What makes for a weak class system?

This post is to put a name on what I feel the purpose of a class is in a TTRPG and where I find certain class systems lacking more than others.

The purpose of a class is threefold, with each game placing greater emphasis on some than others:

1. As genre emulation. Classes give players an avenue to imagine their character's features, traits, tendencies, etc. as said class appeared in movies, literature, or other games. The player chooses to be a barbarian because they wish to imagine their character as Conan and the naming/flavor text/suggested style of play facilitates that imagination in a way that no other class does.
2. As ludic tools for expression. The class provides a set of tools unique to itself that the player chooses because they wish to express themselves through those tools. The player finds a warlock's spell list and class features intrinsically interesting and so chooses it to satisfy what they wish to accomplish.
3. As balancing act/numbers game. The class facilitates a role that is implicitly needed or uniquely valuable in the game. Compared to the previous item, this is an extrinsic motivation where the class' tools provide value to the group and resolve problems that will be faced in gameplay. For instance, a party has a thief in part because without one, they will lack the tools to manage certain problems.

I've chosen to list three examples here for the sake of simplicity, but one could argue that these are really two purposes, intrinsic and extrinsic, yet I think genre emulation and ludic expression differ enough. Genre emulation and ludic expression are both intrinsic to the class itself, but while genre emulation is diegetic (exists within the game's world), the language of ludic expression is non-diegetic, external to the games world.

Now, games mechanics often have bearing on the games world itself. A spell like Grease, for instance, has a mechanical element to it, but also has effects on the imagined world and theoretical utility within that world. However, because the spell (and other things such as "Hide in Shadows" or "Berserker Rage") are predicated on mechanical language, I think it's fair to say that ludic expression as it exists on most class sheets is non-diegetic. And even if it were not, most players distinguish the genre emulation side of their class from their abilities. There is often a dissonance between them (perhaps we'll call it ludonarrative dissonance? haha) as the kind of character the player imagines and wishes to emulate is not adequately facilitated by the non-diegetic ludic elements. A great source of frustration can arise when a character who imagines the rogue as a shadowy assassin finds themselves without the ludic tools to express that vision.

We might think of Type 1 classes as (for instance) Vampire: the Masquerade clans, whose unique style of vampirism is often based on horror literature or films. Type 2 classes--maybe Dungeon World or Whitehack. And OD&D is often considered to have type 3 classes, where it is valuable to have all for the sake of collective balance.

For that matter, pretty much every game with a class system ends up having a Type 3 element to it. Even in cases where the games don't encourage party balance or anything, people will inevitably consider the raw utility of a class. I see it most often in games which encourage certain behavioral loops (like games which emphasize combat).

I believe some class systems are lacking when the differences between the classes are not well defined or do not adequately follow through on the promises of a class's purpose. To use 5th Edition D&D as an example, the Barbarian and the Fighter both have unique abilities (ludic tools) that differ from one another and unique flavor text and suggested visuals which imply different genre emulations. However, the ludic tools provided to either class do not (in my experience) create enough meaningful distinction between the two. Whether you wrap a swordsman in Berserker Rage or Second Wind, their diegetic interactions with the world do not differ too much. They both wade into battle.

(D&D is often criticized for having a variety of classes that, through analysis of their ludic tools, are ultimately just different spins on 'violence-doer'. Not that violence is a problem, just that there's a singular purpose to 5e classes.)

This is not to say that flavor text and the imagined reality of the player characters is not gameplay or that it is irrelevant to discussion about games. A GM who recognizes their player's desires will tailor description and refereeing to validate and encourage the players' vision of the characters. To differ between the barbarian and the fighter, the GM would most likely use diegetic descriptions or backstory characters to reinforce genre-emulation differences. In my 5th edition game, our druid player and I worked very hard to rebuild the class's ludic tools in such a way that matched the player's vision of his character.

I'll conclude with this statement: A weak class system occurs when the ludic tools provided do not diversify diegetic play. They might not actually emulate the "type" of character they seek to emulate, or they might be redundant.

What do you think?