Putting More Class In Your Setting. Part I: Arguing About Class

I've observed and participated in debates about classes in fantasy-historical RPGs for about two years. Typically, a lot of people invest more time and energy in debating the game than playing the game. Disagreements commonly take the form of personal affronts, as if people who disagree with you are telling you how you should play. Or that they are dictating the terms on which you will join their game, which you will not accept. The fact that people are exploring variants and not asking you to play in their game is lost on many, many people. There is no earthly reason why every approach to the game has to be "for everybody" - to my mind, the more different playstyles and interpretations of the rules, the better.

The main stances in the debate - when it is a debate, and not a shouting match, boil down to these two positions. The first advertises itself as the more "modern" one in that it is willing to go beyond the letter of the rules, and thus, to promote player creativity. To gamers who hew to this line, classes (like fighter, wizard, rogue, and so on) are mere conveniences that have become ingrained in rule systems over the years. They have become refined over time by designers concerned that they are "balanced" and don't outshine one another, and that they should have plenty of built in options for players always looking out for new variants. Ultimately, classes are simply congeries of different mechanics that are grouped with one another for reasons of historical accident - once upon a time, thieves had various street-smart skills, and could effectively stab people in the back, so today's rogues are skill-monkeys who have sneak attack. Regardless of how little sense it makes to package these skills together, such players accept them with the proviso that they can portray the character who possesses such skills in any way they want, and that there is no imperative to play such a character as a member of a criminal underground. The notion that classes might be actual categories of people within the game setting strikes the proponents of this position as an anachronistic embarrassment - sure, there used to be names associated with every level in every class, and people would train to advance from a footpad to a cutpurse; but now, the game has evolved, and become much more complex. In the age of "player agency", we don't need built-in mechanisms in the rules to structure our relationship with the game world. My rogue can be a sleuth, my monk - a tavern brawler, and my warlock - a person who woke up with magical powers one day for no explicable reason whatsoever. The existence of monster statblocks, instead of NPC character sheets, for bandits and mages points to the fact that player characters are unique, as no class groupings into which PCs fit exist in the game world: class is a meta-game convention. So long as both you and I accept the mechanics - i.e. rules that stipulate how much damage I do, how fast I run, what spells I can cast - under particular circumstances - no one can tell me any different. And God forbid anyone suggests that a skill-based game may better meet this desire for flexibility - I'm playing this game - presumably, because there is a much larger pool of players available.

"Fluff isn't rules!"

The other position becomes, by default, the "traditionalist" position. Its proponents hold that what's inside the package called class should be, in varying degrees, reflected on the outside wrapper. The "story" part of the class - including membership in organizations (e.g. monasteries, thieves' guilds, temples, etc.), knowledge of class-specific languages (druidic, thieves' cant), explicit relationships with creatures granting powers (gods, warlock patrons, magical schools) - is no less a part of the "rules" than more explicitly "mechanical" aspects of the character. The notion that the two are separable, and the latter is freely jettisoned is a tendentious reading of the rules. Despite a more open design philosophy that allows playing against type, even the more recent 5e iteration, no less than the oldest ruleset (excepting the OD&D rules, where class remained undefined) stipulate that class is a profession - nay, more - a calling (PHB, p. 45). Adherents of this position often view ignoring the "fluffier" aspects of the class as a pretext to powergame - if you are not socially grounded in a class with a distinct structure, traditions, and training, then you can switch classes at the drop of a hat, to build a minimaxed machine - the real reason for the obsession with "mechanics", they aver. Even if a nominally classed character is built as a roleplaying challenge rather than an optimizer's dream, it can be aesthetically awkward. Since the game is a fiction, technically, everything can be reskinned - not only class, but race, weapons, skills, special abilities, etc. In extremis, the ruleset becomes a huge and clunky filter behind which the "real" character is concealed, and at that point, playing the character within that ruleset becomes rather awkward. An additional argument put forward by "traditionalists" is that characters in the setting - whether controlled by players or by the GM - should be able to get a read on someone's class by closely observing them - how often can a particular combatant strike in a given time period? What sorts of spells can a particular caster effect, and how many, per day? Since empirical observation of a character may provide verifiable data, that suggests that real in-game occurrences reflect real class differences, and not just metagame elements that are visible to players, but obscured from their characters behind a narrative veil.

Although my own take is closer to the latter position, I do not gravitate to it out of primarily "traditional" considerations. I also feel that much of the argumentation - on both sides, is quite weak. The idea that simple observation can generally yield accurate information on what class a character belongs to strikes me as quite faulty. Perhaps I can observe the kinds of spells someone casts - if I understand what they are doing (e.g. by making successful Arcana checks). I'm willing to allow that spells are cast pretty differently - they use different formulae and motions, the casting may have to integrate accurate astrological information and the position of the caster - all things that a non-expert would have a difficult time noticing. As for noticing that someone is "striking" several times in a six-second interval, that seems like a particularly weak argument. Hit points have always been an abstract measure, and being "hit" is not equivalent to sustaining a "wound".

But just because hit points or to-hit bonuses are abstract characteristics that should not have clearly defined counterparts in the game world, it does not follow that all in-game characteristics are of this abstract nature. Class seems to me quite distinct from characteristics that have simple measurements. Not only is it explicitly defined as a profession (or more), it is almost certainly the single-most important factor that determines what characters can do, and what makes characters different from one another. People can ignore classes as social structures if they like in their settings, just like some people (including myself) largely ignore alignment, but there is little support in the core texts themselves for treating class as just a bag of abstract mechanics. When the game's lead designers throw together a new variant, they are pretty explicit about the fact that they design the story elements first, and then throw up numbers to represent aspects of these elements. It is also not clear to me why only the "crunchy" elements should be regarded as "rules", whereas the "fluffy" elements are "optional".

Again, every group or GM can make their own decisions about this, but the uncritical "base-superstructure" reading, where numbers are primary, and mere words can be freely ignored, reflects a certain real-world philosophical positivism, and has nothing to do with how to properly approach the text of the rules. Why is the damage caused by a druid spell more of a rule than the druid's ability to speak the Druidic language (which a lot of "modernists" object to as an unnecessary sop to grognards, because their druid has nothing that binds them together with other druids, who may not even exist)? Typically, the "crunch first" positivism tends to run even deeper: ask most "modernists" if race, as opposed to class, is a "real thing" in the game, and they will usually say yes, because race is reflective of the belonging to an in-world biological group, and is therefore not just a "bag of mechanics" that can be interpreted in any way one wants. When you point out to them that they are transposing sociobiological "race realist" ideology from our world onto the game world, that nothing says that biological races in the game world are pure and have genetically hard-wired characteristics, and it's only one small hop away to accusations that you are a Social Justice Warrior, after which the game discussion inevitably breaks down. So, apparently, in D&D (!) race is real, and class is not, because Science.

But class is much more important than race, or any other game characteristic, in determining what characters can do. It is only in a class that one advances in level, and if level is better conceptualized as an abstract measure in many cases, a character's sense of advancement more broadly has a place in the game-world. The advancement is what draws a lot of players - its psychological rewards likely account for D&D's continued success in the face of competition from more "realistic" skill-based systems. While the psychological satisfaction from advancing in a class is felt by players, there is a case to be made that the purely qualitative advance in a class models in-game character satisfaction much better than simply increasing a score in a skill. Skills may be concrete know-how, but a class is an archetype, representing a heroic model developed by a hero, demigod, or important organization. Surely, the official use of archetype to refer to subclasses is not an accident - it represents the mythic quality of classes, meaning that the elements of a class are not some randomly thrown together abilities, but a particular archetypal pattern or style. And archetypes are culturally specific - the notion that every setting must contain exactly the same classes and subclasses - especially in this day and age, when variants are freely available and easy to make - is a bit silly. It may be objected that  non-official variants are not balanced - but that is really a misdirection. The official classes aren't perfectly balanced, either - you hear complaints about this constantly. Variants introduced in the Unearthed Arcana pdfs are in play-test mode, and have not been properly "balanced"; other designs deserve the same kinds of trial periods and considerations. And staking everything on balance does strike me as the kind of crunch-obsession that creative, role-play oriented types that want to redefine classes claim to be getting away from.

The notion that game settings should be open to all official (and semi-official) classes in order to accommodate player desires is a matter of taste, but there is no earthly reason why this has to be every GM's default position. Players do desire options, but there are few players who are committed to playing just this option in your game, no matter its flavor. If you explain your setting and present the new class options to new players, the vast majority of them will find something they like, and will often suggest several variants. There are people who insist that all Japanese restaurants should serve spaghetti and meatballs, but those who operate Japanese restaurants rarely build their menu around such expectations. It is similar with games - just make sure your setting is interesting and coherent, and that your class options work well together. And don't just take away options - provide new taste sensations.

Another argument against class as an in-game reality highlights the introduction of backgrounds as alternatives. If you want characters to fit into a setting - now you have a category explicitly designed for the purpose. The character can come from a noble, or artisan, or soldier, or hermit background. These place the character within a social category, and even provide guidelines for how they shape the characters' outlook on life, personality, and actions. But when it comes to classes, partisans of the no-class-in-the-gameworld approach insist that every character should be sui generis. But why should that necessarily be the case? As a fan of backgrounds precisely because they imbricate characters in society, I see no earthly reason why they obviate classes doing so as well. As characters grow in power, the background from which they originate fades into the background, while class becomes increasingly more important. Why does growth in power automatically make characters anti-social? Would not a real society put a significant amount of effort into regulating and institutionalizing power - regardless of the fact that these efforts would not prove 100% successful? Classes can be professions and callings - just of a different type than those represented by some backgrounds. They are professions and callings for those who wield a lot of power. People who wield a lot of power in our world may be special, but that does not stop them from belonging to (often special) social categories.

When it comes down to it, the underlying attitude of rejecting class as an in-game social structure stems from a real-world anti-societal ideology, rather than a disdain for old-school classes with their level designations as unnecessary crutches for a quality role-playing experience. The emphasis on mechanics (and the parallel belief in races as real in a biological sense) derive from a real-world worship of numbers and natural-scientific categories. But insistence that characters are bound by society is treated with suspicion: my character has completely untrammeled free choice in anything he or she does. In conjunction with the scientistic worship of crunch and race comes the Romantic cult of the genius: my character is a unique Renaissance Person, and possesses peculiar talents, not reducible to any training or structure. This is reflected in the common framing of "good role-playing": a good role-player is good at narrating his or her character's feelings, but not necessarily at figuring out ways to make a character part of the party, or the world. It remains a mystery, however, why people and monsters cluster around power-wielding people, ignoring the sea of mediocrity around them. It is similarly unclear how organized society survives in the face of their onslaught, or how monsters, who supposedly care not a whit for society, are sitting on piles of treasure that society produces and recognizes as valuable. Suspending disbelief to allow a very dense penumbra of powerful creatures around the PC party surrounded by a completely mundane world requires significantly more effort than just supposing that the world has classes to which a significant few belong. Moreover, the party is said to wield earth-shattering powers that developed completely spontaneously, and no one around them can have any sense whatsoever about what these powers are and where they come from - their attitudes toward the PCs can take the form only of overwhelming awe. It is curious that a people who are drawn toward D&D's high magic settings expend so much energy on spinning narratives that deliberately conceal magic's existence as a source of power.

The existence of monster statblocks is hardly a clinching argument in favor of classes being limited only to PCs. The statblocks are mere shorthand conveniences, because creating NPC character sheets is a more involved effort, and can be more difficult for a GM to manage during gameplay, since classes have become more complex, and have more whistles and bells than they did in many earlier editions, and since the GM is commonly managing multiple NPCs and adversaries during a typical encounter. But that doesn't mean they don't belong to classes. NPC statblocks are simplified variants of character sheets, and a glance at the spell powers of those monsters that are casters clearly shows that they possess spell slots, as would characters of a comparable level. On top of that, the DMG itself clearly cites "giv[ing] the NPC a class and levels" (p. 92) as one of the three main options for recording their statistics. The capacity of some people to misread the rules they cite as chapter and verse always astounds, (though I do allow that, as in poetry, religion and theory, there can be "strong" misreadings capable of generating creative approaches to the game).

In short, most of the typical arguments that modern (i.e. 5e) design militates against having classes as elements of game settings fall decisively short of the mark. Claims that even if the letter of the law does allow this (out of consideration for tradition), most people no longer play that way, has no substantial evidence behind it. One (admittedly unscientific) poll suggests that the community is pretty evenly split between "no", "yes", and "it depends" positions on "class in the game". A more scientific poll administered by lead designer Mike Mearls notes that 52% of GMs build their NPCs using the Player's Handbook - that is, with classes and levels. Having classes as in-game elements is therefore not atavistic or anti-modern, but neither does it have to be traditionalist. Every GM is within their rights to decide if the "traditional" classes are well-designed from a narrative point of view, and whether they fit into their settings. If some are found wanting, they can be jettisoned or replaced. Nor does the existence of classes as structures mean that everyone in the gameworld is clearly aware of them, or able to figure out the precise capabilities and levels of particular characters. In our world, the vast majority of the population doesn't really understand the difference between a molecular and a particle physicist, any more than a common peasant knows the difference between various spellcasting classes. But that doesn't mean that the differences don't exist in an institutional sense. Having classes also doesn't mean that they are all closed fraternities with their specific bylaws, symbols and chapter houses. Some classes are clearly more open along their perimeter than others. But again, that doesn't mean they aren't there.

* * *

In Part II of this essay, I will examine the setting-based reasons why classes are likely to be organized as concrete social groups.