What about knowledge for the sake of knowledge? For instance I don’t anticipate that my belief that The Crusades took place will ever directly affect my sensory experiences in any way. Does that then mean that this belief is completely worthless and on the same level as the belief in ghosts, psychics, phlogiston, etc.?
Wouldn’t taking your chain of reasoning to its logical conclusion require one to “evict” all beliefs in everything that one has not, and does not anticipate to, personally see, hear, smell, taste, or touch? After all, how much personal sensory experience do you have that confirms the existence of atoms, for example?
I think Eliezer’s point is less strong than you think: for one thing, reading a history book is a sensory experience, and fewer history books would proclaim that The Crusades occurred in worlds where they had not than in worlds where they had.
I was going to write a more detailed reply, but then realized that any continued discussion will require us to debate what exactly the OP meant to say in his post, which is pointless since neither of us can read his mind. So let’s just call it a day.
I was going to write a more detailed reply, but then realized that any continued discussion will require us to debate what exactly the OP meant to say in his post, which is pointless since neither of us can read his mind. So let’s just call it a day.
This is something of a fallacy of gray. Of course we can read his mind, through the power of human telepathy, by reading more on the same topic. We can’t read minds perfectly, but perfect knowledge is never available anyway, and unless you can point out the specific uncertainty you have that decides the discussion, there is no sense in requiring more detail. You might want to stop the discussion for other reasons, but the reason you stated rings false.
The point is not that the ability is “magical”, but that it’s real, that we do have an ability to read minds, in exactly the same sense as Dpar appealed to the impossibility of.
First of all, calling speech “human telepathy” strikes me as a little pretentious, as well as inaccurate, since the word “telepathy” is generally accepted to have supernatural connotations. Speech is speech; no need to complicate the concept.
Secondly, the article you linked seemed a little rambling and without a clear point. All I was able to take away from it is that the meaning of words is relative. If that’s the case then I respond with “well, duh!”; if I missed a deeper point, please enlighten me.
Finally, when you take it upon yourself to question another person’s purely subjective reasoning, you’re treading very close to completely indefensible territory. If I say that I wanted to stop the discussion because I believe that the author’s intended meaning is ambiguous, it’s a tall order to question that that is indeed what I believe. Unless you can come up with clear evidence of how my behavior contradicts my stated subjective opinion, you more or less have to take my word that that really is what I think.
You misunderstand. Vladimir Nesov was not claiming that you don’t believe that the author’s intended meaning is ambiguous. Rather, he was claiming that your belief that “the author’s intended meaning is ambiguous” is false, or at least not enough to constitute a good reason for stopping the discussion.
The point of calling speech ‘human telepathy’ in this instance is that you claimed there’s no way to know what the author was thinking since we “can’t read his mind”. But there is a way to know what the author was thinking to some extent, so by reading your own reasoning backwards we therefore indeed can read minds.
I stated that taking the OP’s reasoning to its logical conclusion requires one to “evict” all beliefs in everything that one has not, and does not anticipate to, personally see, hear, smell, taste, or touch. RobinZ responded by saying that the OP’s point is less strong than I think. Since two (presumably) reasonable people can disagree on what the OP meant, his point, as it is written, is by definition ambiguous.
Where do we go from here other than debate what he really meant? What is the point of such debate since neither of us has any special insight into his thought process that would allow us to settle this difference of subjective interpretations? I believe that to be sufficient reason for stopping the discussion. I’m not sure what specifically Vladimir takes issue with here.
As to your point of human telepathy—comparing reading what someone wrote to reading his mind is a very big stretch. I can see how you could make that argument if you get really technical with word definitions, but I think that it is generally accepted that reading what a person wrote on a computer screen and reading his mind are two very different things.
I stated that taking the OP’s reasoning to its logical conclusion requires one to “evict” all beliefs in everything that one has not, and does not anticipate to, personally see, hear, smell, taste, or touch.
Right, but RobinZ was not arguing against this claim (depending on what you mean by ‘personally’ here) but rather pointing out that your reasoning was flawed.
For instance I don’t anticipate that my belief that The Crusades took place will ever directly affect my sensory experiences in any way.
RobinZ pointed out that your belief that the crusades took place affects your sensory experience; if you believe they happened, then you should anticipate having the sensory experience of seeing them in the appropriate place in a history book, if you were to check.
If you thought that your belief that the crusades happened did not imply any such anticipated experiences, then yes, it would be worthless and on the same level as belief in an invisible dragon in your garage.
So reading about something in a book is a sensory experience now? I beg to differ. A sensory experience of The Crusades would be witnessing them first hand. The sensory experience of reading about them is perceiving patterns of ink on a piece of paper.
DP
Edit: Also, I think that RobinZ didn’t state that as something that she believed, she stated that as something that she believed the OP meant. It’s that subjective interpretation of his position that I didn’t want to debate. If you wish to adapt that position as your own and debate its substance, we certainly can.
What’s important isn’t the number of degrees of removal, but that the belief’s being true corresponds to different expected sensory experiences of any kind at all than its being false. The sensory experience of perceiving patterns of ink on a piece of paper counts.
Now you could say: “reading about the Crusades in history books is strong evidence that ‘the Crusades happened’ is the current academic consensus,” and you could hypothesize that the academic consensus was wrong. This further hypothesis would lead to further expected sensory data—for instance, examining the documents cited by historians and finding that they must have been forgeries, or whatever.
If you adapt that position, then the belief in ghosts for instance will result in the sensory experience of reading or hearing about them, no? Can you then point to ANY belief that doesn’t result in a sensory experience other than something that you make up yourself out of thin air?
If the concept of sensory experience is to have any meaning at all, you can’t just extrapolate it as you see fit. If you can’t see, hear, smell, taste, or touch an object directly, you have not had sensory experience with that object. That does not mean that that object does not exist though.
So reading about something in a book is a sensory experience now? I beg to differ.
You are disputing definitions. Reading something in a book is a sort of thing you’d change expectation about depending on your model of the world, as are any other observations. If your beliefs influence your expectation about observations, they are part of your model of reality. On the other hand, if they don’t, they are sometimes too part of your model of reality, but it’s a more subtle point.
And returning to your earlier concerns, consider me having a special insight into the intended meaning, and proving counterexample to the impossibility of continuing the discussion. Reading something in a history book definitely counts as anticipated experience.
Very interesting read on disputing definitions. While the solution proposed there is very clever and elegant, this particular discussion is complicated by the fact that we’re discussing the statements of a person who is not currently participating. Coming up with alternate words to describe our ideas of what “sensory experience” means does nothing to help us understand what he meant by it. Incidentally this is why I didn’t want to get drawn into this debate to begin with.
Also—“consider me having a special insight into the intended meaning”—on what grounds shall I consider your having such special insight?
Fair enough. So if, on your authority, the OP believes that reading about something is anticipated experience, does that not then cover every rumor, fairy tale, and flat out non-sense that has ever been written? What then would be an example of a belief that CANNOT be connected to an “anticipated experience”?
I agree wholeheartedly that there are valid beliefs that don’t translate into anticipated experience. As a matter of fact what’s written there was pretty much the exact point that I was trying to make with my very first response in this topic.
Does that not, however, contradict the OP’s assertion that “Every guess of belief should begin by flowing to a specific guess of anticipation, and should continue to pay rent in future anticipations. If a belief turns deadbeat, evict it.”? That’s what I took issue with to begin with.
It does contradict that assertion, but not at first approximation, and not in the sense you took the issue with. You have to be very careful if a belief doesn’t translate into anticipated experience. Beliefs about historical facts that don’t translate into anticipated experience (or don’t follow from past experience, that is observations) are usually invalid.
You seem to place a good deal of value on the concept of anticipated experience, but you give it a definition that’s so broad that the overwhelming majority of beliefs will meet the criteria. If the belief in ghosts for instance can lead to the anticipated experience of reading about them in a book, what validity does the notion have as a means of evaluating beliefs?
When a belief (hypothesis) is about reality, it responds to new evidence, or arguments about previously known evidence. It’s reasonable to expect that as a result, some beliefs will turn out incorrect, and some certainly correct. Either way it’s not a problem: you do learn things about the world as a result, whatever the conclusion. You learn that there are no ghosts, but there are rainbows.
The problem are the beliefs that purport to be speaking about reality, but really don’t, and so you become deceived by them. Not being connected to reality through anticipated experience, they take your attention where there is no use for them, influence your decisions for no good reason, and protect themselves by ignoring any knowledge about the world you obtain.
It is a great heuristic to treat any beliefs that don’t translate into anticipated experience with utmost suspicion, or even to run away from them in horror.
How would you learn that there are no ghosts? You form the belief “there are ghosts” which leads to the anticipated experience (by your definition of such) that “I will read about ghosts in a book”, you go and read about ghosts in a book. Criteria met, belief validated. Same goes for UFOs, psychics, astrology etc. What value does the concept of anticipated experience have if it fails to filter out even the most common fallacious beliefs?
That there are books about ghosts is evidence for ghosts existing (but also for lots of other things). There are also arguments against this hypothesis, both a priori and observational. A good model/theory also explains why you’d read about ghosts even though there is no such thing.
You’re not addressing my core point though. If the criteria of anticipated experience as you define it is as likely to be satisfied by fallacious beliefs as it is by valid ones, what purpose does it serve?
I addressed that question in this comment; if something is unclear, ask away. The difference is between a belief that is incorrect, and a belief that is not even wrong.
Alright, I think I see what you’re getting it, but I still can’t help but think that your definition of sensory experience is too broad to be really useful. I mean the only type of belief that it seems to filter out is absolute nonsense like “I have a third leg that I can never see or feel”, did I get that about right?
I mean the only type of belief that it seems to filter out is absolute nonsense like “I have a third leg that I can never see or feel”, did I get that about right?
Yes. It happens all the time. It’s one way nonsense protects itself, to persist for a long time in minds of individual people and cultures.
With a bayesian twist: things don’t actually get falsified, don’t become wrong with absolute certainty, rather observations can adjust your level of belief.
Ok, I understand what you mean now. Now that you’ve clarified what Eliezer meant by anticipated experience my original objection to it is no longer applicable. Thank you for an interesting and thought provoking discussion.
Slightly OT, but this relates to something that really bugs me. People often bring up the importance of statistical analysis and the possibility of flukes/lab error, in order to prove that, “Popper was totally wrong, we get to completely ignore him and this out-dated, long-refuted notion of falsifiability.”
But the way I see it, this doesn’t refute Popper, or the notion of falsifiability: it just means we’ve generalized the notion to probabilistic cases, instead of just the binary categorization of “unfalsified” vs. “falsified”. This seems like an extension of Popper/falsifiability rather than a refutation of it. Go fig.
I reached much clearer understanding once I’ve peeled away the structure of probability measure and got down to mathematically crisp events on sample spaces (classes of possible worlds). From this perspective, there are falsifiable concepts, but they usually don’t constitute useful statements, so we work with the ones that can’t be completely falsified, even though parts of them (some of the possible worlds included in them) do get falsified all the time, when you observe something.
Isn’t that like saying we’ve generalized the theory that “all is fire” to cases where the universe is only part fire? If falsification is absolute then Popper’s insight that “all is falsification” is just plain wrong; if falsification is probabilistic then surely the relevant ideas existed before Popper as probability theory. It’s not like Popper invented the notion that if a hypothesis is falsified we shouldn’t believe it.
Falsifiability can be quantified, in bits. If the only test you have for whether something’s true or not is something lame like whether it appears in stories or not, then you have a tiny amount of falsifiability. If there is a large supply of experiments you can do, each of which provides good evidence, then it has lots of falsifiability.
(This really deserves to be formalized, in terms of something along the lines of expected bits of net evidence, but I’m not sure how to do so, exactly. Expected bits of evidence does not work, because of scenarios where there is a small chance of lots of evidence being available, but a large chance of no evidence being available.)
Just a note about terminology: “expected bits of evidence” also goes by the name of entropy, and is a good thing to maximize in designing an experiment. (My previous comment on the issue.)
And if I understand you correctly, you’re saying that the problem with entropy as a measure of falsifiability, is that someone can come up with a crank theory that gives the same predictions in every single case, except one that is near impossible to observe, but which, if it happened, would completely vindicate them?
If so, the problem with such theories is that they have to provide a lot of bits to specify that improbable event, which would be penalized under the MML formalism because it lengthens the hypothesis significantly. That may be want you want to work into a measure of falsifiability.
But then, at that point, I’m not sure if you’re measuring falsifiability per se, or just general “epistemic goodness”. It’s okay to have those characteristics you want as a separate desideratum from falsifiability.
That is the criterion which the Bayesian idea of evidence lets you relax. Instead of saying that “you need to be able to define experiments where at least one result would be completely impossible by the theory”, a Bayesian will tell you that “you need to be able to define experiments where the probability of one result under the theory is significantly different from the probability of another result”.
Look at, say, the theory that a coin is weighted towards heads. If you want to be pedantic, no result can “definitely prove” that it is not (unusual events can happen), but an even split of heads and tails (or a weighting towards tails) is much more unusual given that theory than a weighting towards heads.
Edit PS: I am totally stealing the meme that “Bayes is a generalization of Popper” from SilasBarta.
Fair point, and it was EY’s essay that showed me the connection. But keep in mind, the point of the essay is, “Bayesian inference is right, look how Popper is a crippled version of it.”
My point in saying “my” meme is different: “Popper and falsificationism are on the right track—don’t shy away from the concepts entirely just because they’re not sufficiently general.” It’s a warning against taking the failures of Popper to mean that any version of falsificationism is severely flawed.
Steal the meme, and spread it as far and as wide as you possibly can! The sooner it beats out “Popper is so 70 years ago”, the better. (Kind of ironic that Bayes long predated Popper, though the formalization of [what we now call] Bayesian inference did not.)
Example of my academically-respected arch-nemesis arguing the exact anti-falsificationist view I was criticizing.
As Robin’s explained below Bayesianism doesn’t do that. You should also see the works of Lakatos and Quine where they discuss the idea that falsification is flawed because all claims have auxiliary hypotheses and one can’t falsify any hypothesis in isolation even if you are trying to construct a neo-Popperian framework.
Yes, but that still doesn’t show falsificationism to be wrong, as opposed to “narrow” or “insufficiently generalized”. Lakatos and Quine have also failed to show how it’s a problem that you can’t rigidly falsifiy a hypothesis in isolation: Just as you can generalize Popper’s binary “falsified vs. unfalsified” to probabilistic cases, you can construct a Bayes net that shows how your various beliefs (including the auxiliary hypotheses) imply particular observations.
The relative likelihoods they place on the observations allow you to know the relative amount by which those various beliefs are attenuated or amplified by any particular observation. This method gives you the functional equivalent of testing hypotheses in isolation, since some of them will be attenuated the most.
If I remember rightly, that’s where poor old Popper came unstuck: having thought of the falsifiability criterion, he couldn’t work out how to rigorously make it flexible. And as no experiment’s exactly 100% uppercase-D Definitive, that led to some philosophers piling on the idea of falsifiability, as JoshuaZ said.
The key idea is “severe testing”, where a “severe test” is a test likely to expose a specific error in a model, if such an error is present. Those models that pass more, and more severe, tests can be regarded as more useful than those that don’t. This approach also disarms the “auxiliary hypotheses” objection JoshuaZ paraphrased; one can just submit those hypotheses to severe testing too. (I wouldn’t be surprised to find out that’s roughly equivalent to the Bayes net approach SilasBarta mentioned.)
The LessWrong FAQ says that there is value in replying to old content, so I’m commenting in hopes that it is useful to someone in the future, and just for the sake of organizing my thoughts.
I would have phrased this differently than Yudkowsky, but I think I understand the concept he was getting at when he gave this example:
Or suppose your postmodern English professor teaches you that the famous writer Wulky Wilkinsen is actually a “post-utopian”. What does this mean you should expect from his books? Nothing. The belief, if you can call it that, doesn’t connect to sensory experience at all.
His point is that this is just semantics. It makes no difference to the world whether we label something “post-utopian” or “aegffsdfa eereraksrfa” or anything else. The words you read in the book will be the same. The reason I don’t like this example is that, if I actually knew some literary jargon, I might get some real verifiable information that does actually mean I should expect a specific kind of sensory experience. It’s just that the classification scheme is arbitrary, and so is my belief that one classification scheme is “correct”.
The label is just a label, so arguing about classification schemes is just semantics. Using this definition, your belief that the crusades took place would affect what sorts of things you would expect to read, and what sorts of archeological finds you would expect to find if you went looking for them. However, if you believe that the crusades marked the beginning of the high middle ages, that would just be semantics. We could say that the middle ages started at the sacking of Rome, or we could make a label like “dark ages” to describe the intermediary period. What we call it and how we classify it makes no difference in the actual reality of history. It’s just semantics.
Semantic labels are part of the structure of an explicit model. For instance, the Chinese use the same word for both “rat” and “mouse”. A model with a ratmouse vertex will behave differently to a model with separate rat and mouse verteces. The structure and function of model affect what it predicts, what it’s users can notice, how they behave. Agents do not passively receive a stream of predetermined experiences, they interact with the world, and the experiences they can expect depend on the structure and function of their models...
..and more besides. Models contain evaluative weightings as well as neutral structure. For instance, in the English speaking world, mice have the connotation of being cute, rats of being vermin. The professor might not be failing to specify an empirical confirmable concept when describing the writer as a post utopian: she might rather be succeeding in tweaking her students’ evaluative model. She might be aiming at making a social or political point.
There is a long history of the political influence of language ranging from Greek rhetoricIan’s to Orwell’ s essays. A STEM type might consider it pointless, to focus on such issues, rather than what can be proved objectively. A humanities type might also consider it pointless to focus on objective, empirical claims with no social or political upshot. Neither complaint is really about meaningfullness or semantics, in the sense if the meaningfulness of the words, rather they are both about the subjectively evaluated pointfulness of an activity.
By a convoluted meta level irony, the way the way the term “semantics” is often used is itself a way if funneling the reader towards a conclusion. We have seen that there are circumstances where a semantic change would make a difference: where it makes a structural/functional change, and where it makes an evaluative/connotational difference. Since these circumstances don’t always to apply, there are circumstances where a semantic change really is trivial, really “just semantics”. For instance, if the word cat were replaced by the word zeb, in a connotationally neutral way, that would be semantics of a pointless kind that doesn’t change anything. But that situation is atypical. Although the standard rhetoric about what is “just semantic” suggests the opposite., most rewordings make a difference. Indeed, it is likely that people object to recordings because they do make a difference, not because they don’t.
Consider: A: So youre pro abortion?
B: I’m pro choice
A: Thats just semantics.
A has spotted that B’s rewording has strengthened his argument, by introducing a phrasing with a positive connotation, and so she objects to it… using the common apprehension that rewordings are just semantics, and don’t change anything!
Thanks for breaching that topic. I considered pointing out that my “aegffsdfa eereraksrfa” example might be more difficult to pronounce than “post-utopian”, and so actually would have an impact on the world in general. On reflection, I decided to make the assertion that it “makes no difference”, since that would spare a lot of confusion. It’s a good first order approximation. When introducing a topic, it’s important to take the Bohr model view of the world before trying to explain quarks and leptons.
The entanglement of semantic language with our interpretation of reality clouds things. Scientific language is precise, but often dry and hard to understand. However, by de-coupling the two worlds, we study the underlying reality without those (or perhaps with only minimal) distorting effects from our language. That’s what we are doing when we talk about Map and Territory here on LW. We get a better map from this, but if we also compare the collective maps of societies to the best maps of reality, we can look for systematic differences. Some of these are cognitive biases, which we tend to concentrate on here on LW. However, there are also many other interesting or useful things that we can learn about ourselves as mapmakers. For example, the Bouba/kiki effect might help us choose more intuitive vocabulary as we build a more and more extensive set of jargon.
Just studying the way languages evolve can be informative, whether it’s rigorously using Computational Linguistics or informally by an author or artist. The mere existence of a formal scientific understanding of reality allows a poet or philosopher, if they are familiar only with the answers but not the underlying explanations, to look at some facet of human nature and ask “isn’t it odd when people...”. A great deal of social commentary is built from that one question.
What about knowledge for the sake of knowledge? For instance I don’t anticipate that my belief that The Crusades took place will ever directly affect my sensory experiences in any way. Does that then mean that this belief is completely worthless and on the same level as the belief in ghosts, psychics, phlogiston, etc.?
Wouldn’t taking your chain of reasoning to its logical conclusion require one to “evict” all beliefs in everything that one has not, and does not anticipate to, personally see, hear, smell, taste, or touch? After all, how much personal sensory experience do you have that confirms the existence of atoms, for example?
DP
I think Eliezer’s point is less strong than you think: for one thing, reading a history book is a sensory experience, and fewer history books would proclaim that The Crusades occurred in worlds where they had not than in worlds where they had.
I was going to write a more detailed reply, but then realized that any continued discussion will require us to debate what exactly the OP meant to say in his post, which is pointless since neither of us can read his mind. So let’s just call it a day.
DP
This is something of a fallacy of gray. Of course we can read his mind, through the power of human telepathy, by reading more on the same topic. We can’t read minds perfectly, but perfect knowledge is never available anyway, and unless you can point out the specific uncertainty you have that decides the discussion, there is no sense in requiring more detail. You might want to stop the discussion for other reasons, but the reason you stated rings false.
I was expecting the link to be Mundane Magic.
The point is not that the ability is “magical”, but that it’s real, that we do have an ability to read minds, in exactly the same sense as Dpar appealed to the impossibility of.
First of all, calling speech “human telepathy” strikes me as a little pretentious, as well as inaccurate, since the word “telepathy” is generally accepted to have supernatural connotations. Speech is speech; no need to complicate the concept.
Secondly, the article you linked seemed a little rambling and without a clear point. All I was able to take away from it is that the meaning of words is relative. If that’s the case then I respond with “well, duh!”; if I missed a deeper point, please enlighten me.
Finally, when you take it upon yourself to question another person’s purely subjective reasoning, you’re treading very close to completely indefensible territory. If I say that I wanted to stop the discussion because I believe that the author’s intended meaning is ambiguous, it’s a tall order to question that that is indeed what I believe. Unless you can come up with clear evidence of how my behavior contradicts my stated subjective opinion, you more or less have to take my word that that really is what I think.
DP
You misunderstand. Vladimir Nesov was not claiming that you don’t believe that the author’s intended meaning is ambiguous. Rather, he was claiming that your belief that “the author’s intended meaning is ambiguous” is false, or at least not enough to constitute a good reason for stopping the discussion.
The point of calling speech ‘human telepathy’ in this instance is that you claimed there’s no way to know what the author was thinking since we “can’t read his mind”. But there is a way to know what the author was thinking to some extent, so by reading your own reasoning backwards we therefore indeed can read minds.
I stated that taking the OP’s reasoning to its logical conclusion requires one to “evict” all beliefs in everything that one has not, and does not anticipate to, personally see, hear, smell, taste, or touch. RobinZ responded by saying that the OP’s point is less strong than I think. Since two (presumably) reasonable people can disagree on what the OP meant, his point, as it is written, is by definition ambiguous.
Where do we go from here other than debate what he really meant? What is the point of such debate since neither of us has any special insight into his thought process that would allow us to settle this difference of subjective interpretations? I believe that to be sufficient reason for stopping the discussion. I’m not sure what specifically Vladimir takes issue with here.
As to your point of human telepathy—comparing reading what someone wrote to reading his mind is a very big stretch. I can see how you could make that argument if you get really technical with word definitions, but I think that it is generally accepted that reading what a person wrote on a computer screen and reading his mind are two very different things.
DP
Right, but RobinZ was not arguing against this claim (depending on what you mean by ‘personally’ here) but rather pointing out that your reasoning was flawed.
RobinZ pointed out that your belief that the crusades took place affects your sensory experience; if you believe they happened, then you should anticipate having the sensory experience of seeing them in the appropriate place in a history book, if you were to check.
If you thought that your belief that the crusades happened did not imply any such anticipated experiences, then yes, it would be worthless and on the same level as belief in an invisible dragon in your garage.
So reading about something in a book is a sensory experience now? I beg to differ. A sensory experience of The Crusades would be witnessing them first hand. The sensory experience of reading about them is perceiving patterns of ink on a piece of paper.
DP
Edit: Also, I think that RobinZ didn’t state that as something that she believed, she stated that as something that she believed the OP meant. It’s that subjective interpretation of his position that I didn’t want to debate. If you wish to adapt that position as your own and debate its substance, we certainly can.
What’s important isn’t the number of degrees of removal, but that the belief’s being true corresponds to different expected sensory experiences of any kind at all than its being false. The sensory experience of perceiving patterns of ink on a piece of paper counts.
Now you could say: “reading about the Crusades in history books is strong evidence that ‘the Crusades happened’ is the current academic consensus,” and you could hypothesize that the academic consensus was wrong. This further hypothesis would lead to further expected sensory data—for instance, examining the documents cited by historians and finding that they must have been forgeries, or whatever.
If you adapt that position, then the belief in ghosts for instance will result in the sensory experience of reading or hearing about them, no? Can you then point to ANY belief that doesn’t result in a sensory experience other than something that you make up yourself out of thin air?
If the concept of sensory experience is to have any meaning at all, you can’t just extrapolate it as you see fit. If you can’t see, hear, smell, taste, or touch an object directly, you have not had sensory experience with that object. That does not mean that that object does not exist though.
DP
Yes, ghost stories are evidence for the existence of ghosts. Just not very strong evidence.
There can be indirect sensory evidence as well as direct.
You are disputing definitions. Reading something in a book is a sort of thing you’d change expectation about depending on your model of the world, as are any other observations. If your beliefs influence your expectation about observations, they are part of your model of reality. On the other hand, if they don’t, they are sometimes too part of your model of reality, but it’s a more subtle point.
And returning to your earlier concerns, consider me having a special insight into the intended meaning, and proving counterexample to the impossibility of continuing the discussion. Reading something in a history book definitely counts as anticipated experience.
Very interesting read on disputing definitions. While the solution proposed there is very clever and elegant, this particular discussion is complicated by the fact that we’re discussing the statements of a person who is not currently participating. Coming up with alternate words to describe our ideas of what “sensory experience” means does nothing to help us understand what he meant by it. Incidentally this is why I didn’t want to get drawn into this debate to begin with.
Also—“consider me having a special insight into the intended meaning”—on what grounds shall I consider your having such special insight?
I’ve closely followed Yudkowsky’s work for a while, and have a pretty good model of what he believes on topics he publicly discusses.
Fair enough. So if, on your authority, the OP believes that reading about something is anticipated experience, does that not then cover every rumor, fairy tale, and flat out non-sense that has ever been written? What then would be an example of a belief that CANNOT be connected to an “anticipated experience”?
See this comment on the first part of your question and this page on the second (but, again, there are valid beliefs that don’t translate into anticipated experience).
I agree wholeheartedly that there are valid beliefs that don’t translate into anticipated experience. As a matter of fact what’s written there was pretty much the exact point that I was trying to make with my very first response in this topic.
Does that not, however, contradict the OP’s assertion that “Every guess of belief should begin by flowing to a specific guess of anticipation, and should continue to pay rent in future anticipations. If a belief turns deadbeat, evict it.”? That’s what I took issue with to begin with.
It does contradict that assertion, but not at first approximation, and not in the sense you took the issue with. You have to be very careful if a belief doesn’t translate into anticipated experience. Beliefs about historical facts that don’t translate into anticipated experience (or don’t follow from past experience, that is observations) are usually invalid.
You seem to place a good deal of value on the concept of anticipated experience, but you give it a definition that’s so broad that the overwhelming majority of beliefs will meet the criteria. If the belief in ghosts for instance can lead to the anticipated experience of reading about them in a book, what validity does the notion have as a means of evaluating beliefs?
When a belief (hypothesis) is about reality, it responds to new evidence, or arguments about previously known evidence. It’s reasonable to expect that as a result, some beliefs will turn out incorrect, and some certainly correct. Either way it’s not a problem: you do learn things about the world as a result, whatever the conclusion. You learn that there are no ghosts, but there are rainbows.
The problem are the beliefs that purport to be speaking about reality, but really don’t, and so you become deceived by them. Not being connected to reality through anticipated experience, they take your attention where there is no use for them, influence your decisions for no good reason, and protect themselves by ignoring any knowledge about the world you obtain.
It is a great heuristic to treat any beliefs that don’t translate into anticipated experience with utmost suspicion, or even to run away from them in horror.
How would you learn that there are no ghosts? You form the belief “there are ghosts” which leads to the anticipated experience (by your definition of such) that “I will read about ghosts in a book”, you go and read about ghosts in a book. Criteria met, belief validated. Same goes for UFOs, psychics, astrology etc. What value does the concept of anticipated experience have if it fails to filter out even the most common fallacious beliefs?
That there are books about ghosts is evidence for ghosts existing (but also for lots of other things). There are also arguments against this hypothesis, both a priori and observational. A good model/theory also explains why you’d read about ghosts even though there is no such thing.
You’re not addressing my core point though. If the criteria of anticipated experience as you define it is as likely to be satisfied by fallacious beliefs as it is by valid ones, what purpose does it serve?
I addressed that question in this comment; if something is unclear, ask away. The difference is between a belief that is incorrect, and a belief that is not even wrong.
Alright, I think I see what you’re getting it, but I still can’t help but think that your definition of sensory experience is too broad to be really useful. I mean the only type of belief that it seems to filter out is absolute nonsense like “I have a third leg that I can never see or feel”, did I get that about right?
Yes. It happens all the time. It’s one way nonsense protects itself, to persist for a long time in minds of individual people and cultures.
(More generally, see anti-epistemology.)
So essentially what you and Eliezer are referring to as “anticipated experience” is just basic falsifiability then?
With a bayesian twist: things don’t actually get falsified, don’t become wrong with absolute certainty, rather observations can adjust your level of belief.
Ok, I understand what you mean now. Now that you’ve clarified what Eliezer meant by anticipated experience my original objection to it is no longer applicable. Thank you for an interesting and thought provoking discussion.
Slightly OT, but this relates to something that really bugs me. People often bring up the importance of statistical analysis and the possibility of flukes/lab error, in order to prove that, “Popper was totally wrong, we get to completely ignore him and this out-dated, long-refuted notion of falsifiability.”
But the way I see it, this doesn’t refute Popper, or the notion of falsifiability: it just means we’ve generalized the notion to probabilistic cases, instead of just the binary categorization of “unfalsified” vs. “falsified”. This seems like an extension of Popper/falsifiability rather than a refutation of it. Go fig.
I reached much clearer understanding once I’ve peeled away the structure of probability measure and got down to mathematically crisp events on sample spaces (classes of possible worlds). From this perspective, there are falsifiable concepts, but they usually don’t constitute useful statements, so we work with the ones that can’t be completely falsified, even though parts of them (some of the possible worlds included in them) do get falsified all the time, when you observe something.
Isn’t that like saying we’ve generalized the theory that “all is fire” to cases where the universe is only part fire? If falsification is absolute then Popper’s insight that “all is falsification” is just plain wrong; if falsification is probabilistic then surely the relevant ideas existed before Popper as probability theory. It’s not like Popper invented the notion that if a hypothesis is falsified we shouldn’t believe it.
Falsifiability can be quantified, in bits. If the only test you have for whether something’s true or not is something lame like whether it appears in stories or not, then you have a tiny amount of falsifiability. If there is a large supply of experiments you can do, each of which provides good evidence, then it has lots of falsifiability.
(This really deserves to be formalized, in terms of something along the lines of expected bits of net evidence, but I’m not sure how to do so, exactly. Expected bits of evidence does not work, because of scenarios where there is a small chance of lots of evidence being available, but a large chance of no evidence being available.)
Just a note about terminology: “expected bits of evidence” also goes by the name of entropy, and is a good thing to maximize in designing an experiment. (My previous comment on the issue.)
And if I understand you correctly, you’re saying that the problem with entropy as a measure of falsifiability, is that someone can come up with a crank theory that gives the same predictions in every single case, except one that is near impossible to observe, but which, if it happened, would completely vindicate them?
If so, the problem with such theories is that they have to provide a lot of bits to specify that improbable event, which would be penalized under the MML formalism because it lengthens the hypothesis significantly. That may be want you want to work into a measure of falsifiability.
But then, at that point, I’m not sure if you’re measuring falsifiability per se, or just general “epistemic goodness”. It’s okay to have those characteristics you want as a separate desideratum from falsifiability.
Isn’t it an essential criteria of falsifiability to be able to design an experiment that can DEFINITIVELY prove the theory false?
That is the criterion which the Bayesian idea of evidence lets you relax. Instead of saying that “you need to be able to define experiments where at least one result would be completely impossible by the theory”, a Bayesian will tell you that “you need to be able to define experiments where the probability of one result under the theory is significantly different from the probability of another result”.
Look at, say, the theory that a coin is weighted towards heads. If you want to be pedantic, no result can “definitely prove” that it is not (unusual events can happen), but an even split of heads and tails (or a weighting towards tails) is much more unusual given that theory than a weighting towards heads.
Edit PS: I am totally stealing the meme that “Bayes is a generalization of Popper” from SilasBarta.
I’m pretty sure that was handily discussed in An Intuitive Explanation of Bayes’s Theorem and A Technical Explanation of Technical Explanation.
Fair point, and it was EY’s essay that showed me the connection. But keep in mind, the point of the essay is, “Bayesian inference is right, look how Popper is a crippled version of it.”
My point in saying “my” meme is different: “Popper and falsificationism are on the right track—don’t shy away from the concepts entirely just because they’re not sufficiently general.” It’s a warning against taking the failures of Popper to mean that any version of falsificationism is severely flawed.
Ehhcks-cellent!
Steal the meme, and spread it as far and as wide as you possibly can! The sooner it beats out “Popper is so 70 years ago”, the better. (Kind of ironic that Bayes long predated Popper, though the formalization of [what we now call] Bayesian inference did not.)
Example of my academically-respected arch-nemesis arguing the exact anti-falsificationist view I was criticizing.
As Robin’s explained below Bayesianism doesn’t do that. You should also see the works of Lakatos and Quine where they discuss the idea that falsification is flawed because all claims have auxiliary hypotheses and one can’t falsify any hypothesis in isolation even if you are trying to construct a neo-Popperian framework.
Yes, but that still doesn’t show falsificationism to be wrong, as opposed to “narrow” or “insufficiently generalized”. Lakatos and Quine have also failed to show how it’s a problem that you can’t rigidly falsifiy a hypothesis in isolation: Just as you can generalize Popper’s binary “falsified vs. unfalsified” to probabilistic cases, you can construct a Bayes net that shows how your various beliefs (including the auxiliary hypotheses) imply particular observations.
The relative likelihoods they place on the observations allow you to know the relative amount by which those various beliefs are attenuated or amplified by any particular observation. This method gives you the functional equivalent of testing hypotheses in isolation, since some of them will be attenuated the most.
Right, I was speaking in a non-Bayesian context.
If I remember rightly, that’s where poor old Popper came unstuck: having thought of the falsifiability criterion, he couldn’t work out how to rigorously make it flexible. And as no experiment’s exactly 100% uppercase-D Definitive, that led to some philosophers piling on the idea of falsifiability, as JoshuaZ said.
But more recent work in philosophy of science suggests a more sophisticated way to talk about how falsifiability can work in the real world.
The key idea is “severe testing”, where a “severe test” is a test likely to expose a specific error in a model, if such an error is present. Those models that pass more, and more severe, tests can be regarded as more useful than those that don’t. This approach also disarms the “auxiliary hypotheses” objection JoshuaZ paraphrased; one can just submit those hypotheses to severe testing too. (I wouldn’t be surprised to find out that’s roughly equivalent to the Bayes net approach SilasBarta mentioned.)
At the bottom of the sidebar at the bottom, you will find a list of top contributors; Vladimir Nesov is on the list.
Belatedly: Welcome to Less Wrong! Please feel free to introduce yourself.
A belated thanks! :)
DP
The LessWrong FAQ says that there is value in replying to old content, so I’m commenting in hopes that it is useful to someone in the future, and just for the sake of organizing my thoughts.
I would have phrased this differently than Yudkowsky, but I think I understand the concept he was getting at when he gave this example:
His point is that this is just semantics. It makes no difference to the world whether we label something “post-utopian” or “aegffsdfa eereraksrfa” or anything else. The words you read in the book will be the same. The reason I don’t like this example is that, if I actually knew some literary jargon, I might get some real verifiable information that does actually mean I should expect a specific kind of sensory experience. It’s just that the classification scheme is arbitrary, and so is my belief that one classification scheme is “correct”.
The label is just a label, so arguing about classification schemes is just semantics. Using this definition, your belief that the crusades took place would affect what sorts of things you would expect to read, and what sorts of archeological finds you would expect to find if you went looking for them. However, if you believe that the crusades marked the beginning of the high middle ages, that would just be semantics. We could say that the middle ages started at the sacking of Rome, or we could make a label like “dark ages” to describe the intermediary period. What we call it and how we classify it makes no difference in the actual reality of history. It’s just semantics.
Semantic labels are part of the structure of an explicit model. For instance, the Chinese use the same word for both “rat” and “mouse”. A model with a ratmouse vertex will behave differently to a model with separate rat and mouse verteces. The structure and function of model affect what it predicts, what it’s users can notice, how they behave. Agents do not passively receive a stream of predetermined experiences, they interact with the world, and the experiences they can expect depend on the structure and function of their models...
..and more besides. Models contain evaluative weightings as well as neutral structure. For instance, in the English speaking world, mice have the connotation of being cute, rats of being vermin. The professor might not be failing to specify an empirical confirmable concept when describing the writer as a post utopian: she might rather be succeeding in tweaking her students’ evaluative model. She might be aiming at making a social or political point.
There is a long history of the political influence of language ranging from Greek rhetoricIan’s to Orwell’ s essays. A STEM type might consider it pointless, to focus on such issues, rather than what can be proved objectively. A humanities type might also consider it pointless to focus on objective, empirical claims with no social or political upshot. Neither complaint is really about meaningfullness or semantics, in the sense if the meaningfulness of the words, rather they are both about the subjectively evaluated pointfulness of an activity.
By a convoluted meta level irony, the way the way the term “semantics” is often used is itself a way if funneling the reader towards a conclusion. We have seen that there are circumstances where a semantic change would make a difference: where it makes a structural/functional change, and where it makes an evaluative/connotational difference. Since these circumstances don’t always to apply, there are circumstances where a semantic change really is trivial, really “just semantics”. For instance, if the word cat were replaced by the word zeb, in a connotationally neutral way, that would be semantics of a pointless kind that doesn’t change anything. But that situation is atypical. Although the standard rhetoric about what is “just semantic” suggests the opposite., most rewordings make a difference. Indeed, it is likely that people object to recordings because they do make a difference, not because they don’t.
Consider: A: So youre pro abortion? B: I’m pro choice A: Thats just semantics.
A has spotted that B’s rewording has strengthened his argument, by introducing a phrasing with a positive connotation, and so she objects to it… using the common apprehension that rewordings are just semantics, and don’t change anything!
Thanks for breaching that topic. I considered pointing out that my “aegffsdfa eereraksrfa” example might be more difficult to pronounce than “post-utopian”, and so actually would have an impact on the world in general. On reflection, I decided to make the assertion that it “makes no difference”, since that would spare a lot of confusion. It’s a good first order approximation. When introducing a topic, it’s important to take the Bohr model view of the world before trying to explain quarks and leptons.
The entanglement of semantic language with our interpretation of reality clouds things. Scientific language is precise, but often dry and hard to understand. However, by de-coupling the two worlds, we study the underlying reality without those (or perhaps with only minimal) distorting effects from our language. That’s what we are doing when we talk about Map and Territory here on LW. We get a better map from this, but if we also compare the collective maps of societies to the best maps of reality, we can look for systematic differences. Some of these are cognitive biases, which we tend to concentrate on here on LW. However, there are also many other interesting or useful things that we can learn about ourselves as mapmakers. For example, the Bouba/kiki effect might help us choose more intuitive vocabulary as we build a more and more extensive set of jargon.
Just studying the way languages evolve can be informative, whether it’s rigorously using Computational Linguistics or informally by an author or artist. The mere existence of a formal scientific understanding of reality allows a poet or philosopher, if they are familiar only with the answers but not the underlying explanations, to look at some facet of human nature and ask “isn’t it odd when people...”. A great deal of social commentary is built from that one question.