Talk:Wikidebate

Socratic Methods

Thanks for providing this interesting learning project. Perhaps students who are not yet ready to present arguments in the formalized style of the Wikidebate platform would like to sharpen their skills the old fashioned way, using Socratic Methods. Mention of the courses on Socratic Methods and Practicing Dialogue in the description of the Wikidebate project might be helpful. Thanks! --Lbeaumont (discuss • contribs) 12:29, 25 November 2016 (UTC)

Y Done --Sophivorus (discuss • contribs) 14:17, 25 November 2016 (UTC)

Recognizing Fallacies

The course on Recognizing Fallacies can help debaters more clearly identify and describe unsound arguments. It may be helpful to mention this resource in the section on "How to object effectively." Thanks!

Y Done Please feel free to edit the page yourself in the future! --Sophivorus (discuss • contribs) 14:17, 25 November 2016 (UTC)

Explicate the Syllogism

As you point out, a sound argument is of the form: evidence -> premise -> conclusion. In natural language certain words or phrases, called “conclusion-indicators” are often used to introduce the conclusion of an argument. See: Taming Wild Fallacies

Perhaps it would be useful to provide a “macro” that marks the premise, conclusions, and supporting evidence implied by these indicator words. For example, the sentence

Because you promised to wash your car every Saturday, and today is Saturday we can conclude you have washed your car.

Would perhaps be marked as:

Because you promised to wash your car every Saturday, and today is Saturday we can conclude you have washed your car.

Where I have (quite arbitrarily) used bold to mark indicator words, underline to mark evidence, and bold underline to mark the conclusion. (Background colors might provide a more effective mark.)

The book The Logic of real arguments by Alec Fisher introduces such a markup scheme (better suited to pen and paper than text macros) that might be inspiring here.

I recognize this is an ambitious proposal that is only half baked, but it might inspire some thought and more dialogue if you are at all interested.

@Lbeaumont: I think that developing a markup macro is too difficult, because it implies a lot of natural language analysis, which is a particularly difficult thing to do. Marking "conclusion indicators" and other logical terms may be doable, but marking propositions is way, way more difficult, because they may take a myriad different forms. Natural language analysis is one of the hardest branches of computer programming, as far as I'm aware, so the benefits must be huge in order to justify the effort. Also, whatever macro we develop will only work for English, and I hope to extend Wikidebate to Spanish and other languages soon. However, a much more doable (and multilingual) approach would be to create guidelines for a markup system, and then mark up the arguments manually. What do you think? --Sophivorus (discuss • contribs) 19:22, 26 November 2016 (UTC)

What I have in mind is much simpler than natural language analysis. For example, there might be a "macro" called "because" that would take two arguments, the premise and the conclusion. So, using the above example, the syllogism would be written as because(you promised to wash your car every Saturday AND today is Saturday, you have washed your car) (The "Because function" takes two arguments, the premise and the conclusion, each separated by a comma. The "AND" is another keyword in the syntax). I hope this clarifies and simplifies the proposal. --Lbeaumont (discuss • contribs) 00:42, 27 November 2016 (UTC)

Perhaps these two ideas can be beneficially melded. Consider creating "markup primitives" such as: evidence(), premise_indicator(), premise(), conclusion_indicator(), and conclusion(). Each would take a text argument as input and output marked up text according to some simple markup convention. This could help clarify the structure of the argument to the reader, as well as the author. --Lbeaumont (discuss • contribs) 03:49, 2 December 2016 (UTC)

@Lbeaumont: Maybe some optional parameters on the Template:Argument, so that it can be called like so:

{{Argument
|premise1 = All humans are mortal
|premise2 = Socrates is human<ref>Evidence</ref>
|conclusion = Socrates is mortal
}}

I think the "evidence" is more fit for footnotes than extra parameters. What do you think? --Sophivorus (discuss • contribs) 12:53, 3 December 2016 (UTC)

Technically the right conclusion would have been "you have promised to wash your car today". AP295 (discuss • contribs) 17:33, 4 April 2021 (UTC)

A different take

This edit kind of rescopes a debate to have a third "no one knows" option. Kind of an interesting change. —Justin (koavf)❤T☮C☺M☯ 16:15, 3 October 2019 (UTC)

@Koavf Indeed, several debates are evolving towards more numerous and precise sections, rather than just yes/no dichotomies (latest example). I think this is a most desirable development! Sophivorus (discuss • contribs) 16:21, 15 March 2021 (UTC)

Agreed. To visualize it branching somehow could be good. I also think that the way that these should be drafted are that they have individual arguments being refined and responded to (possibly with branching responses) rather than piles of new arguments that are weak. (Just thinking out loud.) —Justin (koavf)❤T☮C☺M☯ 17:56, 15 March 2021 (UTC)

@Koavf Not sure I understand your suggestion. Could you explain a bit more, maybe with an example? PS: please check out the new topic below! :-) Sophivorus (discuss • contribs) 16:30, 16 March 2021 (UTC)

Should we change the title convention?

Currently, wikidebates are titled as questions (e.g. Should cannabis be legal?). However, I'm thinking that maybe we should rename them after the main topic of the debates, and move them as subpages of the main wikidebate page (e.g. Wikidebate/Cannabis, Wikidebate/Capital punishment, etc). The main reason for this is that as debates grow, they tend to go beyond simple yes/no answers. For instance, the debate about cannabis already has 5 different possible answers, so the current title is no longer appropriate. More complex debates are likely to develop even more possible answers. I think the current title convention doesn't scale and we should change it. However, on the other hand, I suspect much of the current traffic comes from users searching the current titles on search engines, as can be seen by comparing with the traffic of the Spanish version. Changing them to more generic ones may reduce traffic. Thoughts? Sophivorus (discuss • contribs) 15:44, 16 March 2021 (UTC)

@Sophivorus: I like the question format. A question doesn't have to have a yes-or-no answer. This might sound strange, but I feel somewhat put upon when the media or a website insists "these are the difficult issues" or "this is what is important". They build this nebulous ethos of social priorities and issues. There always seems to be an odd, tacit insistence behind it, as though there's no way to settle such issues except to pick a side and tow the line. On the other hand, a question is just a question. We can often settle them (or come up with better questions) using measured dialectic. I'll think of a better way to explain this when I have more time. AP295 (discuss • contribs) 12:15, 3 April 2021 (UTC)

To sum up, it would be better to present these debates as a set of queries than as a set of imperatives. To present them as imperatives is presumptuous, and it seems to put people in the wrong frame of mind. An imperative is a problem that must be dealt with. It primes the reader's flight-or-fight response. An amygdala hijack. At that point, dialectic becomes difficult. I'm not saying that this is or was your intention. Unfortunately, it seems that we're conditioned to approach political issues in this manner. It's a flaw that has been introduced into our culture by the media. Let's not perpetuate it though. AP295 (discuss • contribs) 12:56, 3 April 2021 (UTC)

Thinking about it a bit more, I'm beginning to have misgivings about the entire Wikidebate format as a whole. Certainly, Wikiversity can and should be a venue for debate but the current itemized, unattributed format of a debate seems to undermine if not defeat the purpose. For example, see my reply a couple of sections below. AP295 (discuss • contribs)

Collaboration with the Canonizer

The canonizer is a separate on-going project that shares a similar conceptual architecture with Wikidebate. It seems useful to explore collaborations with the canonizer with the intent of improving, or perhaps even merging, both projects. Thanks! --Lbeaumont (discuss • contribs) 11:40, 3 April 2021 (UTC)

Next steps ideas – using free argument map software

First of all, this is a really nice project draft. I had a related vision many years back but haven't had time to get to it so far. Instead of outlining it here (I hoped to make it a FOSS project linked with or part of Wikimedia), I'll just link this Commons cat that I recently created which has subcategories for the best options for structured online debate and conclusion-making currently available.

To make it short, Kialo is currently the best such system but it's not open source – however it seems like they're currently considering making their software open source (if they do so and/or specify a CCBY content license, maybe it would be possible to embed abridged debate trees from there on the relevant Wikipedia articles/sections or sth similar).

While I'm interested in input on any of these topics, I'm posting here to make the following proposal: What do you think about using the free python software Arguman for this project (or possibly a spin-off)? Is it possible to integrate it into Wiki-pages for example? Debates already here could be converted into Arguman syntax and then people could continue with that.

If it is and/or some devs here are interested in implementing this, it may be possible that Wikidebate becomes a standalone Mediawiki project (a Wikimedia sister project like Scholia) rather than an unknown Wikiversity subproject. What do you think? (@Sophivorus @Lbeaumont @AP295 etc) Prototyperspective (discuss • contribs) 16:29, 4 June 2023 (UTC)

What is gained by annotating a debate in such a way, other than a suggestive effect on the reader? I have caution about this format of itemized or sequential argument, which is also encouraged in this resource: "Split distinct arguments ― If one argument is essentially two, split them apart. Keeping them separate will enrich the debate, allow others to object to each argument independently and prevent unnecessary confusion." Suppose P(A|B),P(A|C),P(A|D),P(A|E) are all relatively small but P(A|B,C,D,E)=.99, Person 1 provides evidence B, Person 2 concludes that P(A|B) is small and assigning them "winner" suggests that P(A|B) = P(A), and so on until P(A|B,C,D,E) is reduced to merely the prior P(A). Indeed, much less "confusing" without all that evidence to clutter up the picture. AP295 (discuss • contribs)

Hi AP295, sorry for the delay. I fully agree with your analysis about the limitations of the structures of wikidebates. Wikidebates are not perfect (and Kialo and other systems probably aren't either). Partly because of these limitations, I've been reluctant on developing wikidebates much further. Better structures can surely be constructed, but usually at the cost of increased complexity for the developers and, crucially, for the end users. But wikidebates need not be perfect, only useful and thought provoking, I guess. That persuades me to leave them online, and see what time and people make out of them. Hopefully something more good than evil. Kind regards, Sophivorus (discuss • contribs) 19:54, 27 September 2023 (UTC)

Thanks, Sophivorus.

You are a creative programmer. Thanks alot for creating this. Elominius (discuss • contribs) 18:58, 26 September 2023 (UTC)

@Elominius Thanks for your thanks! I hope you find some debates interesting! Sophivorus (discuss • contribs) 19:56, 27 September 2023 (UTC)

Not only that, but I have created my own wikidebate! Elominius (discuss • contribs) 23:14, 30 September 2023 (UTC)

@Elominius Oh very nice! I just found it and linked it from the Wikidebate homepage, awesome! Sophivorus (discuss • contribs) 21:17, 2 October 2023 (UTC)

Change suggestions

• It makes no sense to let the user choose between Argument for and Argument against – in a Con subsection it should automatically be an Argument against.
• Here Template:Related changes is used but it only shows few changes. I think it would be better if it showed more. I tried to implement this by enabling specifying the number of days shown in the template but the way I did only works when previewing the changed template and clicking on the link, not after saving it.
• It would be best if it showed only recently edited debates instead of an item per edit simply showing a count of recent edits on the page.
• Currently, it's not possible to add supporting claims beneath a claim and to address these separately like it can be done on Kialo where one can add Pros and Cons beneath every claim. I think it would need to be made far more common for people to edit claims to for example add supporting data / explanations / sources / subpoints.
• Likewise, many claims miss explanations and/or supporting sources. I think they should more often be removed if they are incoherent / unexplained and that it may be good to enable creating objection types like a 'Citation needed objection' or 'Xyz fallacy'. Related to that, the claims are not sorted in any meaningful way like the strongest arguments being at the top and also don't have any kind or equivalent to impact ratings like it's possible in most Kialo debates.

I don't know if this can fill some niche that Kialo does not implement in a better way or whether it could be combined with other structured debate websites like that one somehow that makes this useful, e.g. by automatically importing arguments from there or serving as the base for a new Wikimedia project with a UI dedicated to this particular functionality. Prototyperspective (discuss • contribs) 20:27, 20 August 2024 (UTC)

@Prototyperspective Hi!

1. Fixed, thanks for the report.
2. I added parameters to Template:Related changes to control the max number of edits and days shown, and increased the defaults too. Feel free to increase them further if you see any value to it.
3. I improved Template:Related changes so that it groups edits by page.
4. That would require redesigning the entire system. It's a good idea, granted, but then the Wikidebate system would become identical to Kialo? What would be the differential or added value then? Perhaps the collaborative nature of wikis? But Kialo probably has some of that too (not sure though, don't really use Kialo). What then?
5. Feel free to remove or tag claims that you think deserve it. As to sorting, it's possible to do that via JavaScript, but it would happen after the page loads, leading to page shifts which are usually undesirable UX. We would also need to agree on what's the best criteria to sort arguments (more recent first? more objections first? more nested objections first?) The current system is: if you think an argument should be higher and another lower, then edit the page and do it. May sound too manual, and perhaps it is, but sometimes humans are smarter than algorithms.

Kind regards, Sophivorus (discuss • contribs) 14:08, 21 August 2024 (UTC)

Great, thank you! Didn't know grouping edits by page and showing the edit count was also possible with that template so that's nice.

Well one thing is that something else could be done so that people add explanations or sources to their and other claims. For example a template to mark unexplained claims, maybe even with a timer that auto-removes them if not at least some coherent explanation is added.

As for changes making it more similar to Kialo that is what I think could be a problem – it's somewhat redundant to Kialo, for example there are debates about subjects which are also discussed on Kialo but where the debate on Kialo is at substantially higher quality and with far more claims. I think where appropriate it is good to reduce redundancy or combine things – that's why I suggested thinking about things like importing claims from Kialo (they can be exported to csv) or for having this about things that Kialo is not good for. Differences are that here all of it is on one page, that it's open source, and that it doesn't have the problem of debates going inactive or mods not adding claims – here debates are added directly by users and can then only be removed while on Kialo valid claims are frequently not included for various reasons (like mods being inactive or potential bias). Maybe Wikidebates can simply serve as a way to create some dataset that could be used as a starting point by some modern javascript-based nonMediaWiki open source platform or be used for some particular part of argumentative debates like drawing conclusions from large structured debates. Maybe it could also be used in Wiki debates because maybe it could be embedded in talk pages etc directly and users can directly use it without having to sign up or in anywhere.

Well I guess one could use the default citation needed template. Maybe info on e.g. Template:Citation needed span should be added on this page so people know about and are facilitated to use it. I think the most robust most impactful arguments should be at the top like they are on Kialo if people gave arguments adequate the impact ratings. However, least objections or fewest unobjected objections isn't a good metric to calculate that so I don't think there's another way than letting people rate the validity and impact of claims. This could be done in the way that one could choose between argument for and argument against for every claim and I think when giving a claim a low impact rating a main objection would need to be selected that then gets a high impact rating so people see why people consider XYZ not a good argument. Arguments tagged as missing explanation / being incoherent could also be sorted to be placed further down. As an alternative to javascript doing the sorting dynamically there could be some bot updating the sorting every 2 weeks or so. This also enables people to see the main arguments and preferentially engage with these instead of minor arguments which is I think a key benefit of structured debates. All in all this is more an experiment because activity is even lower than on Kialo and both sites aren't used for things like policy-making/changing and better understanding thereof and get only few views. Prototyperspective (discuss • contribs) 15:07, 21 August 2024 (UTC)