IN THE EYES of the DSM, we are a collection of diagnoses. In the eyes of Myers-Briggs, we are one of sixteen personality types. In the eyes of another human, or indeed ourselves, we are a countless number of things that, when taken together, approximate our essence. To be seen is to be seen as something. Computers are incapable of seeing us as people do, but they are very good at seeing us as numbers.
The story of the DSM is the story of fitting human pathology into a quantitative taxonomy. It is a messy business, because the labels never match up exactly, and it makes for a gross simplification. But what if we were to begin from the other side? What if, instead of trying to regiment human complexity into similar forms, we try to build out human complexity from computer-friendly representations?
There was one other famous system for the quantification of human personality and behavior that I became familiar with as a child, which was Dungeons & Dragons (or Advanced Dungeons & Dragons, as it was known in its standard incarnation back then), created by Gary Gygax out of the residue of dozens of less successful antecedents. The mainstreaming of geek culture has turned both computers and role-playing (whether on computer, tabletop, or “live-action” LARPing) into major social phenomena, such that it’s very difficult to tease apart the development of each. Systems like D&D appealed to the analytic mind-set of many engineers. What they provide, after all, are the computational algorithms for an entire fantasy world.*1
D&D and other role-playing games (RPGs) share a common core. Half a dozen or so people each adopt an ongoing alter ego whom they play over multiple game sessions. The Dungeon Master (DM), who referees and drives the game, provides some kind of scenario for these characters to explore jointly. The characters negotiate the traps and monsters laid out by the DM with their individual skills, and the characters develop and become more powerful through their adventures. Characters can also die. A great deal of responsibility is placed on the DM, who has to challenge the players without frustrating them and create suspense without being cruel.*2 Unlike sports games, where the referee is only there to apply rules, DMs have a great deal of latitude to bend or even break rules at their discretion—such as, for example, saving a character when death would disrupt play too much—or alienate a participant.
An explicit set of numerical mechanics, manipulated through simple arithmetic, provides the glue between D&D’s structuring elements. The backbone of D&D’s mechanics is its class and stat system. A player picks a class (fighter, thief, wizard, etc.) to determine a character’s basic skill set (punching, casting spells, pickpocketing), then rolls dice to determine how good the character is at each skill. D&D contains a core set of six character statistics. There have been hundreds of variations on the basic D&D template, but most don’t stray too far from these original six, which divide into three physical and three mental (though since magic exists, mental powers can be pretty physical too):
Strength: How hard do you bash things?
Dexterity: How fast do you bash or avoid being bashed?
Constitution: How much bashing can you take?
Intelligence: How book-smart are you (i.e., how hard do your spells bash, and how quickly do you learn new ways to bash things)?
Wisdom: How street-smart are you (i.e. how’s your bashing-avoidance know-how)?
Charisma: How good are you at convincing people not to bash you?
The algorithms of D&D are simple: a “stat check” consists of rolling a die and not rolling over the character’s current skill level. A character’s ability to learn woodworking from a woodworking book would require a 1d20 roll against intelligence: if the roll of a 20-sided die results in a number lower than the character’s intelligence value, the character learned woodworking. But if my intelligence was 12 and I rolled a 16, my character wasn’t intelligent enough to learn woodworking. Need to cross a rickety bridge? Dexterity check. Decipher ancient runes? Intelligence check. Play the alpenhorn? Constitution check.
* * *
—
Everything in D&D is quantified. Armor has a number indicating how tough it is, a weapon has a number for how much damage it does, and every person or monster in the game has its own set of numerical stats. There are numbers for how good characters are with particular weapons, particular tools, particular skills, particular jobs. Everything functions algorithmically by comparing dice rolls against one stat or another, though players have only partial knowledge compared to the Dungeon Master, who tracks the whole story and keeps certain dice rolls secret. I might know that my Dual-Tipped +3 Enchanted Quarterstaff of Pusillanimity did 2d8+3 damage against the Pulchritudinous Troll, and I could calculate that after rolling a 5 and a 2 with an 8-sided die that I’d done a raw 10 points of damage (5+2+3), but the Dungeon Master might tell me only that the Pulchritudinous Troll was looking a bit more haggard, leaving me uncertain about how much more damage I needed to do while the troll bashed me for 1d20+5 (6 to 25) points of damage every turn. After killing the troll, my character gains experience points and perhaps a level of experience, at which point my character could also earn new abilities and skills. The gradual leveling of a character provides an ongoing sense of progress and mastery that permits greater challenges to be met (tougher monsters, mostly).
For me these challenges weren’t so thrilling. In computerized versions, where progress and acquisition were greatly accelerated compared to the human tabletop game, I could appreciate the novelty of finding an Infinity+1 Sword or acquiring the powerful yet impractical Armageddon spell, but much of the actual gameplay was tedious. Classic series from the eighties like The Bard’s Tale, Ultima, Wizardry, and Might and Magic, which were all based on D&D-like models, consisted of endless battles in twisty dungeon mazes to “level up” your characters so they could survive being bashed by ever more powerful monsters. This process is known as “grinding,” and I never cared for it. The tedium outweighed the strategy. The details I remembered most were the incidental ones, like the bizarre “floors” from Ultima III, enemies you couldn’t see because they looked just like the floor. Invisible moving monster floors: that’s a great touch.
Battling the floors in Ultima III: Exodus
The diabolically difficult mazes of Wizardry IV were more interesting for me to contemplate than they were to play. (This was a game that was so difficult most people were unable to figure out how to get out of the first room.) I marveled at people who enjoyed agonizing trial and error, frustration, and endless character death as player retention strategies. I didn’t enjoy the gameplay, but I was fascinated by the games’ simplified representations of reality.
D&D’s stats are the data structures of D&D. In the 1970s, when D&D was strictly played on paper, the game was already well suited to computers. D&D’s metrics of character and world development, represented numerically (for stats) and geometrically (in dungeon layout and combat), provided a combination of simplicity and precision that made the adaptation of tabletop role-playing to computer gaming inevitable. No other fictional genre, not even science fiction, was as tailor-made for algorithmic computation as the fantasy world that D&D elaborated. D&D relaxed fiction’s burden of storytelling and character psychology in favor of coarse but exact mechanics, simplifying the world into a set of representations that a computer could manage far better than, for example, the mystifying world of mental illness. Yet both D&D and DSM embrace approximate quantified fictions. They are both provisional, extensible, and revisable folk taxonomies.*3
D&D’s extensive auxiliary components quantify its world, borrowing lore from mythologies and religions of every stripe, resulting in a perverse syncretic assemblage.*4 Gygax was far more a synthesist than an original thinker, and his genius lay in his doggedness at creating kitchen-sink worlds capable of engaging role-players of every stripe, worlds simultaneously open-ended and simple enough to allow for variation.*5 There’s the simple yet
uniquely peculiar two-axis alignment system, in which each character is a combination of Good/Neutral/Evil and Lawful/Neutral/Chaotic. Similarly, Gygax had a passion for using the basic algorithms of D&D to regulate every aspect of the created gameworld. Though boiling down every player interaction to numerical dice rolls was hardly necessary for play, the ubiquitous dice rolls provided a centralized gameplay mechanic without restricting how that mechanic could be used. Gygax, for his part, was indefatigable and enthusiastic in constructing these suggestive, loose-fitting analogues to the world.
April 1974: The Minneapa fanzine’s parody of D&D’s system of approximate quantifications
The drollest example I know of Gygax’s omnivorous yet haphazard rage for classification and quantification is the “Harlot encounter” table from the first edition, written by Gygax himself. It is a table that indexes dice rolls from 1 to 100 (using two ten-sided dice), and it demonstrates conventional wisdom and cultural norms being put through the wringer of quantification.
D&D’s Harlot encounter table
The ordering of the harlot taxonomy goes by what could loosely be termed the “classiness” of the harlot. Number divisions usually have an implicit or explicit rank to them. Rolls from 01 to 90 result in harlots of increasing price. Rolls from 91 to 00 produce pimps of four varieties, though for some reason a “sly pimp” is twice as likely to occur as any of the other three types. With the exception of the haughty courtesan, who has only a 5 percent chance to show up, other types of harlots occur with either 10 or 15 percent probability. From this we see that there are nearly as many sly pimps in the world of D&D as there are haughty courtesans. I’m not sure how many Dungeon Masters would be able to assess and depict a meaningful difference between a saucy tart and a wanton wench, which is likely one of the reasons why this table disappeared from subsequent editions of the game. There is, indeed, little reason for this table to exist beyond a sheer mania for classification itself.*6 But what makes a “haughty courtesan” higher-class than an “expensive doxy”? How does a DM distinguish between a “slovenly troll” and a “brazen strumpet”? As with the DSM, the classifications are in large part arbitrary.
But as the harlot table suggests, the quantifications are also soaked in conventional wisdom. Just as homosexuality entered and then exited the folk taxonomy of the DSM’s pathologies, both fashionable and unfashionable ideas fell into D&D as “objective” descriptions of reality. Nowhere is this more apparent than in the treatment of gender. Historically, wargaming and tabletop gaming were overwhelmingly the province of men, a product of the male hobbyist origins of the games. Women entered the workforce and academia—even in computer science—decades before making similar levels of incursions into male recreational activities. As the hobbyist and academic worlds of computers merged after the dawn of the home computer in the 1980s, computer science became more gender-imbalanced, with the number of undergraduate degrees awarded to women nearly dropping in half between 1986 and 1990. The recreationalization of the computer was likely a significant cause. The world of Dungeons & Dragons was very much the product of an overwhelmingly male mind-set that catered overwhelmingly to men. Yet D&D’s flexibility allowed it to evolve to accommodate female players, mostly by eliding the distinctions between male and female characters. Longtime D&D writer/fan Lenard Lakofka’s 1974 system for developing female characters in D&D suggested that female characters be limited in strength to 14, while males were at 18, along with several other recommendations that infuriated many female and male players. Lakofka, alongside Gygax and Dragon magazine’s publisher, were hanged in effigy in a cartoon in the following issue. When a system is such an inexact approximation of reality, any such distinction becomes laden with value judgment and prejudice. As a result, sex distinctions were collapsed within D&D, and only fantasy ones (between elves and humans, for example) were maintained. Ironically, as sex differences dissolved, cultural and physiological differences between races (half-orc, elf, dwarf) grew increasingly broad and extreme.
By far the most fascinating D&D creation I know is the bizarre fever-dream Planescape: Torment, originally released in 1999, which tosses out most of the stat-building and a good chunk of the combat in favor of a twisted and often deranged morality tale of amnesia, immortality, remorse, and psychosis. Paradise Lost is thrown in as an incidental secondary plot. I tore through Torment with a combination of respect and disbelief, amazed at the uncontrolled and unrefined passion that made it difficult to anticipate what craziness it would throw at me next. (There are over a million words of text in the game, and they range very far afield.) There were dream sequences, fake religious cults, magical tattoos, sentient rat kings, doppelgangers, robot mazes. The stats take a backseat, as by the end of the game the main character has “leveled up” so drastically he has godlike powers, a total betrayal of the slow-grind progression that typifies role-playing games like D&D. Torment has a well-deserved cult following. Yet it (as well as similarly bizarre creations like Vampire: The Masquerade–Bloodlines and Mask of the Betrayer) would not have come into existence except as a response to the far duller rudiments of most RPGs (The Bard’s Tale, Diablo, Warcraft, Skyrim). We need our arbitrary frameworks, even the mediocre ones, for creativity to soar.
Planescape: Torment gains its resonance from choosing the right threads to follow in Gygax’s loose tapestry, opting for surrealism and subversion at every juncture and ignoring the less interesting bits. The unsettling psychological derangements within the game owe little to the peculiar and antiquated “Types of Insanity” table from Gygax’s original rules, listing the conditions that could be triggered in a character due to “mental attack, curse, or whatever.”
TYPES OF INSANITY
dipsomania
kleptomania
schizoid
pathological liar
monomania
dementia praecox
melancholia
megalomania
delusional insanity
schizophrenia
mania
lunacy
paranoia
manic-depressive
hallucinatory insanity
sado-masochism
homicidal mania
hebephrenia
suicidal mania
catatonia
Gygax’s table does not derive from a single source. Coming up to a round twenty diagnoses, Gygax must have either combined or trimmed sources to get to that convenient number. In some respects, it is similar to the then-current DSM-II, but it also includes illnesses like dementia praecox and melancholia that had been eliminated in the DSM but look back to psychiatrist Emil Kraepelin’s original taxonomy of psychoses formulated at the end of the nineteenth century. Gygax doesn’t acknowledge, realize, and/or care that dementia praecox and schizophrenia were generally considered identical, or that catatonia and hebephrenia were subcategories of schizophrenia (and dementia praecox). Not that Gygax’s definitions bear much relation to the clinical ones: Gygax defines dementia praecox as ennui, while schizophrenia is based on the then-common conflation with split personalities:
Schizophrenia: This form of insanity has the well-known “split personality” trait. From 1 to 4 separate and distinct personalities can exist in the afflicted—base the number upon the severity of the insanity. Likewise, the difference from one personality to the next should reflect the severity of the affliction. Each “new” personality will be different in alignment, goals, and preferences. (A very severe case might have a different class also but without coincidental possession, the new personality emerging will not have the actual abilities he or she may think that he or she possesses.) The onset of schizophrenia is random, 1 in 6 per day, with a like chance of a new
(or return to the old) personality emerging. However, whenever a stress situation—decision, attack, etc.—arises, the 1 in 6 chance of schizophrenia striking must be checked every round in which the stress continues.
“Naturally,” Gygax writes, “these forms of insanity are not clinically correct.” It seems unlikely that Gygax chose the names in the insanity table for any reason beyond their mythic and exotic sound. In the game, mental disorders are regimented to be more predictable than in real life: manic “lunacy” is literally triggered by the moon’s phase, while sadomasochism causes the player to alternate between sadistic and masochistic periods every three days. The drearier names of the DSM-II are absent from the list: no one in the world of D&D suffers from obsessive-compulsive disorder, hypochondria, or even anxiety.
Longtime D&Der Bob Kruger describes the appeal and practice of the game as follows:
Essential D&D resonates with our evolved cognitive biases, and explores archetypal symbols. It takes practice to do it well, but doing it well is less a function of adhering to the rules than mastering the art of appearing to faithfully apply them.
Kruger strikingly echoes the rationales for provisional psychological taxonomies like the DSM. A folk taxonomy based in human cognitive predispositions becomes an explicitly delineated rule system. Yet the system is full of arbitrary kludges, and the rules are there not for strict obedience but for loose guidance—just as most physicians treat the DSM. And in both cases, transferring the ruleset to computers removes the ability to slacken the rules ad hoc. Computers reify the taxonomy in the most literal sense: computers make the taxonomy more real. The taxonomy becomes more than a set of labels. It becomes a systematic organization of existing things: an ontology.
Bitwise Page 15