...

Here’s an argumentative structure that I have both made and encountered on the internet. It seems to be very straightforward, but I’d like to complicate it. In a rough form:

“Well, Silksong is a Metroidvania. But its platforming is bad. But platforming is a necessary feature of a Metroidvania. Therefore, Silksong fails as a Metroidvania, and this is a count against it as a game.”

Let’s call this the “essence argument,” for reasons that will become clear shortly.

Just to be clear, I do not endorse the above argument (nor do I know if anyone has made it). But I have seen this form of argument made in the game review space, and even among more sophisticated game reviewers (that is to say, not IGN or what have you). Here is a formal version of the argument:

(1) Some game X is of type Y.
(2) X fails at some feature(s) Z.
(3) If some video game A is of type Y, then A has some feature(s) B.
(4) Feature(s) Z are a subset of feature(s) B
(5) If some video game A fails at some feature(s) Z and feature(s) Z are a subset of B, then A has a failure as a video game.
(6) X has a failure as a video game.

This ended up being more complicated than it first appeared. It is trying to state that because a game is of a certain type, it has certain features (say gameplay features like platforming challenges). But then the game fails at these features, so it fails as a game. The argument presented is valid. But there are, of course, problems with the argument as I’ve given it.

First, how are we to interpret (2)? Here is one way to interpret that X fails at some feature: X doesn’t have that feature at all. An FPS doesn’t have platforming challenges and therefore fails at platforming challenges. If a game satisfied (2) in this way, then it seems that we could also reject (1). By modus tollens on (3), we would have a contradiction. If we assume that Call of Duty is a platformer, and that it fails in this way, then we should reject that initial assumption it is a platformer at pains of contradiction. This raises the question: How are we to formulate this argument such that one can’t reject the initial assumption that the game is of a certain type? If I say that Silksong is a Metroidvania and that it fails at platforming challenges, why can’t a dissenter just reject the assumption that Silksong is a Metroidvania? For just because the developers intended for Silksong to be a Metroidvania doesn’t entail that it is (e.g., they could have accidentally made an FPS).

The fix here is in premise (3). A video game being of a certain type entails that it has some essential properties, not merely that it has those properties. To elaborate: a human person with a disability might be said not to have the property of being a biped. But this seems to be glossing over important details. They should be a biped; they just aren’t realizing their biped-ness. This is, of course, not their fault. The “failing” language just has this unfortunate connotation in colloquial use. The idea is that there is something about being a certain type that implies features like being a human implies certain features. But this sense is not merely a logical one. So:

(3) If some video game A is of type Y, then A has some essential feature(s) B.

This is a good fix. Now by showing that a game doesn’t have platforming challenges doesn’t imply that it isn’t a Metroidvania. It can be an incomplete or deformed Metroidvania. The exact way to defend that something is a Metroidvania yet it fails at some essential feature might be tricky. It is easy in the human case. Something born to humans is a human, so if it fails to have an essential feature this is understood as a defect and not as demonstrating it isn’t a human. For games, we might show that it has platforming, but that it doesn’t execute it well.

One might still try to pull the same trick (arguing that, e.g., Silksong isn’t a Metroidvania), but this trick isn’t always logically available. An argument like this would need to proceed by giving reasons why we should reject (1) independent of the justification of (2). It was a problem before because the argument as originally stated could always be interpreted as allowing the trick. But that leak is now patched. So even if the feature is not present (say, platforming in an incomplete platformer), the argument doesn’t permit the objector to imply that (1) is false. In effect, the inclusion of “essence” langauge allows the existence of something like incomplete platformers.

So, the current version of the argument is:

(1) Video game X is of type Y
(2) X fails at some feature(s) Z
(3) If some video game A is of type Y, then A has some essential feature(s) B.
(4) If some video game A fails at some feature(s) Z and feature(s) Z are a subset of B, then A has a failure as a video game.
(5) X has a failure as a video game.

However, premise (4) is false. We are not warranted in saying that because X fails qua Metroidvania that it fails qua video game. For analogy, something failing as a monkey does not imply failing as a mammal. That something might be a human or a platypus and therefore do very well as a mammal. So, we need to modify the argument a tad and add an additional premise:

(1) Video game X is of type Y
(2) X fails at some feature(s) Z
(3) If some video game A is of type Y, then A has some essential feature(s) B.
(4) If some video game A of type Y fails at some feature(s) Z and feature(s) Z are a subset of B, then A fails qua type Y
(5) If A fails qua type Y, then A fails as a video game.
(6) X has a failure as a video game.

But for the same reason mentioned earlier, (5) is false. A platypus is a type of mammal, yet something failing as a platypus does not mean that it fails as a mammal. I fail as a platypus, but not a mammal. But we have located the problem. How do we go from a game failing as a subgenre to failing as a video game as a whole? For example, some hypothetical game mixing up two subgenres might not fit into a subgenre, nor may it succeed as either subgenre. But this obviously doesn’t mean that our hypothetical game must be a failure of a video game. Furthermore, (as previously mentioned) if a game was intended to be type Y, this doesn’t mean that it succeeds at being type Y. Of course, if something fails as a video game, it fails as any subgenre of video games—but this is irrelevant.

Here’s a potential fix:

(5) If A fails qua type Y is and is essentially type Y, then it fails as a video game.

However, my failing as a biped doesn’t imply that I fail as a mammal. Some mammals aren’t bipeds.

At this point we may think to stop here. Let’s just say that A fails qua type Y. What’s the problem with that? I feel that just saying (4) doesn’t carry the normative force that we want it to. Even if Silksong is a Metroidvania and it fails as a Metroidvania, should we consider this to be a count against Silksong? Surely it is a count against Silksong as a Metroidvania, but is it a count against Silksong?

What do I mean by this? Surely, someone with a disability in some sense fails qua human. But this needn’t be a count against that thing understood as a person. In fact, modern sensibilities find this strikingly immoral to even suggest. What could be going on is that that thing falls under two types: human and person. I think this suggestion satisfactorily matches our intuitions in the disability case, and it isn’t merely a philosophical trick to make this move. But this raises a question: can you make a trick like this for video games?

What I mean: Silksong fails qua Metroidvania, but this needn’t count against that game understood as falling under some other category. This move can be pulled as long as there is some category Silksong falls under. Are there constraints about what this category might be? In our real-world scenario, we might think that there are real types that an entity falls under. It is a fact that I fall under the types human and person. But the relevant categories for Silksong are not objective: they are conventional. Even if I say that Silksong objectively fails as a video game, it might succeed as an artwork (for example, that Getting Over It game could be a good example of this). In addition, even constraining this category to be something falling under video game doesn’t do much to limit the possibilities. That category itself is also conventional, and it seems that we might expand it.

I’m going to stop here for tonight, but I’m struggling to see how the essence argument can establish what it wants to. To categorize a video game as type Y and to be negatively evaluated by the standards of type Y does doesn’t entail anything about its overall evaluation.