What happened to the connection between "I want to find my keys" and "I want to look for them in the kitchen"?

Yeah I get that. And my point was that this sort of post hoc fails if I don't act on my preference. You'd have to say my acquiring the preference is what I just do, and that belief figures in my post hoc justification for the new preference I have.

We'd have to work on that. Even if we just dropped all talk of belief or rationality or community norms, what would turn a desire to find my keys into a desire to look for them in the kitchen? Maybe there are substitutes for belief, but something has to get you from one to the other.

Edit: autocorrect

There's surely a difference of some kind. We can say there's A's and there's B's, or we can say there's two kinds of A's. I don't suppose it matters unless we want to say "All A's are F"; then we'd want to be sure we don't mean "All type 1 A's are F."

In recent posts here, I've been kicking the can of outward behavior into another zip code.

It's rather that Pat searches the kitchen, and justifies his behaviour post-hoc by claiming to believe that the keys are there.

The belief is irrelevant. Pat does what Pat does. — Banno

But that can't be right, because of the knife-wielding psycho in the kitchen. I can form a preference to look there even if it's overridden by my preference to go on breathing. Rationality does seem to have a foothold here: given some preferences and beliefs, you should also have this preference. Maybe you don't act on it for whatever reason, or for no reason. Different issue.

All of that assumes by "behavior" you mean outward, publicly observable actions. Are you throwing in what I think as behavior?

I still feel pretty good about the preference version, because I get to say "If you don't want to look for your keys in the kitchen, either you don't think they're there or you don't want to find them (or you don't reason like the rest of us)." That feels solid to me.

So it would be weird if I felt a deep sense of kinship with this man ...

Anywho, this sort of gamesmanship is practically built into the multiple choice test and it would be easy to (ahem) multiply examples.

None of these strategies look like a counter to the random chooser though. (The "all B" sequence messes with our faulty intuitions.)

All of which points to something weird in Jeremiah's puzzle.

My values are halves of the 1 in 3, and 1 in 4 totals, duh. The only reason to look at it by cases as I was -- and I should have done the other two -- is to see if there's any point in a test designer choosing one of these options, and I don't see any such reason, or any reason to follow some other strategy like randomly duplicating, etc.

Dude, I didn't even think of that!

Not sure if I did the simulation right, and I'm at work now. :-(

(I had right answer being chosen from {b, c} and two students: one chooses from {a, b, c, a} and one from {a, b, c}. I don't understand the result though, so must've muffed it.)

Also, I got something wrong about the duplicated correct answer: that will help the student who doesn't read the questions even more than it helps the one who does. So whether the duplication helps the reader or non-reader more flips depending on whether it's a right duplicated or a wrong. Cool.

In any case, it still seems that duplication can only help random choosing, even if it helps in a differentiated way.

What I wanted to get back to eventually was how this puzzle conflicts with our expectations about how tests work...

(Pointless anecdote: I had a professor in college who didn't give multiple choice questions because he said they helped poor students and hurt good ones.)

Why don't multiple choice questions ever have repeated answers? If you were taking a math test and got to a question that had a repeated answer, you'd walk up to the teacher's desk and show him, and he'd make an announcement like, "Sorry, folks, Question 17 has a typo. Answer E should be '33%'."

Presented with a multiple-choice question, there are several methods you can use to choose an answer. Best method is knowing which answer is correct -- a word which here means "will be graded by the test preparer as correct". If you don't know, or don't think you know, you can go with your gut or choose randomly, and you can also eliminate answers you know are wrong before doing either of those to improve your (subjective) chances. There are some other methods, but the main point is that random choice is a fall-back when no better option is available.

Multiple choice tests are designed to test knowledge or reasoning, not luck. Tests are often designed with answers that will appear tempting if your reasoning is faulty, but not to foil a test taker who is lucky. That test takers have the option of making random choices is interesting, but test design needn't take this into account, and it's not perfectly clear that it can.

You could in fact help the random chooser by repeating correct answers, and people do this sort of thing for comic effect. What are the three most important principles of retail? Location, location, location. There are only two rules for working here: 1 is "Do what I tell you" and 2 is "Do what I tell you".

So here's the answer to the question at the top of this post: multiple choice questions never (deliberately) have repeated answers because a duplicated incorrect answer will not foil the random chooser, while a duplicated correct answer will help him. And the test preparer has no reason to help the random chooser.

The question that remains is whether a duplicated incorrect answer also helps the random chooser by changing his chance of getting the right answer from, say, 1 in 4 to 1 in 3. That would seem to depend entirely on the test taker -- that is, on whether he reads the question at all or just bubbles in something on the answer sheet.

And that opens up the possibility of a gap -- noted by several people in this thread -- between the chance of my picking the answer that is correct, on the one hand, and the chance of the answer I pick being correct. In the usual case, with no duplicated answers, these are identical, by design. But if there are duplicated wrong answers, will the random chooser who reads the questions out-perform the random chooser who doesn't?

If I did the simulation right, I get 16.6162 for the test taker who reads the questions, and 12.4914 for the test taker who doesn't.

Bonus thought about constraints: there's prescription and proscription. Foucault talked about this with different styles of morality: you can have the "default" be that everything is allowed and proscribe specific behaviors (don't wed or kill kin, etc.); or you can prescribe The One True Way to live and count every deviation as wrong. Negative versus positive constraints -- don't do that vs do this.

The funny thing about words and concepts is that even though we seem bound to think of them (and teach them) prescriptively -- here is how we use this word, this is what this concept applies to -- we're always ready to rewrite the rules so long as the rule-breaker makes sense. And that's curious, at least because the prescriptive view suggests that being successful in this way isn't even possible. And indeed this is what we will do, add a click. (Metaphors, for instance, are just recently added literal uses.) We always come back to rules, even though the rules are ever changing.

Maybe it's simplest to say that our rules are always open-ended: so far we have found the following uses for this word or concept, there will probably be others. The rules would then in essence all be permissives -- you may at least use this word or this concept in the following way -- which is not how we generally think of them.

There are always frames, whether or not they are recognized. Bringing attention to them is the best way not to be mislead. — T Clark

But -- if there are always frames, it's not that fact alone you'd be relying on to avoid being misled. "Hey wait a minute! This picture has a frame, therefore ..." You still need some other way of evaluating what's in the picture.

There are accidental forced perspective illusions in real life. Baseball has perfect examples: sometimes seen from one angle it can look as if the fielder's glove is laying right on the runner's foot as he slides, but from another angle it's clearly three inches above. Now imagine such a play making the difference in the last game of the world series, and the hometown paper printing the picture with the misleading angle and a banner headline: "WE WERE ROBBED!" It's not framing per se that's the problem here, but the choice of frame.

Yeah, that's good. That's "sense" in the sense of "rational", or I guess more broadly something like "understanding", the "sense" that's in "common sense". That explains my feeling that Willy treats the question as idiotic. (In a big Hollywood action movie, he'd answer "Because that's where the money is, you <string of insulting expletives>.")

It's like Burt Dreben's remark that great philosophers don't argue. (They just lay out their framework and you see how useful it is or isn't by using it -- proof's in the pudding.)

As far as this goes, I think it's unobjectionable. But I have two little concerns:
1. People will tend to leap to some easy relativism here: every theory shows some stuff. hides some stuff, "therefore" no theory is better than any other. And that's BS. Relativism always has this hidden absolutist expectation -- if your theory doesn't show me absolutely everything it's just as deceptive as every other theory.

2. Frames can be misleading. Think of forced perspective tricks. We want theories, frameworks, that reveal relationships that actually hold between objects pictured, not frameworks that make it appear there are relationships there aren't. We do want frameworks that reveal, and they're better than frameworks that don't.

A couple thoughts.

1. Willy Sutton's answer doesn't just shift the sense of the priest's question from one domain to the other, or from taking one kind of answer rather than another. He does also answer the question as asked, because his answer carries the implicature that he wanted money. Wanting money is clearly a sufficient motive for his behavior, but it's a motive that would usually go without saying. Emphasizing it, by cleverly not saying it, suggests that the he thinks the question is pretty stupid. It's very much as if the priest asked Willy why he crossed the road (maybe he'd been arrested for jaywalking) and Willy answered, to get to the other side.

2. Constraints seem to come in analog and digital flavors. Old radio tuner knobs, for example, could be moved continuously between their upper and lower bounds. Everything in between is an option. But the radio tuner in my car goes by steps. I can't tune to 90.2 at all.

Now look at a word like "bank": might mean a financial institution of certain sort, a building housing such an institution, the earthen boundary of a river channel, and lots of other things. But it's "clicky", in the way the duck-rabbit is clicky. It's not a word that covers everything from First National to the side of a creek -- what kind of word would that be? Is a big collection of spotlights somewhere on a continuous spectrum between those? (A duck-rabbit isn't everything from duck to rabbit on some spectrum.)

And this seems to be true even though the application of a word is obviously almost always fuzzy. (The word "bank" as a financial term will have a precise definition set down in law, but everyday folks not engaged in legal matters are just as likely to refer to a savings & loan as a "bank".)

When you creatively "extend" the usage of a word by metaphor, you're usually not enlarging some analog boundary -- I'm picturing a closed curve in a plane -- to take in new things the word can apply to; usually you're jumping to another domain entirely, adding a new click to that word's existing set of clicks.

<SophistiCat already asked my question>

Well, "mechanism" is a somewhat inelegant term -- I just mean the specific way something gets done. In the game theory approach, there's choice and agency, expectations, intentions, preferences, all that. We're not talking about any kind of determinism.

Context is fine, but it's too big. It's all well and good to say that whether "It's raining" is true depends on context, but it's better to say that "It's raining" is true iff at the time and place of utterance it is raining and English is being given a standard interpretation -- and we'd want to fill that out more. There are things about the context that don't matter for the question at hand. Nothing depends on context tout court; there are always features of the context that do the actual work.

Another example is the sort of thing Grice called attention to under the heading of "implicature". Stock example: I ask whether we should stop for gas and you reply, "There's not much between here and [our destination]." In context, that's a "yes". We can say more though: we can fill in the missing premises so that "yes" is actually entailed, and that's what allows me to infer "yes" from what you say. (And that also gives you room to cancel the implicature.) And we can fill in a lot about each participant's expectations and intentions that justify the whole operation.

I think in the long run we might end up in pretty similar places, but I'm for analyzing the crap out of everything, and formalizing everything in sight.

Quick thoughts (since I should really be working):

This sort of "linguistic accent" flattens any hierarchy we might opt for, and blocks outright the kind of distinctions I was contemplating. I get the impulse, but it feels like a rush job, and I'd like to hold off a bit to explore.

I'm tempted to say that the PSR isn't exactly a proposition anyway, and thus isn't exactly true or false. It's almost like an inference rule. (And now I'm reminded that Ramsey wanted to treat conditionals this way in general.) Maybe rather than being a conditional itself ("If x is <whatever> then x has an explanation") it's a generator of conditionals, on the fly, as needed. So an "introduction rule".

So I guess I'm in the neighborhood of, as you suggest, treating it as axiomatic -- or, rather, it looks to me like this is how you could coherently use it. (I've actually never given the PSR any thought at all.)

We could try restricting the domain of quantification to just events, say, in which case the PSR is not something that needs to be explained, since it is not an event. That seems like cheating, though. — MetaphysicsNow

That's curious because my instinct here was to say that the PSR, whatever its status, is not just a fact. Maybe as a first approximation you could say it's a sort of second-order fact, a fact about facts. It's natural to see an infinite regress looming here, but to do that we'd have to recast the PSR as inherently recursive and thus implicitly leave it outside the hierarchy we're constructing. Which is odd. However that works out, I'd at least start by not assuming that the sort of explanation appropriate to garden-variety first-order facts is appropriate to higher-order facts.

Reasonableness is itself not something precisely determinable, but consists in contextual normativities and personal self-knowledge, authenticity and good will — Janus

I think we largely agree, I just want the specifics, so that leaves a lot of work to do. For example, I don't want to lean on a word like "context" without a model of what context will actually do for us. It's why David Lewis's Convention is so important to me these days; the game theory take on norms gives you a mechanism.

For "balance", an example of what looks like a genuine (1)-style report: those people on "Deal or No Deal" who just know the million dollars is in the case they chose, maybe fans (or gamblers) who just know their team is going to win. I think these folks belong to a different epistemic community than I do, one where the idea of special knowledge only available to the faithful makes sense. Like the story of Linus and the Great Pumpkin.

One thing we haven't talked about is how we intend what we say to influence the attitudes of others.

Suppose I am, as always on this forum, looking for my keys. You tell me they're in the kitchen. I look around a little and, not seeing them, ask you if you're sure. You might say, "I'm absolutely certain I saw them in the kitchen." By saying that, you express your certainty, as we've put it, but you also encourage me to have the same attitude toward the proposition that you saw my keys in the kitchen.

There may be some subtle differences here. It's most natural to answer "Are you sure?" with a report of your own degree of confidence (or certainty or certitude). If the question is "They're in the kitchen?" maybe you answer "They certainly are." (I'm having trouble coming up with natural occasions where I'd reach for "It is certain that ..." or "It is a certainty that ...") It's not clear yet that the intended force of such statements is different rather than just grammatically more natural or convenient.

But I can think of one difference, and I'm not sure how much of a difference it is. There are at least two different sorts of claims of confidence available: (1) the simple and exclusive report of your own level of confidence, in some cases explicitly recognizing that others do not share it, even if they have the same access to evidence that you do; (2) a claim that anyone (by which is meant any member of your epistemic community) who had the same access to evidence that you do would have the same degree of confidence.

It's my suspicion that (2) is actually the standard case, and that even when people say things that sound like (1), there's an implication that they have special knowledge, access to evidence others don't. If we're arguing about whether someone will be on time, I might express, somehow, confidence that he will, with the implication that I know him, I know his habits, his record of punctuality, that I know him better than you do, and, again by implication, if you knew everything I know, you'd be similarly confident he'll be on time.

So one the one hand, an expression of confidence might mean, you can take it from me, I'm in a position to know so trust me, you can rely on my being right about this. But it might also mean that if you were in my position, you'd feel the same. But there's one other complication: I ask that you recognize my process as reliable, and suggest that what I'm confident about, I should be confident about. The suggestion regarding you is similarly that, if you were in my position, if anyone were, the right attitude to hold would be the one I hold.

I'm reluctant to admit I hadn't even thought of that, — Pseudonym

I was surprised neither of us had at first and was eager to admit it!

in the context of the authority of language users — Pseudonym

More like that. Grice is our great theorist of conversation and how it relates to logic. (His work would also be the model for Lewis's scorekeeping, etc.)

That doesn't mean that 'fit' automatically gains the status of objectivity, it's still too open to opinion at the finer scales, but it's something like objective. — Pseudonym

This is good, and I'm kicking myself for not sticking with your fashion designer metaphor, because it raises the issue of "fashion" in the other sense, changing tastes over time. I remember when I discovered there was such a thing as fashion, in this sense, in philosophy, and I was not exactly shocked but certainly disappointed.

So yes, "Does this fit?" is exactly the question we ask, and we ask it both of ourselves -- it's a feel -- and of others -- "How does it look? Does it look good?" And though we might aspire to a fashion we'd be willing to call "timeless", that doesn't quite mean what it says on the box.

No longer can we claim that the power of agreement amongst our epistemic peers confers a truth-like value to what we think and at the same time dismiss as nonsense an idea sincerely held by a minority of them. — Pseudonym

This runs deep, and I'd want to pull in Grice here somehow. I'm coming back to this as soon as I have some time. Where does PVI talk about this?

Hmmm. That argument has a funny ring to it in talking about Hume, since many people feel there's another god Hume was interested in arguing against, only he was pretty careful what he said about it while he lived.

Outstanding post. Also a good, serious piece of philosophy.

Regarding option (2), the subjective test, I'll quote Frank Ramsey for the umpteenth time: too many philosophical disputes have the form

A: "I went to Grantchester yesterday."
B: "No I didn't."

I think it's important to recognize this, but I think there's something else to the role of intuition in philosophy, that thing we use as evidence. It needn't be "purely" subjective, in the sense that a native speaker's sense of their native language isn't purely subjective.

As a native speaker, you have a double position of expertise: on the one hand there's just the breadth of knowledge you have about usage; but your speech also contributes to determining what the norms of speaking this language are. That doesn't quite make your pronouncements on usage infallible, but it does mean you are not exceeding your authority in speaking for the entire speech community.

I think there's something quite similar in philosophy. When discussing Gettier, for instance, a philosopher will be inclined to say, this does or doesn't look like knowledge to me. I don't think such pronouncements are purely subjective, or intended to be the expression of personal taste. The idea is to speak on behalf of your epistemic community, to reflect its norms, which you also participate in shaping. In effect, it's something like "This is what we call 'knowledge', isn't it guys?" And such a statement gets to function both as a reference to a norm and an encouragement to define that norm in this way.

So, subjective, yes, but in a very special way.

The difference in meaning between "I" and "it" was not at issue; the question was whether "certain" means something different in "I am certain" than it does in "We're every last one of us certain".

I noted the pragmatics issue, that "I am certain" might count as a report. I don't think we'd want to say that by being used in such a report "certain" gets a different meaning. What should we say about the difference between a report and, I guess, "an observation"?

But it means a completely different thing to say "it is certain" than to say "I am certain". Call it "objective" and "subjective" if you want, it's just the reality of the usage, these phrases mean completely different things. — Metaphysician Undercover

You've shown there's a grammatical difference, in the same way there's a grammatical difference between

• Socrates is wise
• Wisdom is instantiated by Socrates

Nowhere did you show there's a difference in meaning.

I think @Thorongil's claim is wrong in a simpler way, or confused about probability in a way that needs a different kind of response.

If I roll a fair die, it's more likely that I'll not roll a six than it is that I will. Obviously I don't need to know what I'll roll to know this.

Issues:
1. I've loaded the question by making the die "fair".
2. If we want to avoid (1) holding by definition, could we claim empirical certainty that a die is fair? Another probability seems more like what we want. Then we're headed toward the old argument about whether considering anything probable requires considering something certain. (C. I. Lewis, if memory serves, versus Richard Jeffrey.)

The question in the title is a good one, but something seems to go wrong in the way you analyze it.

(For instance, since you reverse Popper's test, "My keys are in the kitchen", being falsifiable, becomes a scientific claim. Maybe that's kinda okay, but is that really what you wanted? Where you say "falsifiable" do you really mean something more like "empirical"?)

I'd like to plump for some probability approach. You argue that we'd have to know the truth to know how close we're getting to it. I can't believe that's right but I don't have a tidy counterargument. (And I won't invoke Apo.) Maybe this would be a good part of the post the explore further.

(No irony here. I recognize I'm expressing a tentative commitment I can't quite justify. That seems to me in the neighborhood of the thread.)

Picking up from the thread on Hume next door, what about our faith in the uniformity of nature, or the validity of induction? What's the overlap between claims for which no justification can be provided and claims that cannot be falsified? (I recall Andrew reading Hume as leaving room for the mystical.)

There are lovers of truth and there are lovers of opinion. Most people are lovers of opinion, they indulge in the self; however, philosophers are lovers of truth, they seek the reality beyond the self. — Jeremiah

I sympathize. (And I'm glad you're back, given your background in mathematics.)

I too have decried what I think of as selecting a philosophical position as you might a breakfast cereal. ("I know this one's supposed to be healthier, but that other one tastes so good!" That sort of thing.) Opinions are boring.

However -- biggish however -- you have to accept the "put up or shut up" challenge. Do the sort of philosophy you aspire to and show it to us. What you're saying in this thread -- what I'm saying in this thread, right now -- is just more opinion.

I think maybe it's necessary to go through a polemical stage to find your allegiances, commit to some methodological principles, etc., but then you have to get on with it.

I admire your tenacity, but there's still an apples-and-oranges problem here.

There's a well-known video of Feynman trying to explain how magnets work:

What's notable here is the idea that explanations only make sense within a given framework. What's more, explanations you give within a framework don't justify that framework. You can point to them as contributing to the coherence of the framework, but how far that gets you is debatable...

You can also point to them as instances of the framework's not yet having been disconfirmed (or "falsified"). And what you make of that -- further non-disconfirmation -- comes down to what Hume is talking about, the framework all frameworks slot into, expectations that nature is uniform and that induction will work. Even if you did want to argue that the success of some bit of mechanics as an explanation contributes to confirming this overarching framework, you'd immediately be engaged in circular reasoning. Unless, Hume allows, there's an alternative deductive justification for accepting induction.

As a matter of English usage, you might be right, but even if you are, it's only for the nouns: the adjective that goes with both "certainty" and "certitude" is "certain".

I still don't see a philosophical point.

And it must happen — Ron Cram

When did you observe this? How did you observe this?

The usual argument here is that we presume nature to be uniform, but we cannot possibly prove that. Thus laws such as those you refer to have everything to do with human habit and custom, just as Hume claims.

I have an attitude, and you have an attitude. The statement that we both have the same attitude is not itself an expression of an attitude — Metaphysician Undercover

(1) The Earth is flat.
(2) I'm certain the Earth is flat.
(3) He's certain the Earth is flat.
(4) You, he, and I are certain the Earth is flat.
(5) Everyone is certain the Earth is flat.

Only (2) is an expression of an attitude, yes? Unless you want to argue that (1) is (2) in disguise...

(1) is a statement about the Earth; (2) - (5) are statements about people, attributing attitudes to them. (2) might be a special case - if candid it counts as a report.

Now what's the point you're making?

Our understanding of kinetic energy and electricity has nothing to do with human habit. — Ron Cram

When billiard ball A strikes billiard ball B, must (some of) A's kinetic energy be transferred to B?

Wisdom is not forgetting that there are other ways to look at things, knowing some of those ways, and being able to rank them. Wisdom often has this form: sure, that's true, and that's important, but looked at this other way, you can see there's something else that's more important. Wisdom is inclusive this way, doesn't need to deny any of reality.

It's the antidote to the tunnel-vision we all naturally fall into.

Here's a different version.

Suppose I want to have my keys. Then we might say

(A) Given that Pat wants his keys, if he believes they are in the kitchen, then ceteris paribus he will look for them in the kitchen.

Now (A) can fail if, say, there's a knife-wielding madman in my kitchen. It could also fail if I just happen to find them in the living room on my way to the kitchen. It can fail in lots of ways.

We could say (A) is like any other prediction, that it's really a claim that the consequent is probable, and it can be defeated by unlikely occurrences. But we might prefer something more like this:

(B) Given that Pat wants his keys, if he believes they are in the kitchen, and if he is rational, then ceteris paribus he will look for them in the kitchen.

That's still a prediction, and still probable. So maybe we need this:

(C) Given that Pat wants his keys, if he believes they are in the kitchen, and if he is rational, then he will wish to look for them in the kitchen.

And that can be true whether it's safe for me to enter the kitchen, and right up until the point that I find my keys elsewhere.

That only gives us an expected connection between a belief and an intended behavior, rather than an actual behavior, and lots of shadows can fall between the intention and the act. Does that bother us?

ADDENDUM

Beliefs here are how we get from one preference to another. How we get from preference to action is left for another day.

Suppose I'm looking for my keys in the kitchen. If asked why I'm doing that, I might say that I think I left them there. Someone else asked why I'm doing what I'm doing might say I believe my keys are in the kitchen. In each case, some belief is attributed to me as at least part of an explanation of my behavior.

The twin problems that arise are that (1) some other beliefs might also work as part of an explanation of my behavior, substituting for the belief previously offered; and (2) I might have that belief and yet not engage in the observed behavior.

Is that where we are?