Surely someone already tried this: automated, iterative, argument builder

The Something Awful Forums > Discussion > Debate & Discussion > Surely someone already tried this: automated, iterative, argument builder

Mrs. Wynand: Nov 23, 2002; DLT 4EVA

Basically, just some website where you can walk people through an argument for some specific clam - e.g. "vaccines do not cause autism", or "you should/should not vote for _____", or "TNG is objectively the best Star Trek". The subject isn't really important here.

Briefly put, the idea is:

An opponent of some claim begins by selecting among multiple counter-points to that claim, sorted by popularity (in terms of how often they are selected).
If the counter-point you had in mind is not listed, you can enter it now and it will be made available to future participants. The exercise ends until someone bothers responding to it.
Otherwise, once you selected a counter-point, a response to your counter-point is shown to you. This response is chosen by way of an effectiveness score (explained shortly).
If you find the response convincing, you select it as such,and all counter-arguments that brought you to this outcome have their effectiveness score increased. The exercise ends.
If you do not find the response convincing, you must select among multiple counter-points explaining why, same as in the first step, and as was the case there, you may enter a new counter if the one you had in mind isn't listed.
You are now shown another response, which will be either a response to your original counter-point, or the one you've just chosen, depending on which one has a higher effectiveness score.
The process repeats (increasing the search-space of replies each iteration) until you've either marked a response as convincing, or given up in frustration, or you run out of replies to respond to.
If you did give up (or, and this is really very unlikely to happen often, you actually ran out of argument/replies) all responses you've viewed up to that point have their effectiveness score decreased. Arguments with a sufficiently low negative score are removed from the system.

There's a lot of detail-tweaking that could be figured out on top of that, like, some amount of moderation is going to be required to keep bots out, remove what are essentially duplicate responses or counter-arguments, controlling creepy little groups trying to game the system for whatever reason and so forth. The actual scoring is probably going to have to be a different too, as the above +1/-1 system is going to greatly favor newly created coutners/replies (assuming they start with score=0). Maybe the failure score can be separate and expressed as standard-deviations from the mean failure rate or something like that...

It really strikes me as something some techno-libertarian somewhere is going to have tried (because software is the answer to all of the world's problems, of course!), but I can't really find any references to such a thing... probably because i'm not exactly sure what to look for.

That aside, this is something that's been in the back of my mind ever since I heard how Al Gore basically crafted his Inconvenient Truth presentation over the years. He basically just iterated over it over many many years - doing his speil, engaging the audience after, cutting out what ended up not really working and addressing new concerns that his opponents kept bringing up, one by one, point by point, in a long, painstaking process.

# ? Apr 22, 2014 23:34

Adbot: ADBOT LOVES YOU

# ? Apr 25, 2024 00:06

mdemone: Mar 14, 2001

Victor, is that you?!?

# ? Apr 23, 2014 01:35

falcon2424: May 2, 2005

This seems a bit like the FAQ-format that's used at Slatestar: http://slatestarcodex.com/2013/10/20/the-anti-reactionary-faq/ and Prussian: http://www.skepticink.com/prussian/2014/03/31/the-anti-racialist-q-a/

Basically, they try to reduce their opponent's arguments into specific-claims. Then they present a strong-as-possible formulation of the stance. And then they address it.

What I like is that the FAQ breaks things down into points that are specific enough to be debated. A lot of other debates come off as booing or cheering conclusions. The problem is that a lot of people don't like putting their opponents arguments into an ideally-strong form.

A CYOA format would probably end up looking like a FAQ, except that most content would be hidden on most paths.

# ? Apr 23, 2014 01:41

Mo_Steel: Mar 7, 2008; Let's Clock Into The Sunset Together; Fun Shoe

Mr. Wynand posted:

Basically, just some website where you can walk people through an argument for some specific clam - e.g. "vaccines do not cause autism", or "you should/should not vote for _____", or "TNG is objectively the best Star Trek". The subject isn't really important here.

Briefly put, the idea is:

An opponent of some claim begins by selecting among multiple counter-points to that claim, sorted by popularity (in terms of how often they are selected).

If the counter-point you had in mind is not listed, you can enter it now and it will be made available to future participants. The exercise ends until someone bothers responding to it.

Otherwise, once you selected a counter-point, a response to your counter-point is shown to you. This response is chosen by way of an effectiveness score (explained shortly).

If you find the response convincing, you select it as such,and all counter-arguments that brought you to this outcome have their effectiveness score increased. The exercise ends.

If you do not find the response convincing, you must select among multiple counter-points explaining why, same as in the first step, and as was the case there, you may enter a new counter if the one you had in mind isn't listed.

You are now shown another response, which will be either a response to your original counter-point, or the one you've just chosen, depending on which one has a higher effectiveness score.

The process repeats (increasing the search-space of replies each iteration) until you've either marked a response as convincing, or given up in frustration, or you run out of replies to respond to.

If you did give up (or, and this is really very unlikely to happen often, you actually ran out of argument/replies) all responses you've viewed up to that point have their effectiveness score decreased. Arguments with a sufficiently low negative score are removed from the system.

There's a lot of detail-tweaking that could be figured out on top of that, like, some amount of moderation is going to be required to keep bots out, remove what are essentially duplicate responses or counter-arguments, controlling creepy little groups trying to game the system for whatever reason and so forth. The actual scoring is probably going to have to be a different too, as the above +1/-1 system is going to greatly favor newly created coutners/replies (assuming they start with score=0). Maybe the failure score can be separate and expressed as standard-deviations from the mean failure rate or something like that...

It really strikes me as something some techno-libertarian somewhere is going to have tried (because software is the answer to all of the world's problems, of course!), but I can't really find any references to such a thing... probably because i'm not exactly sure what to look for.

That aside, this is something that's been in the back of my mind ever since I heard how Al Gore basically crafted his Inconvenient Truth presentation over the years. He basically just iterated over it over many many years - doing his speil, engaging the audience after, cutting out what ended up not really working and addressing new concerns that his opponents kept bringing up, one by one, point by point, in a long, painstaking process.

The closest non-programmatic example I can think of is the massive video project "How It All Ends" about human caused global climate change. A science teacher put together hours upon hours of video explaining the topic, the framing, possible solutions, as well as a number of videos explicitly focused on pre-addressing common criticisms of his original video on the topic.

It's an interesting idea; I'm envisioning something in a similar form to Akinator where users suggest improvements and it weights criteria to find an answer, perhaps?

# ? Apr 23, 2014 01:46

Mrs. Wynand: Nov 23, 2002; DLT 4EVA

mdemone posted:

Victor, is that you?!?

I know right?

# ? Apr 23, 2014 01:50

Mrs. Wynand: Nov 23, 2002; DLT 4EVA

Anyway, yes the point basically is that sort of "browbeating them with arguments" that can sort of work sometimes with a sufficiently interested audience and a heaping ton of effort, but combined with the sort of creepy crowd-pleasing/mind-reading you can do out of thousands of bits of human input, as with 20-question bots and reddit.

# ? Apr 23, 2014 05:04

Judakel: Jul 29, 2004; Can't post for 9 years!

mdemone posted:

Victor, is that you?!?

Victor was a very sophisticated prototype of a similar concept. We had to terminate since there was not much that could be done to correct its flaws.

# ? Apr 23, 2014 05:06

Mrs. Wynand: Nov 23, 2002; DLT 4EVA

I've actually always sort of looked Victor's crazy-posts with a certain "there by the grace of god go I" attitude because the general thrust of his ideas definitely appealed to me, until I actually start to read the full posts, eyes slowly widening in horror.

So anywho, I take it the answer to my OP question is basically no, nobody has quite done this, though people certainly have tried this style of argument before, not necessarily with great success... (To be perfectly honest, all those FAQs linked so far sort of creep me out... I think it's the sheer volume of material coming from a single person's steel-eyed conviction that they are right beyond the shadow of a doubt, if only you could just see it, here let me spell it out for you in excruciating detail... I'm hoping democratizing the process can alleviate this somewhat... Al Gore certainly didn't come off as creepy, but then, we can't all be the charismatic sexual tyranosaurus that is Al Gore)

# ? Apr 23, 2014 19:07

Judakel: Jul 29, 2004; Can't post for 9 years!

Mr. Wynand posted:

I've actually always sort of looked Victor's crazy-posts with a certain "there by the grace of god go I" attitude because the general thrust of his ideas definitely appealed to me, until I actually start to read the full posts, eyes slowly widening in horror.

So anywho, I take it the answer to my OP question is basically no, nobody has quite done this, though people certainly have tried this style of argument before, not necessarily with great success... (To be perfectly honest, all those FAQs linked so far sort of creep me out... I think it's the sheer volume of material coming from a single person's steel-eyed conviction that they are right beyond the shadow of a doubt, if only you could just see it, here let me spell it out for you in excruciating detail... I'm hoping democratizing the process can alleviate this somewhat... Al Gore certainly didn't come off as creepy, but then, we can't all be the charismatic sexual tyranosaurus that is Al Gore)

Victor never discussed things, he just wanted to win arguments and not abandon the tenets of his thinking. A lot of people do that, it is that they usually have the courtesy to helldump instead of endlessly argue for their points.

# ? Apr 23, 2014 19:40

ducttape: Mar 1, 2008

Mr. Wynand posted:

Basically, just some website where you can walk people through an argument for some specific clam - e.g. "vaccines do not cause autism", or "you should/should not vote for _____", or "TNG is objectively the best Star Trek". The subject isn't really important here.

Briefly put, the idea is:

An opponent of some claim begins by selecting among multiple counter-points to that claim, sorted by popularity (in terms of how often they are selected).

If the counter-point you had in mind is not listed, you can enter it now and it will be made available to future participants. The exercise ends until someone bothers responding to it.

Otherwise, once you selected a counter-point, a response to your counter-point is shown to you. This response is chosen by way of an effectiveness score (explained shortly).

If you find the response convincing, you select it as such,and all counter-arguments that brought you to this outcome have their effectiveness score increased. The exercise ends.

If you do not find the response convincing, you must select among multiple counter-points explaining why, same as in the first step, and as was the case there, you may enter a new counter if the one you had in mind isn't listed.

You are now shown another response, which will be either a response to your original counter-point, or the one you've just chosen, depending on which one has a higher effectiveness score.

The process repeats (increasing the search-space of replies each iteration) until you've either marked a response as convincing, or given up in frustration, or you run out of replies to respond to.

If you did give up (or, and this is really very unlikely to happen often, you actually ran out of argument/replies) all responses you've viewed up to that point have their effectiveness score decreased. Arguments with a sufficiently low negative score are removed from the system.

There's a lot of detail-tweaking that could be figured out on top of that, like, some amount of moderation is going to be required to keep bots out, remove what are essentially duplicate responses or counter-arguments, controlling creepy little groups trying to game the system for whatever reason and so forth. The actual scoring is probably going to have to be a different too, as the above +1/-1 system is going to greatly favor newly created coutners/replies (assuming they start with score=0). Maybe the failure score can be separate and expressed as standard-deviations from the mean failure rate or something like that...

It really strikes me as something some techno-libertarian somewhere is going to have tried (because software is the answer to all of the world's problems, of course!), but I can't really find any references to such a thing... probably because i'm not exactly sure what to look for.

That aside, this is something that's been in the back of my mind ever since I heard how Al Gore basically crafted his Inconvenient Truth presentation over the years. He basically just iterated over it over many many years - doing his speil, engaging the audience after, cutting out what ended up not really working and addressing new concerns that his opponents kept bringing up, one by one, point by point, in a long, painstaking process.

The biggest problem that I can think of is that it seems susceptible to organized trolling. People 'arguing' against their actual beliefs, only to give up at an argument they want to strengthen (or to get rid of arguments that they don't like), or arguing their side but deliberately selecting the worst counter-arguments as the 'most convincing' to discredit the whole thing.

# ? Apr 23, 2014 19:52

Mrs. Wynand: Nov 23, 2002; DLT 4EVA

ducttape posted:

The biggest problem that I can think of is that it seems susceptible to organized trolling. People 'arguing' against their actual beliefs, only to give up at an argument they want to strengthen (or to get rid of arguments that they don't like), or arguing their side but deliberately selecting the worst counter-arguments as the 'most convincing' to discredit the whole thing.

Yes, as mentioned, some oversight for dealing with "gaming the system" will be needed, as is always the case with anything involving internet points. Hiding said internet points might be a start actually... I guess it's sort of difficult to do this manually using trusted humans (which is usually my go-to answer when it comes to any sort of internet community oversight, one need only look at the difference between the SA approach to moderation and reddit to see why) - even with infinite resources it would be hard to tell the genuinely willfully ignorant from the organized trolls. So probably the more promising approach is along the lines of not making the target worth the effort. I mean there isn't any real value in any of this anyway - it's just a fun little exercise in internet-debating, the most it can hope to deliver is a handful of interesting (though not necessarily useful or reliable) insights, that's all.

I suppose in theory it could be used to fine-tune an actual political argument or campaign, though in that case you could probably be assed to perform the exercises from a statistically representative pool of pre-selected participants as when doing actual market research, not internet strangers (which are a notoriously poor sample anyway - it really would be a non-starter to try anything serious with them)

# ? Apr 23, 2014 20:59

Ghost of Reagan Past: Oct 7, 2003; rock and roll fun

Is this the kind of thing you're thinking of? Sure sounds a lot like it.

# ? Apr 25, 2014 01:02

KernelSlanders: May 27, 2013; Rogue operating systems on occasion spread lies and rumors about me.

It sounds like what you want is a choose your own adventure book for political debating. I can see that working great as satire, but would probably just piss off anyone with half a brain. It might be fun to write, though.

# ? Apr 25, 2014 01:43

emfive: Aug 6, 2011; Hey emfive, this is Alec. I am glad you like the mummy eating the bowl of shitty pasta with a can of 'parm.' I made that image for you way back when. I’m glad you enjoy it.

If you get anything put together as a prototype be sure and test it on Eripsa.

[edit] oh and also something about the way you worded the OP reminded me of that loon designing (patenting?) the video game about making decisions that either lead to "mankind uplifted" or "godless socialism", plus he took lots of cheesy pictures of girls in cowboy boots.

vvvvv "exalted" was the word I couldn't remember

emfive fucked around with this message at 04:02 on Apr 25, 2014

# ? Apr 25, 2014 02:15

Malcolm: May 11, 2008

I remember that patent diagram, it was rather entertaining in a tube-cube sort of way.

Well, I had to go find it, still makes me laugh. https://www.google.com/patents/US20090017886

# ? Apr 25, 2014 02:30

Mrs. Wynand: Nov 23, 2002; DLT 4EVA

Ghost of Reagan Past posted:

Is this the kind of thing you're thinking of? Sure sounds a lot like it.

That is very close! It's a little thin in both quantity and quality, but I guess that's not a problem easily avoided either way. I do think it could benefit from more focused guidance. Also it seems a lot of the popular points are more just people agreeing with each-other, which their format sort of encourages...

Either way though, thanks for posting this - it's very much the sort of thing I was hoping to find with this thread.

# ? Apr 25, 2014 05:05

Kilty Monroe: Dec 27, 2006; Upon the frozen fields of arctic Strana Mechty, the Ghost Dads lie in wait, preparing to ambush their prey with their zippin' and zoppin' and ziggy-zoop-boppin'.

What happens with this is that as discussions on a topic get forked into individual talking points, you get arguments on one point that aren't compatible with arguments being made about another point. Positions ought to logically follow from first principles, and you can't just make a substitute for actual thinking so easily.

# ? Apr 25, 2014 07:14

duck monster: Dec 15, 2004

Malcolm posted:

I remember that patent diagram, it was rather entertaining in a tube-cube sort of way.

Well, I had to go find it, still makes me laugh. https://www.google.com/patents/US20090017886

He did a post on Gamasutra and it confused the *poo poo* out of everyone

http://www.gamasutra.com/blogs/DrEl...nologyGames.php

# ? Apr 25, 2014 11:30

Mrs. Wynand: Nov 23, 2002; DLT 4EVA

Kilty Monroe posted:

What happens with this is that as discussions on a topic get forked into individual talking points, you get arguments on one point that aren't compatible with arguments being made about another point. Positions ought to logically follow from first principles, and you can't just make a substitute for actual thinking so easily.

What do you mean? The arguments all follow a directed graph, each topic isolated. Like you don't re-use a talking point just because it sounds similar - I mean, as you say, it only really makes sense in context.

# ? Apr 25, 2014 17:01

Phyzzle: Jan 26, 2008

Kilty Monroe posted:

What happens with this is that as discussions on a topic get forked into individual talking points, you get arguments on one point that aren't compatible with arguments being made about another point.

AKA, arguing like a lawyer. List as many separate arguments as reasonable in hopes that one sticks.

# ? Apr 25, 2014 17:28

Krotera: Jun 16, 2013; I AM INTO MATHEMATICAL CALCULATIONS AND MANY METHODS USED IN THE STOCK MARKET

Hey, this seems like a cool idea.

I'd hack something out if I felt like I could reasonably commit to hosting/maintaining it -- as it stands, though, the most I'd be able to do is write some code.

# ? May 7, 2014 05:27

OwlBot 2000: Jun 1, 2009

An Israel/Palestine one could net you a lot of money.

# ? May 7, 2014 05:34

rudatron: May 31, 2011; by Fluffdaddy

D&D goon project: object-oriented shitposting.

# ? May 7, 2014 11:48

Spoke Lee: Dec 31, 2004; chairizard lol

Can I get some background on Victor? It's not ringing any bells.

# ? May 11, 2014 16:54

Mrs. Wynand: Nov 23, 2002; DLT 4EVA

He's just some poster (that hasn't actually posted since 2012 as far as I can see). It's actually kind of hard to describe his exact brand of nuttyness. Basically imagine the OP, but easily 10 times longer and no paragraph breaks and an unflinching, steely-eyed, utterly inexorable belief that it is not only an unbelievably fantastic idea but also the cure to some combination of (or just, all) of the world's problems.

Also imagine a reply to every single person responding to the OP that is of similar size and conviction. And the hardest to imagine part: this weird veneer of rationality that pulls otherwise reasonable people into engaging in the nuttiness.

# ? May 11, 2014 18:19

Spoke Lee: Dec 31, 2004; chairizard lol

Ohh, those are fun. Kinda like those people that instead of letting an idea go in the face of a million reasons why it's a bad idea, they either deny or accept your criticism as valid and just come up with some convoluted workaround?

Spoke Lee fucked around with this message at 21:35 on May 11, 2014

# ? May 11, 2014 21:32

Judakel: Jul 29, 2004; Can't post for 9 years!

Spoke Lee posted:

Ohh, those are fun. Kinda like those people that instead of letting an idea go in the face of a million reasons why it's a bad idea, they either deny or accept your criticism as valid and just come up with some convoluted workaround?

Proto-bears. They were around before the great flood and that is where current bears come from you see.

# ? May 12, 2014 00:08

Captain_Maclaine: Sep 30, 2001; Every moment I'm alive, I pray for death!

Spoke Lee posted:

Can I get some background on Victor? It's not ringing any bells.

He was a colossal idiot who was staggering obstinate about the dumb poo poo he believed, and argued, ad nauseum. Among his many peculiarities was he coded up a post-aggregator to handle his arguing rather than, you know, actually addressing points made against him. Allow me to bring up the classic summation of what arguing with Victor was like:

"hurrrr[super posted:

2[/super]"]
A better analogy would be if someone walks into a championship tournament, says "GEE I THINK I MAY HAVE TRANSCENDED THE UNDERSTANDING OF SOME OF YOU GRANDMASTERS HERE, WANT TO JOIN MY NEW SCHOOL OF CHESS STRATEGY?", then loses by scholar's mate twice in the first round.

This person then refuses to leave his seat, claiming that he needs additional proof that the queen in f7 actually ontologically exists before he will admit defeat, and that the rules of the CHESS ESTABLISHMENT were unfairly biased against him by disallowing the possibility of his king being able to leapfrog pieces.

Then he pulls out an ancient shopping list from 1905 and claims that "1. Eggs" means 'The King', "2. Butter" means 'can', and "3. Milk" means 'leapfrog'. This is admissible evidence for his case because he has lived according to the dictates of this list since he was a teenager, and it has drastically improved his quality of life. When the referees tell him that this makes no loving sense, he drags them into a three hour debate over the precise meaning of the words 'makes', 'no', 'loving', and 'sense'.

When people point out that there is more than enough evidence to suggest his list is just a scrap of paper from some long-dead housewife's purse, he rather proudly points out how close-minded they are in dismissing outright the possibility that the list was in fact a secret coded message on the best way to live life, originally formulated by Atlanteans and passed down through the ages disguised as everyday documents. After all, if one starts with the presupposition that such a document exists, then it would be very fair to argue that it is indeed in the form of his shopping list.

Never mind that his previous interpretations of the list led to three convictions and time served for robbery, hate crimes, and murder. These were just unfortunate misinterpretations on his part of the list's true intentions, he says. The list itself is blameless. In fact, the Atlanteans deliberately obfuscated the true meaning of the list in this way, so that it would require multiple failed misinterpretations before one would happen across its TRUE meaning, and in doing so appreciate it all the more.

In fact, he does have some evidence to back up his claims. Why, just last week during his daily meditation on the list, he felt it telling him that something good was about to happen in his future. And yesterday, wouldn't you know it, he found a twenty dollar note on the sidewalk! Evidence of the list's prophetic powers if I ever saw one. And believe him, he has many more stories where that came from.

By now, the debate has splintered off into innumerable tangents, with the one man against literally every other player and referee present at the tournament. Finally, he graciously accepts the possibility of defeat in some of the myriad topics now being covered. OK, maybe the tallest player doesn't always get to go first. Fine, I will concede that there isn't much evidence to support my third-invisible-knight hypothesis. But that's all irrelevant. What he wants to concentrate on, and what nobody has yet been able to disprove, he adds, is the ability of the king to leapfrog over other pieces.

The argument drags on for weeks. Finally, one afternoon, the beet-faced referee exhausts his last reserves of decency and throws his arms up in frustration and despair. "YOU loving RETARD, HOW CAN YOU LAY CLAIM TO KNOWING ANYTHING ABOUT CHESS STRATEGY WHEN YOU DON'T EVEN GRASP THE MOST BASIC RULES!?" He shouts, just as a new entrant walks through the door. "I'm sorry," replies the man calmly, "I simply cannot discuss the rules of chess with such an 'official' if you insist on using such strong and uncouth language. Please retract your insults or I will be forced to plug my ears whenever you say anything from now on."

Seeing only this last exchange, the new entrant pipes up. "He's right, you know. If he did something wrong, then you as the referee have every right to tell him he is so, but it should be done with a patient and thorough explanation of the details of his error. Hurling ridicule at him solves nothing and won't change anyone's mind."

The lazy eye of the retarded List-following, King-leapfrogging man twitches almost unnoticeably, as he cranes his head towards the source of this new voice. A welcoming smile cracks, inch by beaming inch, across his face. He licks his lips. He clears his throat.

"So glad to know decent people like you still value a polite discussion. Care for a game?"

# ? May 12, 2014 01:07

Adbot: ADBOT LOVES YOU

# ? Apr 25, 2024 00:06

Stanos: Sep 22, 2009; The best 57 in hockey.

Already been done, OP

http://thomasfriedmanopedgenerator.com/about.php

# ? May 12, 2014 01:21

The Something Awful Forums > Discussion > Debate & Discussion > Surely someone already tried this: automated, iterative, argument builder