Abstract
I propose that we build the SQL code that would facilitate an online forum. This forum would use a relational database to track reasons to agree and disagree with conclusions. It would also allow you to submit a belief as a reason to support another belief (see image 1 below):
Figure 1: Arguments used to support other arguments
Arguments are currently made on websites, in books, and even in videos and songs. It would be powerful to outline all the arguments that agree or disagree with a conclusion and put them on the same page as seen below:
Figure 2: Arguments go from websites, books, songs, videos, into a relational database and are presented with their structure
Having the structure of how all these arguments are used to support each other, could allow us to automatically strengthen or weaken a conclusion's score based on the score of their assumptions.
The purpose of the Idea Stock Exchange is to find ways to give conclusions scores based on the quality and quantity of reasons to agree or disagree with them with an open sourced SQL database.
Pros and Cons are a tried and true method to evaluate a conclusion
Many people, including Thomas Jefferson and Benjamin Franklin advocated making a list of pros and cons, to help them make decisions. The assumption is that the quantity and quality of the reasons to agree or disagree with a proposed conclusion has some bearing as to underlining strength of that conclusion. I wholeheartedly agree.
No one has yet harnessed the power of Pros and Cons in the information age. We can.
However, now that we have the internet, we can crowd source the brainstorming of reasons to agree or disagree with a conclusion.
The only trick is how do you evaluate the strength of each pro or con? Many people suggest putting the strongest pros or cons at the top of the list. Also, if we had enough time we might make a separate list FOR each pro or con.
For instance, FDR had to decide if we should join WWII or not. One pro might be that the German leaders were bad. There were many reasons to support this belief, and this belief was used to support another belief.
Not very many people have enough time to do a pro or con list for each pro or con. But on the internet we keep making the same arguments over and over again. For thousands of years we have been repeating the same arguments that Aristotle and Homer have made. Most of our arguments have been made thousands or millions of times. However no one has ever taken the time to put them into a database, and outline how they relate to each other. We can change this.
I propose that we find algorithms that attempt to promote good conclusions and arguments. This simplest and best method of scoring conclusions is to counting the number of reasons to agree, and subtracting the number of reasons that disagree. Because some arguments are better than other arguments, we should repeat this process for every argument until we reach verifiable data. The following equation represents this plan:
· n = number of “steps” the current arguments is removed from conclusion
We can use algebra to represent each term, and make it look a little more mathematical, with the below formula:
· n: Number of “steps” the current arguments is removed from conclusion
· A(n,i)/n: When n=1 we are looking at arguments that are used directly to support or oppose a conclusion. The 2ndsubscript is “i”. This is used to indicate that we total all the reasons to agree. So when n=1, we could have 5 “i’s” indicating there are 5 reasons to agree. These would be labeled A(1,1), A(1,2), A(1,3), A(1,4), and A(1,5). N on the bottom indicates that reasons to agree with reasons to agree only contribute ½ a point to the overall conclusion. Thus reasons to agree with reasons to agree with reasons to agree would only contribute 1/3 of a point, and so on. If we decide to make the bottom of the equation n x 2, then these would contribute 1/6 of a point. It is obvious that some of their score should contribute to the conclusion scores, because weakening an assumption should automatically weaken all the conclusions built on that assumption. We could continually update n to give reasonable result, or each website could use its own secret sauce.
· D(n,j)/n Ds are reasons to disagree, and work the same as As but the number of reasons to disagree, are subtracted from the conclusion score. Therefore, if you have more reasons to disagree, you will have a negative score. “J” is used, just to indicate that each reason is independent of the other.
· The denominator is the total number of reasons to agree or disagree. This normalizes the equation, resulting the conclusion score (CS) representing the total percentage of reasons that agree. The conclusion score will range between -100% and 100% (or -1 and +1)
The above equation would work very well, if people submitted arguments that they honestly felt supported or opposed conclusions. We could probably find informal ways of making this work, similar to how Wikipedia trusts people, and has a team of editors to ensure quality. However, we could also introduce formal ways to discourage people from using bad logic.
For instance, people could submit that the “grass is green” as a reason to support the conclusion that we should legalize drugs. The belief that the grass is green, will have some good reasons to support it, and may have a high score. At first, to avoid this problem, I would just have editors remove bad faith arguments. But a formalized process would be to have for each argument a linkage score, between -1 and +1 that gets multiplied by the argument’s score that represents the percentage of that argument’s points that should be given to the conclusions points.
I believe the most elegant way to come up with a linkage score would be to just make a new argument, that “a” supports “b”, with all the normal reasons to agree and disagree. However, I also propose the percentage of up-votes compared to the percentage of down-votes and other good idea promoting algorithms below.
Also, without editors, you would run into the problem of duplication. If we had this at the time of the Gulf Wars, people could have been submitting the belief that Saddam Hussein was a bad person as a reason to support the belief that we should go to war. People would submit the belief that we don’t go to war with everyone who is bad, as a way of weakening the linkage between this conclusion and argument. But someone might also submit the belief that he was “evil”. How much is the world “evil” and “bad” the same thing? Is Evil just a worse kind of bad? These questions could be quantified, if for each argument, we brainstormed a list of “other ways of saying the same thing”. Of course we would use all of our algorithms to determine to what degree they are the same thing. If we determine that two items are 85% the same thing, then when both of them are used as reasons to support the same thing, then they would only count as 1.15x their two scores, not 2x.
Examples
We might be arguing the conclusion that “It was good for us to join WWII.” Someone may submit the argument that “Nazis were doing bad things” as a reason to support the conclusion about entering the war. The belief that Nazis were doing bad things might already have a score. Let’s suppose that this idea score has a high ranking of 99%. This might be awarded a linkage score of 90% (as a reason to support the conclusion that we should have gone to WWII). In this situation it would contribute 0.495 points (0.99 X 0.5) to the conclusion score for the beliefs that “It was good for us to join WWII”. Someone else might submit a belief that “Nazis were submitting wide scale systematic genocide” as a reason to support the belief that “It was good for us to go to WWII”. Because we don’t go to war with every country that “does bad things”, we would assume that this linkage score would be higher, perhaps a 98%.
For example the belief that Nazi Germany leaders were evil, is a belief with many argument to support it. However it can also be used as an argument to support other conclusions, such as the belief that it was good of us to join WWII.
Assumptions
· Reason ≝ Belief used to support another belief(For example the belief that Nazi Germany leaders were evil, is a belief with many argument to support it. However it can also be used as an argument to support other conclusions, such as the belief that it was good of us to join WWII).
· Good Belief ≝ Good Reasons to Agree > Good Reasons to disagree
· Bad Belief ≝ Good Reasons to Agree > Good Reasons to disagree
· Great Belief ≝ Good Reasons to Agree >> Good Reasons to disagree
· Terrible Belief≝ Good Reasons to Agree << Good Reasons to disagree
