This is ostensibly a review of Peg Tittle’s What If…, Collected Thought Experiments in Philosophy (Pearson Longman, 2005) with a narrow focus: post-Gettier epistemology. (In these days of “post-” anything, I had better clarify: For me, “post-Gettier epistemology” means the same as “epistemological inquiry informed by Gettier’s seminal paper”.) But this frisky little post is trying to be all of the following: (a) a lightning-fast commentary on the tools of the epistemology-teaching trade, (b) a lightning-fast commentary on philosophical genres, (c) a very brief catalog of very offensive errors in the study of epistemology, (d) a very brief catalog of very insidious errors in the study of epistemology, (e) the antidote to all those errors, offensive or insidious, and (f) a bit of stern advice for instructors and teaching assistants who wouldn’t call themselves “card-carrying epistemologists” but are occasionally assigned an introductory course in the field.
1. Overview, motivation, rules of engagement and miscellaneous thoughts
Tittle’s book may well be the largest collection of philosophical thought experiments ever assembled – 117, from the original sources, where an excerpt can be obtained. (As far as I have been able to determine, the excerpts are well-edited throughout the whole book.) Each of them is followed by a short commentary designed to give prominence to the philosophical issue to which the thought experiment speaks most obviously. The comments aim at being both intelligible to beginners and a good starting point for discussion in the classroom.
Although the book will appeal to professional philosophers generally as a reference source, it is cleverly marketed as a classroom tool by one of the largest publishers in the education business, Pearson Longman. The blurb on the back cover pushes all the right marketing buttons:
“Featuring a clear, conversational writing style that doesn’t dilute the ideas, the value of this book is in its simplicity… An ideal supplement, this brief book can be used in a variety of courses, including Introduction to Philosophy, Critical Thinking, Introduction to Ethics, and Applied Ethics… Organized by area, the thought experiments explore the major topics found in Introduction to Philosophy courses: Logic, Ethics, Epistemology, Aesthetics, Metaphysics, Philosophy of Mind, Political Philosophy, etc.”
The organization by areas – that’s where the problems begin. Some people might think it inappropriate for the author to include the whole section on philosophy of religion as a topic in the metaphysics section. That’s a questionable complaint, yes. But who, among professional philosophers, would take personal identity as an area of philosophy (as opposed to a topic in metaphysics)?
I will not pursue that line of criticism in this brief review. I’m determined to be as permissive as technically possible. Otherwise, there would be, to begin with, the never-ending rant about her choice of epistemological thought experiments.
Contemporary epistemology is allotted a total of seven thought experiments (in her own descriptions): “Gettier’s Smith and Jones (and Brown in Barcelona)”, “Skyrm’s Pyromaniac”, “Harman’s False Report”, “Goldman’s Fake Barns”, “Bonjour’s Clairvoyants”, “Plantinga’s Epistemically Inflexible Climber” and “Lehrer’s Mr. Truetemp”. Assuming that a book of this kind must contain only seven thought experiments from post-Gettier epistemology, we could forever argue about which those should be, of course. Right off the bat, I’d submit that Stewart Cohen’s New Evil Demon (to use Ernie Sosa’s label) must be one of those.  I’d next suggest that a good case could be made for the inclusion of Fred Dretske’s Disguised Mules.  Since a case would have to be made against the most vulnerable of Tittle’s choices (some of those are beyond reproach), I’d start by suggesting that she got the wrong thought experiment from Plantinga’s deep well. It’s an easy case that an argument against the sufficiency of coherence for justification is much less exciting than a case against its necessity. I’d have Keith Lehrer himself on my side.  This could go on and on. But I’m not doing it here. (It makes for great conversation, though – ideal for parties, coffee breaks with graduate students, funerals…)
My aim will be to identify inexcusable problems, either those that will twist one’s understanding of a given point beyond repair (be it a glaring distortion or a stealthy one) or those that will at least wreck any sensible plan for an introductory class in the field. There will be absolutely no nitpicking in my criticism.
There’s righteous thunder in my voice here. It’s a safe bet that the problems in Tittle’s book won’t command attention from our best epistemologists, those who are guaranteed to be heard. It’s not review material for our best journals. Dismissed as a classroom tool, Tittle’s unassuming volume will fly under the radar and land into the hands of unsuspecting, ill-equipped teachers and teaching assistants, where it can do considerable damage (in spite of the rewards it offers). My populist heart screamed bloody murder. I thought: The Certain Doubts watchtower must not fail our abused crowds now!
Before getting into the philosophical action, I thought I’d take a moment to consider the genre to which Tittle’s book belongs.
A conservative look will probably assign it to the reference shelf. It is a reference book after all. But I tend to keep it elsewhere on my shelves. Not among the introductory books, though a sensible case can be made that that’s where it belongs, as an accessory of sorts. Not with books that seek to explain what a thought experiment is and how important an instrument it is for theory building in philosophy and the sciences, most definitely not there. Tittle is not seeking company with Roy Sorensen, Tamar Gendler and James Brown.  No. I’m tempted to keep it with books that excel in offering thought experiments (either new ones or recycled versions of old ones) in a philosophically thought-provoking manner.  I’m thinking along the lines of a philosophical genre pioneered by Lewis Carroll, a genre characterized by a refreshingly “ill-behaved”, humorous, often chaotic mixture of paradoxes and thought experiments, the kind of book that Wittgenstein would have written if he had a sense of humor. Raymond Smullyan is, of course, the prime keeper of the Carrollian flame. His books This Book Needs No Title (Prentice-Hall, 1980) and 5000 B.C. and Other Philosophical Fantasies (St. Martin’s Press, 1983) are the pinnacle of that genre. But, stretching a bit, I would also include the more classroom-friendly There Are Two Errors in the the Title of this Book (Broadview Press, 1992), by Robert M. Martin, and Labyrinths of Reason (Doubleday, 1988), by William Poundstone, in that very same category. Stretch it some more and Tittle’s book won’t seem obviously out of place on that shelf. These are all books that would make for an adventurous introduction to philosophy, given some careful selection of excerpts. (Careful selection of excerpts for an adventurous course? Isn’t this like the oxymoronic “an adventurous safe trip”?! Is it just me or do you also feel inclined to think that kids are intellectually overprotected in the classroom these days? Why not let them read those books without any guidance whatsoever for a few weeks and see what happens? Aren’t we doing too much teaching, too much safe teaching, fearing the mighty judgment of the mediocre among our students in cahoots with the worst instincts in bad administrators and how difficult it gets to explain any incentive for autonomy and creativity?)
I should get back to work now.
2. Smith, Jones and Brown
Tittle’s section on “the conditions of knowledge” (her title for the post-Gettier epistemology section) gets off on the right foot by giving pride of place to Gettier’s job application case. The Brown case has always been a bad idea, though it might understandably have seemed alluring to 1963 eyes (and we’re happy to have it anyway, of course). From where I look, nobody needs two problems rolled into one. Undergraduate students, in particular, should be spared the crash course on the charms of relevant logic. Still, it would be totally unfair to criticize Tittle for feeling that the Brown case should not be ignored in her commentary. She does report on it, and her report is not problematic. But the paragraph where the Brown case is discussed deserves criticism for two important errors (or three, or four).
This is how she introduces the Brown case:
“In a similar case, Gettier shows that the problem (a justified belief happening to be true even though it was derived by using what turns out to be a false premise) can arise not only with conjunctive propositions…but also with disjunctive propositions…”
Is there a problem with “a justified belief happening to be true even though it was derived by using what turns out to be a false premise”? What problem is that?! Unless one takes the derived belief to be a case of knowledge, there most definitely is no problem. Why is that a problem for Tittle? Because it reinforces a novice’s tendencies to be confused about either what logical implication is or how it transmits justification. Both the tendency to think that there is something illegitimate in deducing a truth from a falsehood (since implication is “truth-preserving”) and the tendency to think that any such deduction deprives the deduced truth from justification for the believer have to be checked in the classroom. By leaving any reference to the real problem – the presumed acquisition of knowledge through deduction from what happens to be a justified falsehood – Tittle deals a blow to the good plan that the epistemology teacher might have had for his class. For the beleaguered teacher will now have to make a number of cautionary remarks that he might otherwise be able to do without. 
Another inconvenience is the incautious reference to disjunction and conjunction. You can’t blame a kid for thinking that the moral of all that is that conjunction and disjunction are treacherous connectives! We could certainly do without the extra worry.
Towards the end of that paragraph, we find the following: ‘Surely, says Gettier, contrary to the standard account, Smith cannot have claimed to know that “Either Jones owns a Ford, or Brown is in Barcelona”.’
Well, Gettier did not say what Tittle claims he did! And it’s a very good thing that he didn’t. The Gettier problem is not a problem about the justification of knowledge claims, which is the charitable interpretation of what Tittle puts into Gettier’s mouth. Nor is it a problem about higher-order knowledge (knowing that one knows), which is another charitable interpretation of what she is attributing to Gettier. There’s absolutely nothing in Gettier’s paper about whether Smith is entitled to claim that he knows this or that. It’s all just about Smith’s not being in a position to gain (first-order) knowledge where a traditional account of the concept would have him succeeding.
3. The Pyromaniac
Any suspicion that the problem in the last paragraph of the preceding section might be a one-off slip on Tittle’s part (and that this reviewer may be unduly concerned with the matter) will certainly be lost in view of the following:
“[I]t seems a causal connection between truth and justification is required…before one can make a claim to knowledge. Is that so for all cases in which we want to say we know something?”
For crying out loud, it’s possession of knowledge, not entitlement to make a knowledge claim, that should be the focus of our students’ attention when these thought experiments are brought before them. And, if there will ever be any fruitful hint at a conceptual entanglement of those issues, it had better be considered in light of that distinction.
You’ll be trying to catch your breath when Tittle hits you with this:
“And what sort of causal connection? How strong a causal connection?”
Oh, my!… How can those questions help move an epistemology lesson forward? Are we now supposed to take a long detour into metaphysics, into an analysis of causation, in order to gain some minimal degree of understanding of how a causalist approach gets a foothold in the analysis of knowledge? No, we’re not. But now you may be looking kind of cowardly (or worse) to those students who are slower to acknowledge the supreme wisdom of your quick evasive move when the metaphysical assault is suggested by Tittle. And you really don’t deserve to be.
4. The False Report
Here’s Tittle’s summary of Harman’s problem:
‘[H]ow can what you don’t know (especially if what you don’t know is false) “weaken” your claim to know what you do know?’
Yes, the old knowledge/knowledge claim confusion reappears in full swing. But let’s look beyond that. How can we charitably interpret the parenthetical remark? Not meaning to belabor the point, I’ll say simply that I can see no way of making charitable sense of what she writes that doesn’t imply or presuppose one of the following: Either a falsehood may be an object of knowledge or “knowing” and “believing” may properly be used interchangeably. I really can’t go any further into this point without a drink.
5. Fake Barns
There are three problems here. Tittle opens her brief discussion of the fake-barn experiment as follows:
“Goldman is exploring with this thought experiment the traditional account of knowledge as justified true belief.”
But that’s terribly unfortunate. Goldman is not repeating Gettier! He’s not concerned with refuting what has already been refuted. He does make a brief reference to the uselessness of the traditional account in generating an explanation of the epistemic situation when fake barns enter the picture. But the ostensive target – the one you can’t miss if you so much as read the first paragraph of his paper – is his own previous causal account of knowledge.
The second problem is painfully displayed here:
“Goldman initially considers that, unlike in the first case [where there are no fake items], in the second case [where there are], Henry was accidentally right – but Goldman isn’t convinced that being accidentally right is a sufficient criterion for evaluating claims of knowledge in all cases.”
For the absolute beginner, the above passage will probably seem to present an impenetrable mystery. If you have Goldman’s paper in front of you, you will understand that Tittle is trying to let you know that Goldman is not satisfied with a certain (uninformative) explanation put forward by Peter Unger. But, since the quotation above is the full extent of what she gives you about the matter, the beginner’s best bet is that he has been denied some crucial piece of information. But you know the grief that usually precedes any such bet in the classroom, don’t you?
The third problem: In a short and convoluted paragraph, Tittle tries to give you a report on why the defeasibility theory (not identified by her as such) seems unacceptable to Goldman. Here, Tittle fails by feeding her reader a lot more than he can chew. Her brief description of a defeasibility analysis is the most compressed report on the theory to be found in the literature! Moreover, her reader doesn’t get any minimal sense of the importance of Goldman’s target. Her reader is not warned that, at the time, Goldman was ill-positioned to offer his reader a useful critique of the defeasibility theory. On that very same issue of The Journal of Philosophy where Goldman’s 1976 paper was published, Peter Klein put forward a version of the theory that immediately made Goldman’s (Harman’s) objection obsolete.  Tittle seems oblivious to such “minutiae”.
Peg, Samantha is the wrong clairvoyant! The year before BonJour published his paper on “Externalist Theories of Empirical Knowledge”, which is where his clairvoyants come from, Alvin Goldman, one of BonJour’s main targets in the paper, had already considered a case that is relevantly analogous to the Samantha case and conceded that one does not know that p if one is in possession of non-overridden counterevidence for the belief that p (though this is not his way of putting it).  So, the Samantha case had been preempted. And two-thirds of Tittle’s commentary suffer from our lack of interest in that case.
Fortunately, Tittle does acknowledge the Norman case in her commentary. Norman is the right clairvoyant. Norman is not in possession of evidence against the reliability of clairvoyance (in general and on that particular occasion). Her report on the Norman case is unproblematic.
However, toward the end of the section, Tittle offers the following:
“Is it irrational to believe X for no reason? If so, how good a reason is good enough?”
Some version of the second question is fair game for the epistemology teacher. But the first one is inadmissible. At the very least, if you dignify such a simple-minded question, externalism as a whole may instantly become uninteresting. The question is not just biased in favor of internalism. It is biased in an uninteresting way. It will obviously invite an affirmative answer from unsuspecting minds. (“Irrational?! You’d have to be a raving lunatic…”) But it will be a simple-minded answer. On the most popular concept of a reason (a proposition; a reason “one has” being a proposition in one’s belief system), even internalist foundationalism becomes unfairly targeted. Only coherentism and infinitism are favored. But nobody will want that kind of cheap advantage. We would like our students to acknowledge the allure of the connection between the availability of reasons, on the one hand, and the possession of knowledge, on the other, in full view of the crucial questions: Why is the availability of reasons a good thing when knowledge is the prize? And can’t we have that good thing even when reasons are not available (or may not reasonably be expected to be appealed to)? But I can’t see that Tittle’s question is a sensible prelude to these ones. (Her question is certainly rendered useless by these questions.) 
7. The Epistemically Inflexible Climber
That hapless character helps Plantinga show that coherence is not sufficient for epistemic warrant. According to Tittle,
“Plantinga uses this thought experiment to challenge the coherence theory of knowledge, which says that a belief is justified as long as it is coherent (consistent) with other beliefs (in the relevant system or structure of belief).”
Where did she get the idea that “coherence” and “consistency” are synonyms? Not from Plantinga. Not from any competent professional epistemologist. So, right here, the epistemology teacher must step in and explain the distinction to his students. The teacher will need some time for this one.
It gets much worse when, in a single sentence, Tittle notes that, for Plantinga, coherence is not necessary for “justification” either: ‘we are often justified in believing something that “doesn’t fit” with the rest of our beliefs’. Under the very best of circumstances, this is unintelligible as it stands to even the most brilliant of readers. So, the teacher should be prepared to unpack the point. It may be a nice series of classes. If you want to do it right, the Lottery and Preface paradoxes will be discussed at some point.
The closing remark for the section is another problem. Tittle writes:
‘An interesting question might be “Which source ‘trumps’ the others when conflicts arise?” For example, what if your reason tells you to believe one thing but your experience tells you to believe another? (And one of those beliefs is coherent with the rest of your beliefs.)’
To my mind, the question is so far from becoming interesting that the most obvious response seems to be as follows: “Other things being equal, there is no trumping. Withholding belief is what is justified for you.” Ask a volunteer to explain your answer to the class and go have a coffee. They won’t need you for about ten minutes. Why only ten minutes? Because, if you’re doing the job you’re supposed to do, by the time your class gets to the second paragraph in Tittle’s commentary, you will already have given them at least some rough idea of what “epistemic justification” is supposed to mean. But any rough – even very rough – idea about the meaning of the term will have required your explaining that it applies to three propositional attitudes, which, in turn, will have required your explaining how suspension of judgment may be justified. So, it’s hard to see how that question can be a very exciting one at that stage in the conversation.
8. Mr. Truetemp
Tittle’s commentary on Keith Lehrer’s Mr. Truetemp is another bewildering tangle of banality and complexity put together in the rough-and-tumble manner that we’ve come to expect. The highlight here is the extreme philosophical complexity hidden behind some simple-minded claims and questions. My points about the section:
8.1. The commentary opens as follows:
“Lehrer is examining an epistemological position known as externalism, which says that in order for true belief to count as knowledge, there must be an appropriate connection between belief and truth.”
Well, you won’t want to have your students believe that the truth connection is an externalist cause, will you? The truth connection is the conceptual core of epistemic justification (entitlement, warrant, rationality, the right to be sure, etc.) since time immemorial. What’s new, after Gettier, is how wrong we had been since time immemorial as regards the expected simplicity of the truth connection. Externalism is only a novel account of the truth connection. And, in the wake of externalist admonition, we also get internalist views which either reject the truth connection in a conservative way (Chisholm et al.) or else reject the truth connection where it is expected, aiming at rearranging the whole field (Foley et al.).  A number of other internalist views agonize over the need to account for the truth connection somehow.
8.2. Consider this passage in Tittle’s commentary: “But, Lehrer asks with this thought experiment, what if the person has no idea how his or her beliefs are formed?” This is entirely okay. But, for Tittle, it becomes the springboard for the following question: “How knowledgeable must we be about our cognitive processes?” Now, how likely is it that you will get the philosophically proper response to this from your undergraduate students, to wit: Knowledgeable for what? (a) To be regarded as a first-order knower? Or (b) to be in possession of metaknowledge?
(a) is the interesting question, and there is a clear answer to it in Lehrer’s work: Not knowledgeable at all. Any beliefs about reliability that Lehrer’s epistemology may require (I’m not discussing it here) will only affect what he calls “personal justification”, which will, in turn, become “undefeated justification” (the defining characteristic of “acceptances” which are cases of knowledge) when additional requirements are satisfied. But those additional requirements will not include the possession of higher-order knowledge.
8.3. Tittle’s commentary segues into this: “And if knowledge about our cognitive processes is obtained through those cognitive processes, how can it be knowledge?”
The only objection I have here is: That good question is out of character with its environs. I can’t imagine its being a question that would fall into place in any introductory lesson. From the top of my head, I can think of two places where one will find important discussions leading to a non-skeptical answer to Tittle’s question: chapter 9 of Lehrer’s Theory of Knowledge (2 edn., Westview Press, 2000) and Ernest Sosa’s “How to Resolve the Pyrrhonian Problematic: A Lesson from Descartes” (Philosophical Studies 85, 1997).
8.4. But I’m afraid this review is destined to end with rolling eyes and a sense of hopelessness. Just when my respect level for Tittle had soared at the end of the previous paragraph, she closes her section on contemporary epistemology with this:
“Another possibility is that the belief is formed on the basis of a reliable third-person source… How do we determine its reliability? (Especially if our own cognitive processes or faculties are unreliable?) And how reliable must the source be?”
Have mercy, people! I’m going home now! I mean, I’ll leave you with this take-home assignment: How many problems do you see in the quotation above? Justify your answer.
9. Concluding remarks
I close with an open letter to Peg Tittle.
Peg, don’t be misled by my brand of tough love: I strongly urge people to buy your book. It’s a well-conceived project, and I hope you make the New York Times best-seller list. I really do. But, when the time comes to put out the second edition, I trust you won’t want to look as if you were barging into this sanctuary we call “epistemology” like a herd of buffaloes. Zealots like me will give you trouble you don’t need. Pearson Longman-grade simplicity is safely attainable in our field (as you came so close to proving yourself). Here’s the hot tip: Ask the experts! Just ask.
1. See Keith Lehrer and Stewart Cohen, “Justification, Truth, and Coherence” (Synthese 55, 1983), where it is indicated that that objection to externalism is Cohen’s. You won’t regret checking out the discussion of Cohen’s thought experiment in Gerald Vision’s new paper “Truly Justified Belief” (Synthese 146, 2005).
2. See Fred Dretske, “Epistemic Operators” (The Journal of Philosophy 67, 1970).
3. See Lehrer’s “Proper Function versus Systematic Coherence” in Jonathan Kvanvig, ed., Warrant in Contemporary Epistemology (Rowman & Littlefield, 1996).
4. James Brown, The Laboratory of the Mind (Routledge, 1991); Roy Sorensen, Thought Experiments (Oxford University Press, 1992); Tamar Gendler, Thought Experiment: On the Powers and Limits of Imaginary Cases (Garland Press, 2000). I was disappointed to see that Tittle makes no reference to Sorensen’s book on the issue. I don’t think that reference is optional. Brown is ignored too. She does make a reference to Gendler.
5. There are a few other new books in that genre out there that are primarily pitched for the education market, designed to appeal to undergraduate courses. A noteworthy addition to that class is Martin Cohen’s Wittgenstein’s Beetle and Other Classic Thought Experiments (Blackwell, 2005), an interesting middle ground between Tittle’s reference-market appeal and Sorensen’s erudition and depth of analysis. The distinguishing feature of Tittle’s book in that lot is that it is exclusively concerned with philosophical thought experiments and seems to offer the largest number of them.
6. Whenever this issue comes up, I feel duty bound to report that I don’t think there is any general problem involving the acquisition of knowledge through inference from a justified falsehood, since some false beliefs are epistemically benign. The problem is, instead, one of accounting for the benign/malignant distinction. See my post “On Useful Falsehoods” (December 2004) in Certain Doubts. Peter Klein has led the way here. (There is a philosophical point in his preference for the term “useful falsehoods” as opposed to my “benign falsehoods”. I have discussed it in unpublished talks since 2001.) See his forthcoming “Useful Falsehoods”, in Quentin Smith, ed., Epistemology: New Essays, Oxford University Press. See also Ted Warfield’s “Knowledge from Falsehood”, in John Hawthorne, ed., Philosophical Perspectives 19, 2005 Epistemology (Blackwell, 2005). Here’s one of the benign falsehood examples I used in a 2003 talk in Chicago (where EJ Coffman offered valuable commentary):
The Spokesperson: The very reliable spokesperson for the president assures me that (q) the president is in Jordan. Based on my belief that q, I infer that (p) the president is not in the Oval Office. Suppose that p is true and q is false: A last-minute change in his Middle Eastern tour schedule now has the president in Israel, not Jordan. Don’t I know that p?
7. See Klein’s “Knowledge, Causality, and Defeasibility”, The Journal of Philosophy 73, 1976.
8. See Goldman’s “What Is Justified Belief?”, in G.S. Pappas, ed., Justification and Knowledge (Reidel, 1979).
9. But maybe I should note, on Tittle’s behalf, that the ease with which one can “load the dice” against foundationalism, as regards the relevant intuitions, remains one of the most exciting factors in epistemology. In fact, that’s just one half of the very powerful intuitive traction that infinitism initially gets (whatever may later happen to those intuitions when you consider the implications). Keith Lehrer and Peter Klein have exploited that advantage masterfully.
10. Roderick Chisholm, “The Indispensability of Internal Justification” (Synthese 74, 1988); Richard Foley, Intellectual Trust in Oneself and Others (Cambridge University Press, 2001) chapters 1 and 2.