subject plainly has, which does not need true premises. The other is what they call justification proper, the standing that inferential knowledge (knowledge reached by reasoning from other beliefs) requires, and justification proper needs true premises. Gettier subjects have the first but not the second, and so, they argue, turning to externalism in answer to the Gettier problem is neither useful nor necessary.
- Suppose a man has watched his coworker drive a Ford to work for years and concludes, "Someone in this office owns a Ford." Unknown to him, the coworker sold it last week, but another colleague owns one. His conclusion is true, and he reached it rationally, but it rests on a false premise. On the McGrews' account, that false premise is why he lacks justification proper.
- Whether their account succeeds is debated, and (P1) through (P3) do not depend on it. What matters here is the point already made: an outside condition on knowledge is not an outside condition on justification.
See also:
Philosophy / Epistemology: Justified True Belief (JTB)
+The metaregress proves too much. Whether perception is reliable is a contingent fact too, so the same argument would make knowledge of the outside world impossible for the internalist as well.
1. The reliability of perception is indeed contingent.
An argument "proves too much" when, if it worked, it would also knock down something nearly everyone accepts, and that is the charge here. The premise behind the charge is true. Whether our senses are generally trustworthy is not true by definition, and it could have been otherwise.
2. The internalist reaches the world by a priori principles applied to experience.
On the McGrews' view, a person knows the outside world by applying principles of probable reasoning, which are knowable a priori, to beliefs about present experience, which are certain. The contingent fact, that the world is as it seems, is the conclusion of the reasoning rather than a hidden condition on it. So the regress stops at what the person can see: their experience, and the principles that connect it to the world.
- Think of astronomers in 1846 reasoning from small irregularities in the orbit of Uranus to an unseen planet. Neptune's existence was a contingent fact, but they reached it by reasoning anyone could check.
3. The externalist's contingent fact sits in a different place.
On proper functionalism, the contingent fact is not a conclusion the believer reasons to. It is a condition the believer must meet without being able to tell whether they meet it, and that is the difference the metaregress turns on.
- Go back to the astronomers. For the internalist, Neptune is where the reasoning ends up. For the proper functionalist, whether their belief is warranted depends on whether their own minds were working as designed, in the right environment, while they did the sums, and no amount of rechecking the sums can show that.
See also:
Philosophy / Epistemology: Classical Foundationalism
+Michael Bergmann argues that epistemic circularity is often benign. Every source of belief can be checked only by trusting some source, so a faculty vouching for itself is unavoidable, and harmless when nothing has given us cause to doubt it.
1. Bergmann's distinction is a serious one.
The strongest form of this objection rests on two kinds of situation.
- In the first, nobody has called a source of belief into question. A farmer trusts her eyes and her memory as she always has.
- In the second, someone has. Her doctor has told her that her eyesight may be failing.
Bergmann argues that in the first kind of situation, a belief that a source is reliable can be perfectly well justified even when it was formed by using that very source. He adds that demanding a check from outside every source asks for something no theory can supply.
- Much of this should be granted. Most people never ask whether their faculties work, and their ordinary beliefs are none the worse for it (C1).
- Think of a traveler who has never had reason to doubt her watch. She does not need to compare it with the station clock before she believes she is on time.
2. The metaregress begins when the question is asked.
The metaregress starts the moment someone asks, "Is this belief of mine warranted?" That puts her in the second kind of situation, where the source has been called into question.
- Bergmann himself grants that once a source is seriously in doubt, reassurance drawn from that same source cannot settle the doubt. The man who checks his watch against his watch (P2) learns nothing, and Bergmann does not say otherwise.
- So even on Bergmann's own terms, benign circularity explains only how a person who never asks can have warranted beliefs about warrant. It gives nothing to the person who asks, and the person who asks is the one looking for a reason (P2).
- The McGrews go further and argue that epistemic circularity is vicious, not benign. The metaregress needs only the narrower point.
3. Not every view needs a source to vouch for itself.
Circularity is unavoidable only on a view where every standard of good reasoning is contingent, so that each must be checked by a source whose own reliability is in question. That is the very feature of proper functionalism the metaregress targets (P1).
- The internalist does not certify perception by perception. She reasons from experience to the world by principles she can see to be true (D8), and seeing that a principle is true is not using it to vouch for itself.
- The externalist answers that rational insight is one more source vouching for itself. Whether it is, or whether a truth plainly present to the mind needs no further vouching, is the question at the heart of (D2).
See also:
Philosophy / Epistemology: Internalism
+Christians believe God designed the human mind to find truth. So on proper functionalism a Christian does have reason to think her beliefs warranted, and the metaregress never gets started.
1. The belief is true, and it matters.
Christians believe that God made the mind to know truth, and the McGrews, as Christians, believe it too. Plantinga's evolutionary argument (D5) turns on this very contrast: the believer in God has grounds to trust her faculties that the naturalist lacks.
- This should be granted. A person who knows that a good God designed her faculties has good reason to trust them in general.
2. On proper functionalism, the assurance comes through the same channel.
On Plantinga's model, belief in God is warranted when the sensus divinitatis produces it while working as designed. So the belief "God designed my faculties" is warranted only if a faculty is working properly, and that is the very thing in question.
- Ask whether that belief is warranted, and the climb begins again (P2).
- Think of a warranty card that is valid only if the factory was working properly on the day it printed the card. However reassuring its wording, the card cannot vouch for the factory.
3. The assurance does its work for a believer with reasons she can weigh.
If belief in God rests on evidence a person can examine, then God's goodness gives her a further reason to trust her mind, since a good God would not make minds built to deceive. Descartes argued in just this way.
- Descartes was charged in his own lifetime with arguing in a circle, using reason to prove God and then God to vouch for reason. The internalist's reply is that the principles of reasoning are seen to be true directly and do not wait on God's guarantee. What the guarantee adds is confidence in the senses and memory, where error most often creeps in.
- On that route, the assurance about God's design rests on reasons the believer has, not on a fact she cannot reach.
See also:
Philosophy / Epistemology: InternalismPhilosophy / Epistemology: Classical Foundationalism