Making a Debate Game: the Design Challenges of Socrates Jones

When I first played Phoenix Wright: Ace Attorney back in 2009, I - like many others - instantly fell in love with its unique gameplay. And then, for some bizarre reason, I had this temporary fit of insanity: "You know, it would be kind of cool if they were talking about philosophy instead!”

And so, a year and a half later, I formed a team to flesh out the possibilities of combining Ace Attorney style gameplay with philosophical debate. With the help of Carnegie Mellon philosophy professor Andy Norman and some of my close friends from Game Creation Society, we nailed down the following goals for the game that would become Socrates Jones: Pro Philosopher:

With those goals in place, we began to make adjustments to the Ace Attorney template. Like the witness testimonies, we broke up arguments into individual statements. But instead of a vague “press for details” action, we gave the player three concrete ways of asking questions about each statement:

These moves were based on Andy Norman’s “How to Play the Reason-Giving Game,” which was meant to teach the fundamentals of critical thinking to children. “Presenting evidence” likewise became “challenging with contrary ideas.” Instead of grilling witnesses, the player debates historical philosophers; instead of looking for holes in their testimonies, he is seeking out logical flaws.

Let’s get one thing out of the way: Socrates Jones makes no effort to hide its inspiration. Philosophers yell “NONSENSE!” with possibly egregious frequency and the general flow of gameplay is very similar to that of the Ace Attorney series. But despite these obvious surface similarities, the things that were different about Socrates Jones: Pro Philosopher created quite a few unique design challenges. 

Where do you get your ideas?

Early in the brainstorming process, we decided that players would accumulate ideas which they could present as counterpoints to their opponents’ reasoning. However, a major concern was the question of where these ideas would come from. In Ace Attorney, the tools through which the player overcomes problems are neatly handed to them in the form of evidence.  Ideas, on the other hand, are… kind of abstract, and the palette applicable to even just one argument can be quite diverse.

One of the major sticking points for us was the fact that the players couldn’t introduce their own ideas into an argument, as people often do in a real debate. At one point, we were considering implementing an “idea cloud” that simulated the process of pulling thoughts out of thin air, where words floating around in Socrates’s mind could be grabbed for “inspiration.” But even on paper, it was apparent that was a poor substitute for the process of coming up with ideas in a real debate, not to mention the scope issues it would have created.

So what were we to do? We were on the verge of pulling our hair out. No matter how you sliced it, this seemed like an irreconcilable limitation of any computer-based “debate game.”

Well, we eventually realized we could work within this limitation, instead of fighting against it. An important goal of the game is teaching Critical Thinking, and a big part of that happens to be the Socratic Method – investigating the validity of arguments through questions, rather than objections. So we decided that the ideas necessary would all have to come from questions you ask of the person you are debating. Asking Kant about the Categorical Imperatives, for instance, will cause him to give you an example, which you can later throw in his face later.

Importantly, this attitude was reflected in the behavior of the player character, who rarely adds ideas to the archive himself. Because Socrates is more concerned with vetting the ideas of others than in presenting his own, the players’ inability to present their own random ideas is not discordant. It all came together to further emphasize understanding the ins and outs of what your opponent is saying instead of rushing to present your own views.

Ultimately, I believe if we had tried to implement something like an “idea cloud,” it would only have served to create the very frustration we were trying to avoid – giving the player the illusion of freedom without being able to guarantee that the idea they had in mind was actually present. By sticking to a method of debate that discouraged such behavior, we reinforced our learning goals and came up with a better solution all at once.

Abstract Progression

On the surface, the mechanics of Socrates Jones are very similar to its inspiration. But Ace Attorney’s mechanics can’t really be summed up with what is on the system side -- much of the game, and what is going on there, is inside the player’s head. In the shift from facts into the more abstract realm of philosophical debate, that mental gameplay changed dramatically. Most notably, the path forward in a philosophical argument was no longer always clear or singular.

In Ace Attorney, the key to progression is almost always in finding factual contradictions between the evidence and the testimony. In Socrates Jones, however, problems with philosophies can come in many forms – internal inconsistencies, problematic implications, and faulty assumptions, to name a few. Finding such problems is a much broader, less concrete objective – one which the player can approach in many equally reasonable ways, even with a limited number of ideas. But we couldn’t account for every single one of those routes – for practical reasons, we constrained the game to a single path.  So how do you maintain the illusion that the player is dictating progression in a linear game? Every time the player presented an “incorrect” idea that reasonably should have made an argument, the immersion would be broken. 

As it turns out, one of our simplest solutions wound up going a long way. Every time we watched playtesters present an idea, we asked them to note what argument they were trying to make. We quickly discovered that there was, in many cases, consistency across players in regards to the arguments they thought particular counterpoints would raise. So, in these instances, we added custom responses to the game.

For a comical example, in the tutorial section of the game, a salesman claims that “deer steal our jobs” - a point that can be undermined by merely Asking for Backing. However, several players presented the idea that “Deer live in the Woods” on this node, trying to make the argument that they couldn’t steal our jobs because of their location. This is… a valid move, if not the best one, so we wrote up a custom response for it:

This particular set of changes went a long way toward maintaining players’ immersion in the game – even if a custom response doesn’t let the player leave the argument, the player at least feels like the game is acknowledging his thought process, often giving him more information to work with.

Dramatic Tension

In the Ace Attorney series, the stakes are always clear – if you fail to defend your clients, they will be sent to prison for a crime they did not commit. On the gameplay side, you are only allowed to make a few mistakes before the Judge stops listening to you. As a result, each gameplay misstep brings your clients closer to a terrible fate, while each success brings them closer to freedom.

We struggled to provide a similar level of tension for the player in Socrates Jones, both in the narrative and in the gameplay. In early versions of the game, we had no system for penalizing “wrong” moves. While there were certainly moves that weren’t as well thought out as others, they were all treated the same way – so the player could theoretically just try EVERYTHING, and not attempt to process the argument at all.

Taking another cue from Ace Attorney, we added a credibility bar – if the player asks a question or presents a counterpoint with little reason or basis, they lose credibility. If the bar becomes empty, the game ends. Our hope was that this would discourage simply “trying everything,” forcing the player to think about their moves more carefully.

However, rather than discouraging brute force, that simple change made the game cripplingly stressful for those trying to play it properly. Suddenly, while they were asking for backing or clarification, the little green bar would decrease, and it wasn’t always clear to them why. When players ran up against our judgments about what warranted a credibility loss, they felt unfairly punished and were hesitant to ask further questions. As a result, the change ended up undermining an important part of our core.  Oops.

Ultimately, only having the player risk credibility when presenting a counterpoint proved to be the right balance between penalties and lenience. It gave the player freedom to explore an argument, while still requiring him to think before acting carelessly. So yes, the right answer was to do exactly what Phoenix Wright did. Go figure.

Oh, and on the narrative side, we went with a simple solution to the tension issue. Socrates is dead! Finding the nature of morality is part of a wager to win his life back. Bam! Problem solved. Cliché’d, but effective.

Real Philosophers, Set Ideas

Most of our characters are real philosophers. Most of the ideas you debate with them are their real philosophies. We did take some liberties – For example, Euthyphro was adapted so that his argument spanned across religions, rather than just dealing with the Greek Pantheon – but for every philosopher we needed to be careful to A) be true to the philosopher, B) make the philosopher understandable, and C) make the ultimate counterpoints logical. This, of course, presented its own unique challenges.

Take for example, Immanuel Kant, a philosopher whom players face off against towards the end of the game. Kant is notoriously difficult to understand in full – in part due to his intricate web of logic, and in part due to his tendency to invent terminology. When translating Kant for this interactive form, however, we see-sawed between being true to his original ideas and simplifying them for consumption– resulting in many panicked emails to Andy Norman clarifying Kant, cross-referencing with his in game arguments, and massive rewordings and restructurings.

In the end, I suppose we must have had some degree of success. In a recent playthrough, my friend Dana Shaw said that she didn’t understand why the characters kept saying Kant was confusing. “I don’t think Kant is as confusing as your characters think he is,” as she put it. While I am worried even now that this could indicate that we strayed from Kant’s true ideas, if we truly managed to make even the basics of Kant easily understandable Socrates Jones has earned its keep.

Obviously, this is a challenge in which having a philosophy professor on our project proved invaluable. We worked with Andy Norman endlessly – running the game by him, the arguments, the progression – in order to make sure we were accurately representing our subject matter, and that the problems ultimately raised with each philosopher were well reasoned. Ultimately, the best lesson here is that if you are dealing with a specific subject, don’t be afraid to reach out to those with greater knowledge in the field than you.

Using real philosophers had one unavoidable effect on the game: the “Question Relevance” move is not nearly as useful as it is in real life, as our philosophers have most of their thoughts nicely organized. This isn’t ultimately that much of a problem, but a few of our players did pick up on this fact. It might have been nice to have a few more invented arguments in order to balance things out.

On the flip side, working with historical philosophers also gave us one obvious benefit. Thinkers such as Kant and Hobbes were often quite eccentric, meaning that little modification was needed to make them fun to interact with – the line pictured above, for example, is a real, (very) slightly modified Hobbes quote. The guy was a grump.

In Conclusion 

I don’t think we’ve even come close to solving all the problems inherent in making a game about philosophical debate; we inherently limit ourselves by turning the often abstract flow of an argument into more concrete mechanics and moves - a logic puzzle with a single solution. But I do think we made significant progress on the issues, and learned several things that are applicable to anyone trying to make a game about something as free-flowing as a debate.

Ultimately, from the feedback we have received, I know our game is successful at conveying the principles which the game was built on to its players. And if the game helps with the proliferation of basic reasoning skills, it can only be considered a success. Down the line, I hope more games try to do similar things, and do them even better – but I am proud of what we have created here.

Socrates Jones was recently (finally) uploaded to Kongregate and can be found here. The two primary developers are on twitter -- Myself and Valeria Reznitskaya – in case you are interested in hearing our other random thoughts.