We don't need to speculate about what Edwards said about
incoherence and
dysfunction. His writings on the topic are easily available online.
In the original GNS essays, Edwards says some things about incoherence
here:
In terms of design, the issue is incoherence, defined here as failure to permit any Premise (or any element of Exploration) to be consistently enjoyed. I think that any and all RPG designs have some identifiable relationship with the GNS modes, out of the following possibilities.
*Focused: the design facilitates a specific, identifiable Premise (or area of Exploration).
*Semi-adaptable: the design is at least compatible with more than one Premise and/or Exploration across GNS goals. (Whether this category even exists, or whether it merely reflects correctable incoherence, is debatable.)
*General: the design facilitates a specific mode, but permits a range of Premises or Explorations within that mode.
*Kitchen sink: the design utilizes layers and multiple options such that any specific point of play may be customized to accord with GNS goals. (This design often ends up being a general Simulationist one, however.)
*Incoherent 1: the design fails to permit one or any mode of play. In its most extreme form, the system may simply be broken - too easily exploited, or internally nonsensical, or lacking meaningful consequence, to pick three respective possibilities for Gamism, Simulationism, and Narrativism.
*Incoherent 2: more commonly, the design presents a mixed bag among the modes, such that one part of play is (or is mostly) facilitating one mode and other parts of play facilitate others.
In terms of actual play, yes, one "can" bring "any" GNS focus to "any" RPG - but I argue that in most cases the effort and informal redesign to do so is substantial, and also that the effort to keep focused on the new goals as play progresses is even more substantial. This chapter discusses why that effort needs to be there at all. . . .
Can multiple GNS goals be satisfied by a single game design? It may be possible, but it is not easy. As mentioned before, merely aligning topics of Exploration with those of Premise is probably not effective. I conceive of two types of
hybrid: (1) two modes are simultaneously satisfied in the same player at the same time, of which I am highly skeptical; and (2) two modes can exist side by side in the design, such that differently-oriented players may play together, which might be possible. Some possible candidates for the latter include these.
*G + S: Rifts.
*N + G: Champions 1st-3rd editions; I'm interested as well in seeing the upcoming Elfworld and a proposed game from Hogshead Publishing regarding fantasy weaponry.
*N + S: Little Fears and UnderWorld (these games' degree of "abashedness" exists squarely on the border of the two modes).
Drift is a related issue: the movement from one GNS focus to another during the course of play. I do not think that "drift" reflects hybridized design (in which both modes are indeed present), but rather correctable incoherence (moving toward coherence in one mode). Historically, drifting toward Gamism is very common; it isn't hard to understand that a frustrating and incoherent context can be turned into an arena for competition. Internet play has illustrated some distinctive drifting: Amber moves from abashed Narrativism either to Simulation with Exploration of Character or to Gamism with the emphasis on interpersonal control; Everway moves from abashed Narrativism to Simulationism with the emphasis on Exploration of Situation. . . .
Unfortunately, functional or nearly-functional hybrids are far less common than simply incoherent RPG designs.
The "lesser," although still common, dysfunctional trend is found among the imitators of the late-1970s release of AD&D, composed of vague and scattered Simulationism mixed with vague and scattered Gamism. Warhammer is the most successful of these. Small-press publishers pump out these games constantly, offering little new besides ever-more baroque mechanics and a highly-customized Setting (Hahlmabrea, Pelicar, Legendary Lives, Of Gods and Men, Fifth Cycle, Darkurthe: Legends, and more). Another, similar trend is the never-ending stream of GURPS imitators.
The "dominant" dysfunctional system is immediately recognizable, to the extent of being considered by many to be what role-playing is: a vaguely Gamist combat and reward system, Simulationist resolution in general (usually derived from GURPS, Cyberpunk, or Champions 4th edition), a Simulationist context for play (Situation in the form of published metaplot), deceptive Narrativist Color, and incoherent Simulationist/Narrativist Character creation rules. This combination has been represented by some of the major players in role-playing marketing, and has its representative for every period of role-playing since the early 1980s.
*AD&D2 pioneered the approach in the middle 1980s, particularly the addition of metaplot with the Dragonlance series.
*Champions, through its 3rd edition, exemplified a mix of Gamist and Narrativist "driftable" design, but with its 4th edition in the very late 1980s, the system lost all Metagame content and became the indigestible mix outlined above.
*Vampire, in the early 1990s, offered a mix of Simulationism and Gamism in combat resolution, but a mix of Narrativism and Simulationism out of combat, as well as bringing in Character Exploration.
The design is hugely imitated, ranging from Earthdawn, Kult, and In Nomine, to the mid-1990s "shotgun attack" of Deadlands, Legend of the Five Rings, and Seventh Sea.
All of these games are based on The Great Impossible Thing to Believe Before Breakfast: that the GM may be defined as the author of the ongoing story, and, simultaneously, the players may determine the actions of the characters as the story's protagonists. This is impossible. It's even absurd. However, game after game, introduction after introduction, and discussion after discussion, it is repeated.
Consider the players who were excited about the vampire concept for role-playing. What happens when they try to play Vampire: the Masquerade? Well, they try to Believe the Impossible Thing, and in application, the results are inevitable.
*The play drifts toward some application of Narrativism, which requires substantial effort and agreement among all the people involved, as well as editing out substantial portions of the game's texts and system.
*The play drifts toward an application of Simulationism in which the GM dominates the characters' significant actions, and the players contribute only to characterization. This is called illusionism, in which the players are unaware of or complicit with the extent to which they are manipulated.
*Illusionism is not necessarily dysfunctional, and if Character or Situation Exploration is the priority, then it can be a lot of fun. Unknown Armies, Feng Shui, and Call of Cthulhu all facilitate extremely functional illusionism. However, it is not and can never be "story creation" on the part of all participants, and if the game is incoherent, illusionism requires considerable effort to edit the system and texts into shape.
*Most likely, however, the players and GM carry out an ongoing power-struggle over the actions of the characters, with the integrity of "my guy" held as a club on the behalf of the former and the integrity of "the story" held as a club on behalf of the latter. . . .
The often-repeated distinction between "roll-playing" and "role-playing" is nothing more nor less than Exploration of System and Exploration of Character - either of which, when prioritized, is Simulationism. Thus our players, instead of taking the "drift" option (which would work), may well apply themselves more and more diligently to the metaplot and other non-Narrativist elements in the mistaken belief that they are emphasizing "story." The prognosis for the enjoyment of such play is not favorable.
One may ask, if this design is so horribly dysfunctional, why is it so popular? The answer requires an economic perspective on RPGs, in addition to the conceptual and functional one outlined in this essay, and is best left for discussion.
In the next chapter he adds
this:
When AD&D was released in its late 1970s form, its content encouraged a "more is better" approach. The more players, the better. The more time spent, the better. The longer the sessions, the better. The longer the sessions continued, the better. Nearly all role-playing games used AD&D as the starting point for presentation purposes, even those with vastly different systems and philosophies of play, and so this dysfunctional approach remains with us to this day. The term "campaign" is especially misleading, as in wargaming it denotes a specific set of events from point A in time to point B in time, whereas in role-playing it denotes playing indefinitely.
For those forms of role-playing that emphasize "story" in the general sense (see Chapter Two), this approach is completely unsuitable. What is a "story" to be, in terms of individual sessions and all-sessions? In role-playing culture, one is often assumed either to be playing a "campaign," which means it should go on forever, or a "one-shot" session which aside from the connotation of being superficial is simply too short for many sorts of stories. The functional intermediate of playing the number of sessions sufficient for the purpose of resolving a story is nowhere to be found in the texts of role-playing. . . .
I think the most common dysfunction . . . is GNS incompatibility. At the highest-order level, if the people simply have entirely different goals, then actual play continually runs into conflicts about priorities and procedures based on those different goals. I think everyone who's familiar with the theory knows that this is a "no fault, no blame" criterion. I like potatos, you like pink lemonade, have a nice game with your own group.
More difficult incompatibilities also exist within each of G, N, or S. People may share the the large-scale GNS goal, but be accustomed to or desire different standards for [various features of play]. In this case, dysfunction arises from (a) trying to resolve the differences during play itself, and (b) anyone being unwilling to compromise about the differences.
Drift is the usual method for dealing with this level of discord. It is a fine solution for resolving within-mode differences, if everyone is willing to give a little. However, drift has a dark side, or degeneration, the disruption or subversion of the social contract such that what is happening is not more fun, at least not at the group level. . . .
The tragedy is how widespread GNS-based degeneration really is.
In the subsequent three essays, Edwards says some things about "incoherent" game design
here:
Abashed" refers to design that must be Drifted in order to play because incompatible priorities are present among different parts of the rules. It's different from Incoherent design in that such Drift is easy and minor. Technically, an Abashed game is already at least two modes (or sub-modes) . . .
As far as I can tell, Simulationist game design runs into a lot of potential trouble when it includes secondary hybridization with the other modes of play. Gamist or Narrativist features as supportive elements introduce the thin end of the metagame-agenda wedge. The usual result is to defend against the "creeping Gamism" with rules-bloat, or to encourage negatively-extreme deception or authority in the GM in order to preserve an intended set of plot events, which is to say, railroading. In other words, a baseline Simulationist focus is easily subverted, leading to incoherence.
Whether this issue can be resolved by future designs and Social Contracts is unknown. Speaking historically, though, AD&D2, Vampire, and Legend of the Five Rings are especially good examples of incoherent design that ends up screwing the Simulationist. You have Gamist character creation, with Narrativist rhetoric (especially in Vampire). You have High Concept Simulationist resolution, which is to say, easily subverted by Gamism because universal consistency is de-emphasized. And finally, you have sternly-worded "story" play-context, which in practice becomes game-author-to-GM co-conspiracy. The net result is a fairly committed Simulationist GM presiding over a bunch of players tending toward more agenda-based play of different kinds. . . .
I think that we need to distinguish between Simulationist elements vs. coherent design - the former have certainly been widespread, but mainly in incoherent games, with AD&D and Vampire as the chief examples.
There is some elaboration on that last point
here:
The most striking feature across role-playing history is the astonishing shift in the late 1980s from assuming that Gamist play was the default to practically nothing - limited mainly to "old AD&D," various D&D imitators, Shadowrun, or Rifts.
I think this rarity is mainly a matter of rejection by texts that facilitated other preferred modes of play. I specifically include AD&D2 to be included in this shift, as I consider it to be mainly incoherent with various and sometimes-contradictory doses of Simulationist design scattered throughout, going all the way back to the Wilderness Survival Guide and the Dragonlance modules. I also think that the various setting-derivative AD&D2 boxed sets of the early 1990s (Al-Qadim, Dark Sun, Planescape, et al.) explicitly facilitate Illusionist Simulationist play.
The first of the later essays has no use of the word "dysfunction" that I can find. The second does, and so does
this one. The diagnoses of dysfunctional play about competing agendas among participants - eg gamists who wreck or break simulationist-oriented play, or would-be narrativists who struggle with the GM over control of the story in a game that the GM is intending to be high concept simulationist with the GM making the major plot decisions. I think this is pretty consistent with what is said in the original GNS pieces: dysfunction, when it occurs, is often (not always) a result of incompatible creative agendas.
The problem with incoherent game texts is not that they make such dysfunction inevitable - no claim of that nature is made that I can see - but that they make it more likely, by encouraging RPGers to refrain from asking
What is their creative agenda? and
How are the particular techniques and procedures set out in this rules text going to help achieve that agenda?
Edwards is obviously not unaware that a very common way to resolve these questions in the real world, playing a system like AD&D 2nd ed, is for the GM to take charge and impose their will. He does doubt whether that's the only resolution that's possible.