Monday, April 25, 2011

Mailbag Monday: What Do We Know?

Mailbag Monday: A weekly segment that covers readers' questions and concerns about all things Philosophy, Bro, and Philosophy Bro that don't quite fit anywhere else. Send your questions to with 'Mailbag Monday' in the subject line.


Emanuel writes,
I was curious to see if you could do a post about justified true belief and whether it counts as knowledge. As I am sure you know, the writings on this range from the MenoTheatetus, all the way to the more contemporary writings of Edmund Gettier (Is Justified True Belief Knowledge?) and some of Goldman and Plantinga's writings. There are a whole host of positions on this question, and I have no specific requests for which ones you might cover. I would very much enjoy reading your analysis of this issue.
Great question, and one that aspiring philosophers, six-year olds, and  intractable douchebags everywhere are fond of asking: How do you know? The question of what counts and doesn't count as knowledge is one of the central problems of epistemology, the philosophy of knowledge. Chalk up another point for Plato in the "Who's your daddy?!" column, because, big surprise, he pretty much started it all; the Theatetus is where he puts in the most work on what it means to know something, and in the Meno he pretty much lays out his theory.

Now, knowledge is a pretty fucking important thing to philosophy, since what we know dictates how we act in so very many areas of our lives. Knowledge is central to human functioning; a lot of people think that it's our ability to systematically build knowledge that sets us apart. Dolphins are pretty fucking smart, but we invented the Internet with our knowledge, no big deal or anything. So we've tried to hold knowledge per se to a pretty high standard. There are a lot of concepts related to knowledge - certainty, belief, conclusiveness, and so on, and some of those are like participation ribbons - easy to have, don't mean a thing. But when we say someone knows something, that's the blue fucking ribbon. And you don't want to hand out blue ribbons to just anybody. You've got to earn that shit.
Obviously before you can know something, you have to believe it. You know that asshole who answers the question, "So you believe God is real?" with "No, I know it."? He makes my eyes roll into the back of my fucking head. If you supposedly know that God exists, you obviously also believe itBut for you to know something, it also has to be true - you can't know something that's false. It's just not how we use the word. If somebro was like, "Dude, I just know that Elvis is still alive!" you'd be like, "...really, bro? Seriously?" We don't want to say that any old strong belief that a crackpot has is knowledge.

Plato started with True Belief, and he pointed out a problem: what if someone accidentally believes something true? If some guy coughed for ten minutes straight and then was like, "No worries, it's just a cold!" And you asked, "Are you a doctor?" And he replied, "No, I'm a homeopath." You would get the fuck away from him, because he could have anything from a common cold to fucking tuberculosis for all he knows, and unless he has crushed oyster shells on hand (or whatever the fuck they use) he's fucked. If it turned out he did just have a cold, he got lucky - he didn't know. No blue ribbon for that asshole. To get past these difficulties, Plato added that we have to have a justification for believing what we believe; we have to be able to give a sufficient account of why we believe a thing. So even a strong belief that happens to be right isn't really knowledge.

So Justified True Brolief, or JTB, was the gold standard for millennia. We mostly just argued over what counts as justification. Even that, we had pretty much figured out. Philosophers were all like, "I mean, we can't be sure about most of the shit we say, but at least we figured out that knowledge is JTB. No one can take that from us. It's a slam dunk." But of course, that jinxed it, because in 1963 a bro named Edmund Gettier was up for tenure, and he hadn't published anything at all; he hated the process. So all his friends bugged him until he published a little three-page paper, which he humbly titled "Is Justified True Belief Knowledge?" In which he offers a couple of simple counterexamples.

For example, let's say you were going to be out of the country, and you would definitely miss the NBA Finals where the Bulls take on the Jazz for the second year in a row. So you tape the finals, where the Bulls wreck the Jazz, and when you get home, you watch it. Except you accidentally pop in a tape of last year's Finals, the first time the Bulls beat the Jazz, and you watch that instead without realizing your mistake. Now, you believe that the Bulls won the championship, and you're justified in believing that, since you watched games where the Bulls won; it's also true. But do you know the Bulls are champs? It seems like once again, the information you have is just lucky in that it corresponds to the truth, since you would have no idea had the Jazz actually won the second time around.

Here's another (famous) one: You're driving through a small town, where every year they have a paper-mache Barn competition. Fake-barn builders from all over the world come to show off their mad fake-barn-building skillz. Now, you have no idea that this competition is going on; you just see a bunch of barns. But you look at what happens to be the actual barn they're all emulating and think, "How quaint! A real barn!" Again, justified, true, and you believe it. But you got lucky as shit. You would have thought that no matter which barn you were looking at - after all, some competitors are fucking great at making fake barns look real. This is the world championships - no amateur looking barns here.

Examples like those are called "Gettier problems," even though he only published two himself, and it looks like Gettier was right - JTB isn't sufficient for knowledge. They gave him tenure and he never published anything ever again like a boss; he just teaches whatever he wants now. That was a revolution in epistemology - seriously, it was like the fall of the USSR, bro. Once the dominant theory collapsed, the land-grab started and suddenly everybody had a theory as to what constitutes knowledge. Tons of bros have proposed different solutions to the problems Gettier raised. Tons. A couple famous ones (you mentioned two):

Robert Nozick proposed something he called truth-tracking  to take the place of justification - if you do believe P, and P is true, (in other words, you have a True Belief) then you know P if and only if, were P false, you wouldn't believe P. So even if Elvis happens to be alive, the guy who believes Elvis is alive still doesn't know it; he would believe it even if Elvis is dead. You don't know that the Bulls are champions, since you watched the wrong tape; you would believe that they won even if they had actually lost to the Jazz. You would have thought that barn was a barn, even if it were a copy. So you don't know anything in the above examples. But your friend who watched the actual Finals does know that the Bulls are champions, since had the Bulls lost, he would believe that the Bulls aren't champions based on the actual matches.

The big problem with Nozick's account is the counterfactual, the part about how things would end up if the facts were different. How the fuck would you know how the world would be if Elvis was alive? Events don't happen in isolation; maybe if the Bulls had lost you would have heard the news, because it would be bigger news than another win, and you would have seen a newspaper or some shit. We can't just say "if this one thing had been different, and nothing else changed, would this other thing way the fuck over here stay the same?" Counterfactuals come with a lot of metaphysical baggage, bro.

Goldman tries to do something similar, but avoids the problem of counterfactuals by saying that we know P if the fact that P caused the belief that P. So, for example, the Bulls win didn't cause you to believe that the Bulls won. But Goldman also faces some problems, like what counts as a good cause? How do facts cause beliefs? What if the Bulls win caused a fucking riot, and someone broke into your house, looted your TV, and switched the labels on your tapes? Then the win caused me to watch tapes which caused me to believe that the Bulls win, but I still watched the wrong tapes.

Plantinga has tried to replace 'justification' with 'warrant', a stronger notion that requires that the cognitive processes that form beliefs are functioning "as intended," in an environment "sufficiently similar" to the environment the faculties were designed to work in, according to a working design plan, and producing statistically probably true beliefs.  It's a controversial thesis, and as with almost all of his work, Plantinga ties it to his religious convictions; it also has a lot of floating terms that aren't very clearly defined. "As intended"? What the fuck is the "intended" function of logic, if we evolved? Because if it's just to help us survive, pick good mates, and make babies, then it seems that any logically formed belief that would help us do that most of the time meets his standard; I think he would have to say that we're warranted in saying the Bulls are champs, for example. Of course, he wrote three books on it and maybe he covers my brief criticism; I'm just saying, it doesn't look good.

In fact, there are probably answers for all of my objections - I know that David Lewis has a pretty complete  metaphysics of counterfactuals, and that Goldman has a form of reliabilism that works for him. Hell, not everyone thinks we have to get rid of the JTB standard: some bros just strengthened the notion of justification to infallibility. They say that you're not justified unless your "justification" absolutely, infallibly entails the truth you believe. So "I watched a tape of the games in which the Bulls won" doesn't even count as a justification for the belief "The Bulls won."

What happened? How could we be so wrong for so long? What went wrong? Introducing justification in the first place was supposed to stop lucky shit like that from being knowledge. Requiring justification stopped us from calling any lucky belief "knowledge"; the problem is, sometimes we're justified in believing false things. There's that scene in Sherlock Holmes where Watson watches a guy get hanged, and then confirms that he has no detectable pulse and isn't even fucking breathing. Based on that, Watson believes that this bro is fucking dead, and he's perfectly justified. If some asshole went, "how can you be sure, doctor?" Watson would be like, "What the fuck else would you like as proof? HIS GODDAMN HEART ISN'T BEATING." ...except Watson was wrong. How embarrassing. The key insight that Gettier had was, "What if the thing happened to be true, just not for the expected reason?" The pieces were all there - he just pulled them together.

A successful solution will provide the strength we want 'knowledge' per se to have, but there's disagreement over how strong that is. We used to agree, but in the wake of Gettier we're not so sure what we want knowledge to be, since it looks like we have to make some sacrifices somewhere or other on our perfect conception of knowledge - we want a rigorous but practical definition, and that's a tough balancing act.

There's way more on the question of JTB than I can cover here; the point is, it's no longer the slam dunk we once thought it was.

For more on Nozick's theory, including his solution to the problem of counterfactuals, check out his Philosophical Explanations.

Plantinga wrote the Warrant Trilogy, which lays out his theory. Don't expect my paragraph to have defeated it:
Warrant: The Current Debate
Warrant and Proper Function
Warranted Christian Belief

You can read Gettier's original paper online: Is Justified True Belief Knowledge? Seriously, it's so short it hurts.

Wikipedia also has a great article on Gettier Problems.


  1. The problem with J is that it is very much a conjecture -- justification can come in any form, individual delirium or the common knowledge by acquaintance. But under all the examples that appear to be fully justified, which turns out to be false (the Sherlock Holmes thang), it's all based upon knowledge by acquaintance.

    In that case, if knowledge by acquaintance is not reliable -- the statement, 1 results in elephants, 2 results in elephants, 3 results in elephants... therefore x+1 results in elephants, is false (I'm an equations noob) -- then no belief has justifiable warrants for its justifications. So then, even normal justified true beliefs that have no faults in them are reduced down to luck.

    Epistemology is depressing.

  2. Yeah the problem is that even in cases where our reasons for believing something seem plenty good enough to draw a conclusion, there still isn't 100% certainty, which means that sometimes you can form a "justified" belief and have it turn out to be false. And if you can be justified and wrong, then you can also be justified and right but only because you got lucky. I think we just have to accept that the statement "I know X" as it's commonly used is really shorthand for something more probabilistic, like "Based on the available evidence, it is exceedingly likely that X" (except in cases where we really do have certain, infallible justification, such as analytically true statements). I guess that puts me in the revised JTB crowd.

  3. +1 for epistemology! wooo!

  4. People typically think that the "problem" spawned by Gettier is that you can have a justified false belief. In all Gettier counter-examples (GCEs), there is always a false belief - either explicit or tacit - in the reasoning that then arrives at a belief that just happens to be true. But the fact that you can have a false belief that is justified in some sense (specifically in the sense that the believer is not epistemically blameworthy - all their relevant beliefs and inferences are quite reasonable under the circumstances) has long been recognised; that is why JTB has the T in the middle. It is stipulated that the belief must be not only justified but true.

    The actual problem post-Gettier is to describe accurately the cases where JTB really is knowledge in a way that distinguishes them accurately from cases that are not knowledge. After all, everyone (or almost) agrees that the GCEs are NOT knowledge. So some JTB is, and some is not, knowledge. The mistake of a lot of epistemologists is to try to shore up J in a way that makes it impervious to counter-examples. That is a mistaken strategy, because it usually makes most of our ordinary and scientific empirical beliefs not knowledge. We need to leave justification with the possibility of getting false or Gettiered (that is, only luckily true and therefore not knowledge) beliefs, but then accurately describe what is going on when justification is not being Gettiered but is actually producing knowledge.
    On another matter, the colloquial employment of "bro" seems to derive from gang/rap culture, and not surprisingly leaves out "sis"ters, who in that culture are normally referred to as "bitches". Is it possible for the host of this site to be cool and funny without endorsing discriminatory exclusion? I hope that he - as a philosopher - can come up with the right answer.

  5. 14:17 says:

    "The mistake of a lot of epistemologists is to try to shore up J in a way that makes it impervious to counter-examples. That is a mistaken strategy, because it usually makes most of our ordinary and scientific empirical beliefs not knowledge."

    Your analyses are quite good, but you stumbled on this one.

    If you accept that knowledge can never be false, then anything that counts as knowledge must necessarily be beyond the possibility of revision.

    No good scientist would make that claim about anything within the sciences. If so, then it's unproblematic to exclude "scientific empirical beliefs" from being knowledge.

    Really, what seems to be going on is that 'knowledge' in the classical sense is a pipe dream; we might be best served by abandoning the term.

  6. I posit that the only knowledge we can truly "know" are direct experiences. I suppose the term would be qualia. The knowledge that we think of as "knowledge" is more of a cognitive abstraction of things we can't really prove, ultimately. Even proof and justification ultimately fail, which would leave us floating in an infinite sea of uncertainty, except for qualia. Qualia may not prove the existence of an objective "reality" of any sort, but they don't need to. They assert their existence in and of themselves and really cannot be doubted. If I put my hand in a bucket of water, I can still doubt the objective existence of the water, but I cannot doubt the existence of "wet" and "cold."

    Sensations then, or qualia, are the only things we know for certain.

  7. Anonymous poster who recommends abandoning the term 'knowledge':

    What do you suggest we replace it with? Something like 'understanding'?

  8. My issue with justified true belief isn't with justification. To me, at least, justification can be saved by adding the rider that "there be no true belief pertaining to your justification that would lead you to repudiate your belief."

    My problem is with the epistemic regress implied by including truth. It is trivially true that if a person holds a particular belief, P, he also holds the belief that P is true. He may not be certain of its truth, but he certainly holds it to be true, or what does it mean for him to believe it at all?

    Now, if truth realy just means certainty, that is not a problem, but most of us hold truth to higher standards than that - we believe that it has something to do with the way the world is.

    In the end, we have a definition that is only meaningful from a viewpoint that no one can ever hold. From a subjective viewpoint, truth is indiscernable. There is only belief, and any knowledge we have, we are unable to discern from our other beliefs.

    After all, it was once considered trivially true that the Earth was flat. Everyone KNEW that, even children. Except, of course, that we now know they didn't know shit. What do we believe we know that people will one day know we didn't know? What do we know that people in the future will one day believe we didn't know, when actually we did?

    It's all pretty ridiculous, realy.

  9. Ah yes, how I love epistemology.

    Visit my blog at:

  10. Anonymous poster commenting on my anonymous post:

    What do you suggest we replace ['knowledge'] with? Something like 'understanding'?

    Well, functionally, "model" is the best fit since it captures the notion of approximation without identity, but it's too awkward--imagine trying to make sense of "I model you've been sleeping around on me."

    The problem with knowledge is that we often use and think of it as if it means something absolutely true--we treat things as if they were decisively either true or false, forgetting that everything falls somewhere between those two points and nothing on them.

    The problem would be solved if we all recognized that something being 'knowledge' means two things:

    1) It fits our experience.

    2) It has no substantive challengers at this time.

  11. @Anonymous, above:

    Uhm... You seem to be misaprehending the definitions of true and false. Things are one or the other, always. If they in any way fall short of being true, they are false. True/false is a binary, not analogue, concept. To claim otherwise is to not understand what the terms mean.

    Now, you can create an analogue logical system. Indeed, math is such a system. But if you do so, you're using the terms true and false in very different meanings from their conventional ones. You may as well claim that frogs are huge, dangerous mammals, with the rider that you are using the term 'frog' to refer to the animals conventionally described as 'tigers'.

    Also, your solution has a few problems. First, you need to substantially expand upon what is meant by 'substantive'. No matter what specifications you arrive at for 'substantive', one side or the other of the evolution debate is going to disagree with you. Second, many people would have a problem with the idea that people can know things that are false, but your definition leaves that possibility open.

    Finally, if your definition were accepted, then killing everyone who attempts to mount a (substantive) challenge to your belief would be sufficient to qualify it as knowledge, no matter how ridiculous that belief might be. I'm sure you can construct a few science fiction thought experiments that allow people to 'know' crazy things under your definition. If you really can't, let me know and I'll be happy to furnish some.