Argument Mapping

I am wrestling with is what are the types of "boxes" that would be in an Argument tree.

I started looking at Stephen Toulmin's work. He divides an argument into Conclusions (which you are trying to prove), Grounds (which can be a fact or evidence) and Warrants (which is logic that connects grounds to conclusions).

It seems to me that Conclusions and Grounds are the same thing (since you can use a conclusion as grounds to prove another conclusion). I also want to distinguish between types of Grounds.

There are sourced facts (i.e. FBI report [link] states that there were 14916 homicides in 2007). Sources, of course, can be disputed.

There are "axioms" that are just stated as fact without proof. This is fine in an argument as long as the axioms are understood and agreed upon.

And there are previously proven conclusions.

I am also playing around with the idea of a "rationale" box, where you could explain the reason that some collection of facts actually led to your conclusion.

Good luck with this group.

And leave a box for insults, tangents, jokes, and sabotaging

That is kind of the point Punkey. There are several people here that I like to debate with (Thomas for example), including a couple of conservatives who sit on the opposite side of the political spectrum (MysteryMan and Finn). I learn things by talking to them and I sharpen my own ideas. But this process is inefficient and the format doesn't help.

There are several things that get in the way of what I am calling "efficient" debate.

- There are tangents (which include humor, previous grudges and friendships and truly interesting but unrelated ideas).
- There are repeats. The discussion format encourages people (including me) to post the same idea over and over again because we know that only the past few posts matter.
- People talk around each other, ignoring points that the other side thinks is important.

I am wondering if these inefficiencies could be addressed in software.

I can mark this one point as really important to my conclusion. This would be prevalent and anyone who wants to debate my conclusion would be encouraged (even obligated) to address the really important point. It would let me structure my argument based on what is important.

This would also also change the "last post wins" principle. Since I am setting the points that are most important (and the time the post was made doesn't matter) I won't feel obligated to repeat the important points to my argument.

By the way... I like the idea of a "tangent" box.

When lawyers present their evidence in court, they need to clear two hurdles with each piece of their evidence: (1) They have to produce the physical object on which their argument relies. (2) They have to persuade the jury that the object actually proves what they say it proves. I can see how a piece of software could do #1. It seems like the same kind of structure that mind-mapping software uses. But #2? Computers have never excelled in the rhetorical and dialectical skills of persuading people. Maybe that's the reason your vision has never been implemented well.

The software I envision will have nothing to do with the reasoning. It won't even evaluate the reasoning. People will do the reasoning. The software I am writing will provide a framework where people will be asked to provide supporting evidence with their own reasoning to each point they make.

This is a point that has long interested me. The way that discussion software works, for example Able2know, effects the types of discussion and the culture of the participants. I contribute to Able2Know and DailyKos. Able2know is based on the last response (no one reads earlier responses which are largely forgotten). DailyKos features the original post (which acts as an article). I enjoy both forums, but they are very different experiences.

My idea is that by predominantly featuring the supporting points and primary evidence, these will be much more important to the discussions. The order that posts are made won't matter.

Have you looked at rbutr? It's not quite what you're doing, but it's in the same neighborhood. Rbutr is a browser plugin that alerts you when the information on a site you are reading has been disputed or rebutted (or both, of course). And if several websites rebut the one you're looking at, it apparrently ranks them according to some vote-counting or link-counting algorithm.

I say "apparrently" because I haven't used the plugin myself yet, I've only seen a link to it on RichardDawkins.net and remember it was interesting. Also, it doesn't act specifically on discussion groups. But this approach, too, will lead you to the evidence, pro and con, on any sufficiently-interesting proposition. Perhaps you can use it for inspiration on your own thing.

A problem could be that some people might refuse to accept that facts really are facts, and the two debaters would then endlessly argue about whose material is obtained from the correct source!

Thanks Thomas. That is really interesting. I will play around with rbutr.

One of the other sites has a really good idea to prevent endless back and forth arguments. To each point you can make only one criticism. And to each criticism the original poster can make one rebuttal. This means there is at most three layers of argument. The key is that each of these points is editable, so that you can make the best criticism or rebuttal possible and adapt it as it is challenged. Readers will have a limited amount to digest to see clear, succint and (hopefully) well-polished arguments on both sides.

My idea is to let contributors self-select a "world-view" (e.g. progressive, conservative or biblical Christian etc.) that readers can filter on. I am not sure whether this is a great idea, or a terrible one. But I will provide tools that will allow anyone to easily compare the differences between the world views.

Since the basic format of the site will be a tree, where arguments descend from earlier arguments which descend from facts which are all human edited, I think seeing the exact point where different world views depart would be really interesting (if this idea works).

People on the internet already do that manually, and the outcome in practice seems to be that everybody ends up talking to like-minded people, blocking out everybody else. Personally I consider that a bad idea. (For extensive details, see Eli Pariser's book The Filter Bubble: How the New Personalized Web Is Changing What We Read and How We Think .)

Thomas,

Everyone blocks out everyone else. Most of us don't want to admit that we are blocking out everyone else.

My experiment is that if I give you a big prominent switch, placed right in your face... one choice is "block out everyone else" and the other is "see other points of view", what would you choose?

I am also interested for myself to see that if the arguments were set up in a tree (where each point is connected to the points it is based on), where would they diverge? Could we get people of different world views to agree on a certain part of the tree?

My hope is that the "filter out other opinions" button will be subversive. It that is what you want to do... then you will have to actively do that. I might filter out different perspectives as long as I can do so without thinking about it. But if I have to push a button, admitting specifically that I am filtering out any different perspectives, that will give me pause.

If people visit the site, flipping between perspectives to see the differences... then I will consider this experiment a success.

I would ignore the switch, because the line-items in my political opinions aren't correlated in the same way they apparently are for most American. It appears to be the case that if I know you're an American who opposes capital punishment, I can make fairly confident predictions of where you stand on progressive vs flat income taxes, the minimum wage vs the Earned Income Tax Credit vs neither, nuclear energy, genetically-engineered crops, Walmart vs Whole Foods, and a lot of other issues that are socially-correlated in America even though there's no logical correlation. I don't fit this pattern --- possibly because I'm from Germany --- and so I wouldn't know who "everyone else" is.


I agree that's an interesting question, and I think the answer would surprise many participants. For example, Noam-Chomsky liberals and Tea-Party libertarians have a lot of overlap about the role of special interests vs. regular citizens in the political system. They're mostly unaware of it because of cultural aversions, but I'm pretty sure the overlap is there.


I wish you luck, but I'm skeptical that this attempt at reverse psychology would work. People's tribal instincts tend to be awfully strong.