6 ms·

(I'm a philosophy professor.) Some training in logic -- reflecting on what make good reasons and justified conclusions -- does have value. It has the same sort

(I'm a philosophy professor.) Some training in logic -- reflecting on what make good reasons and justified conclusions -- does have value. It has the same sort of value that learning grammar does. It enables you to make certain things explicit, to put them under conscious scrutiny and control, when needed. But I think this sort of enumeration of logical fallacies is of very limited value. In the abstract there are only a few ways that arguments fail, and informal fallacies are all iterations on an extremely similar theme. Moreover, the kind of argument for which it is easiest to characterize fallacies is deductive, and many instances of reasoning that appear deductive are in fact statistical, inductive, inferences to the best explanation, or some combination thereof. Even for deductive arguments, what if any fallacies are being committed will depend upon how one interprets suppressed premises. The best thing is to learn about different sorts of reasoning and how they work in good cases, and pick up a feel for some fallacies along the way. At the end of the day, reasoning is an art.

humanrebar 13y ago
What about looking at these lists of fallacies as vocabulary building tools? Perhaps fallacies are all variations on a handful of errors, but it certainly expedites conversation to be able to say, "No, that's just begging the question." Otherwise I would have to waste my breath disagreeing in a more clumsy way.
- mbateman 13y ago
  I agree, there's some value in that. But two things: (1) For begging the question, I think it's more perspicuous to go up a level of abstraction and talk about circular reasoning. And something analogous is true of most of these fallacies. (2) For any real argument you're going to have to spend some breath anyway. It's very rare when you can just stop at, "Isn't that question-begging/circular?", or, "But aren't you equivocating on the word 'freedom'?" This is part of what I mean by saying that reasoning is an art: the classification of an error in reasoning in any particular case will involve nuance and controversy.
KiwiCoder 13y ago
Your last sentence feels unfinished. At the end of the day, reasoning is an art, and an art worth your learning. The question is rather, whether you be capable of learning it? [0] [0] http://www.gutenberg.org/cache/epub/683/pg683.html http://www.gutenberg.org/cache/epub/683/pg683.html
- mbateman 13y ago
  I'm not sure exactly what you mean, but I meant "art" to imply a contrast to something more algorithmic or top-down.
dragonwriter 13y ago
> But I think this sort of enumeration of logical fallacies is of very limited value. In the abstract there are only a few ways that arguments fail, and informal fallacies are all iterations on an extremely similar theme. I think the value of enumerations -- in addition to creating a vocabulary which is often useful in discussing problems with an argument, even if, as you correctly note, exactly which fallacy is most applicable is a subject of interpretation -- is that it helps recognize instances of potentially problematic arguments. Even if, on a certain level of abstraction, a lot of the named fallacies are just different views of a smaller number of real problems, understanding the different manifestations of those common problems helps to recognize them in practice.
- syntaxfree 13y ago
  The limit case of this is lesswrong.com, which is a kind of cultish/Vulcan purge of all fallacies. (Ask around about Roko's basilisk)
  - mchusma 13y ago
    I read both HN and LW, and enjoy both. I've noticed people on HN describing LW as "cultish" a couple of times now, and am constantly suprised by this. Can you shed any light on why you feel this way?
    - pliny 13y ago
      Yudkowsky pushed a virgin into a volcano once.
    - Crito 13y ago
      Roko's Basilisk itself is fine. An interesting idea that is fun to consider. The notion that the idea is too dangerous to discuss publicly is rather cultish. If I understand correctly, such violent reactions to Roko's Basilisk only come from a minority of LW people, but a prominent minority...
    - gjm11 13y ago
      I think there are two main things about LW that strike some people as cultish. (There are others, less important.) Both are less true than they were, say, a year ago. 1. Its distinctive brand of rationalism grew out of this huge long series of blog posts by Eliezer Yudkowsky, conventionally referred to on LW as "The Sequences". So: we have a group of people united by their adherence to a set of writings by a single person -- a mixture of generally uncontroversial principles and more unusual ideas. It's not a big surprise if this reminds some people of religious scriptures and the prophets who write them. 2. The LW culture takes seriously some ideas that (a) aren't commonly taken very seriously in the world at large, and (b) share some features with some cults' doctrines. Most notably, following Yudkowsky, a lot of LW people think it very likely that in the not too distant future the following will happen: someone will make an AI that's a little bit smarter than us and able to improve itself (or make new AIs); being smarter than us, it can make the next generation better still; this iteration may continue faster and faster as the AIs get smarter; and, perhaps on a timescale of days or less, this process will produce something as much smarter than us as we are smarter than bacteria, which will rapidly take over the world. If we are not careful and lucky, there are many ways in which this might wipe out humanity or replace us with something we would prefer not to be replaced by. -- So we have a near-omnipotent, incomprehensible-to-us Intelligence, not so far from the gods of various religions, and we have The End Of The World (at least as we know it), not so far from the doomsdays of various religions. Oh, and LW is somewhat associated with Yudkowsky's outfit, MIRI (formerly the Singularity Institute), and Yudkowsky is on record as saying that the Right Thing to do is to give every cent one can afford to them in order to reduce the probability of a disastrous AI explosion. Again, kinda reminiscent of (e.g.) a televangelist telling you to send him all your money because God is going to wrap things up soon. On the other hand, I do not believe that's his current position. For the avoidance of doubt, I do not myself think LW is very cult-like.
dionidium 13y ago
I think I get what you're saying, but I also think it more usefully applies to the conversations you might have with your philosophy friends. Meanwhile, there seem to be a lot of people on my Facebook wall who could benefit from a list of basic fallacies.
- saraid216 13y ago
  Not really. I'd bet the people on your Facebook wall could benefit from learning how to construct a cogent argument. A list of "DO NOT"s rarely benefits anyone.
SkyMarshal 13y ago
Fwiw, in my experience, simply knowing all or as many as possible logical fallacies better enables one to recognize and discard invalid arguments quicker. It facilitates a more systematic process of elimination. It also helps one realize and appreciate just how difficult it is to construct a correct argument, when you can quickly spot the fallacies in your own assertions. As a result, on a personal level, this has caused me to tend to read, listen, and question more, and assert less.
tokenadult 13y ago
Just so. (My late dad was an industrial engineer whose recreational reading was all about the philosophy of science, so I grew up in a home with logic textbooks on the home bookshelf.) Plenty of incorrect arguments encountered here on Hacker News or elsewhere in cyberspace are incorrect not so much because they fit on a checklist of classical logical fallacies, but rather because they are based on false factual premises. One has to keep seeking knowledge of the world to recognize mistaken premises of arguments. One thing I enjoy about Hacker News is that it includes people from all over the world, including people who have seen places I have never seen, and done things I have never done, so that I can learn new facts here as I read discussion threads. That is often every bit as valuable as knowing lists of logical fallacies, although I did do a bit of teaching of formal deductive logic back in the days when I was a high school debate coach. A previous comment of mine https://news.ycombinator.com/item?id=5835453 https://news.ycombinator.com/item?id=5835453 from a thread recommended in another comment on this thread goes into more detail about why logical fallacies are not the only problems to look for when evaluating claims for truth or falsehood.
turnersr 13y ago
> The best thing is to learn about different sorts of reasoning and how they work in good cases, and pick up a feel for some fallacies along the way. At the end of the day, reasoning is an art. Do you have good suggestions for learning about different sorts of reasoning?
- mbateman 13y ago
  I do but if you're a senior U of C student (as your profile says), I'm sure you've already been very well-served in this regard by your curriculum. :)
  - turnersr 13y ago
    I'm sure others in the community would like to know. I'm also always looking for good books full of ideas and modes of explanation and augmentation. :)

humanrebar 13y ago

What about looking at these lists of fallacies as vocabulary building tools? Perhaps fallacies are all variations on a handful of errors, but it certainly expedites conversation to be able to say, "No, that's just begging the question." Otherwise I would have to waste my breath disagreeing in a more clumsy way.

mbateman 13y ago
I agree, there's some value in that. But two things: (1) For begging the question, I think it's more perspicuous to go up a level of abstraction and talk about circular reasoning. And something analogous is true of most of these fallacies. (2) For any real argument you're going to have to spend some breath anyway. It's very rare when you can just stop at, "Isn't that question-begging/circular?", or, "But aren't you equivocating on the word 'freedom'?" This is part of what I mean by saying that reasoning is an art: the classification of an error in reasoning in any particular case will involve nuance and controversy.

KiwiCoder 13y ago

Your last sentence feels unfinished. At the end of the day, reasoning is an art, and an art worth your learning. The question is rather, whether you be capable of learning it? [0] [0] http://www.gutenberg.org/cache/epub/683/pg683.html http://www.gutenberg.org/cache/epub/683/pg683.html

mbateman 13y ago
I'm not sure exactly what you mean, but I meant "art" to imply a contrast to something more algorithmic or top-down.

dragonwriter 13y ago

> But I think this sort of enumeration of logical fallacies is of very limited value. In the abstract there are only a few ways that arguments fail, and informal fallacies are all iterations on an extremely similar theme. I think the value of enumerations -- in addition to creating a vocabulary which is often useful in discussing problems with an argument, even if, as you correctly note, exactly which fallacy is most applicable is a subject of interpretation -- is that it helps recognize instances of potentially problematic arguments. Even if, on a certain level of abstraction, a lot of the named fallacies are just different views of a smaller number of real problems, understanding the different manifestations of those common problems helps to recognize them in practice.

syntaxfree 13y ago
The limit case of this is lesswrong.com, which is a kind of cultish/Vulcan purge of all fallacies. (Ask around about Roko's basilisk)
- mchusma 13y ago
  I read both HN and LW, and enjoy both. I've noticed people on HN describing LW as "cultish" a couple of times now, and am constantly suprised by this. Can you shed any light on why you feel this way?
  - pliny 13y ago
    Yudkowsky pushed a virgin into a volcano once.
  - Crito 13y ago
    Roko's Basilisk itself is fine. An interesting idea that is fun to consider. The notion that the idea is too dangerous to discuss publicly is rather cultish. If I understand correctly, such violent reactions to Roko's Basilisk only come from a minority of LW people, but a prominent minority...
  - gjm11 13y ago
    I think there are two main things about LW that strike some people as cultish. (There are others, less important.) Both are less true than they were, say, a year ago. 1. Its distinctive brand of rationalism grew out of this huge long series of blog posts by Eliezer Yudkowsky, conventionally referred to on LW as "The Sequences". So: we have a group of people united by their adherence to a set of writings by a single person -- a mixture of generally uncontroversial principles and more unusual ideas. It's not a big surprise if this reminds some people of religious scriptures and the prophets who write them. 2. The LW culture takes seriously some ideas that (a) aren't commonly taken very seriously in the world at large, and (b) share some features with some cults' doctrines. Most notably, following Yudkowsky, a lot of LW people think it very likely that in the not too distant future the following will happen: someone will make an AI that's a little bit smarter than us and able to improve itself (or make new AIs); being smarter than us, it can make the next generation better still; this iteration may continue faster and faster as the AIs get smarter; and, perhaps on a timescale of days or less, this process will produce something as much smarter than us as we are smarter than bacteria, which will rapidly take over the world. If we are not careful and lucky, there are many ways in which this might wipe out humanity or replace us with something we would prefer not to be replaced by. -- So we have a near-omnipotent, incomprehensible-to-us Intelligence, not so far from the gods of various religions, and we have The End Of The World (at least as we know it), not so far from the doomsdays of various religions. Oh, and LW is somewhat associated with Yudkowsky's outfit, MIRI (formerly the Singularity Institute), and Yudkowsky is on record as saying that the Right Thing to do is to give every cent one can afford to them in order to reduce the probability of a disastrous AI explosion. Again, kinda reminiscent of (e.g.) a televangelist telling you to send him all your money because God is going to wrap things up soon. On the other hand, I do not believe that's his current position. For the avoidance of doubt, I do not myself think LW is very cult-like.

dionidium 13y ago

I think I get what you're saying, but I also think it more usefully applies to the conversations you might have with your philosophy friends. Meanwhile, there seem to be a lot of people on my Facebook wall who could benefit from a list of basic fallacies.

saraid216 13y ago
Not really. I'd bet the people on your Facebook wall could benefit from learning how to construct a cogent argument. A list of "DO NOT"s rarely benefits anyone.

SkyMarshal 13y ago

Fwiw, in my experience, simply knowing all or as many as possible logical fallacies better enables one to recognize and discard invalid arguments quicker. It facilitates a more systematic process of elimination. It also helps one realize and appreciate just how difficult it is to construct a correct argument, when you can quickly spot the fallacies in your own assertions. As a result, on a personal level, this has caused me to tend to read, listen, and question more, and assert less.

tokenadult 13y ago

Just so. (My late dad was an industrial engineer whose recreational reading was all about the philosophy of science, so I grew up in a home with logic textbooks on the home bookshelf.) Plenty of incorrect arguments encountered here on Hacker News or elsewhere in cyberspace are incorrect not so much because they fit on a checklist of classical logical fallacies, but rather because they are based on false factual premises. One has to keep seeking knowledge of the world to recognize mistaken premises of arguments. One thing I enjoy about Hacker News is that it includes people from all over the world, including people who have seen places I have never seen, and done things I have never done, so that I can learn new facts here as I read discussion threads. That is often every bit as valuable as knowing lists of logical fallacies, although I did do a bit of teaching of formal deductive logic back in the days when I was a high school debate coach. A previous comment of mine https://news.ycombinator.com/item?id=5835453 https://news.ycombinator.com/item?id=5835453 from a thread recommended in another comment on this thread goes into more detail about why logical fallacies are not the only problems to look for when evaluating claims for truth or falsehood.

turnersr 13y ago

> The best thing is to learn about different sorts of reasoning and how they work in good cases, and pick up a feel for some fallacies along the way. At the end of the day, reasoning is an art. Do you have good suggestions for learning about different sorts of reasoning?

mbateman 13y ago
I do but if you're a senior U of C student (as your profile says), I'm sure you've already been very well-served in this regard by your curriculum. :)
- turnersr 13y ago
  I'm sure others in the community would like to know. I'm also always looking for good books full of ideas and modes of explanation and augmentation. :)