Thursday, 28 February 2013

On Theory (Re)Defined, Railroad (Part III)

Other Objections & Responses

Now, obviously, some people agree with –C.

Jason Packer, for instance, writes, 
Restriction on action by the player based on the codified rules would never, in my opinion, constitute railroading.   Railroading is much more the province of the story itself, where players are given "freedom to choose" but only one choice will actually advance the plot, and all other choices are rebuffed.
This, of course, assumes that the game and the story are separate things, as with a role-playing game with many scenarios, and that the decisions that comprise the scenario are something different than “the game”.  We discussed this last post, when we noted that statements (2), (3), (8), and (21) from the original proposition cannot all be true.

As came up in the comments from the last post, it is easy to imagine a game in which everything is narrated by the rules, except your choice at a single point, which determines how you reach the next section, but does not otherwise affect the following sections, which then reach a pre-scripted conclusion.

If it is true that you are playing a game so long as you have any agency, then this is a game. If it is true that the rules cannot railroad, then this is not a railroad.  But if you make these claims, you will be attempting to redefine the term in a way that precludes the common meaning.  And this is not an irrelevant point, because –C proposes that being able to choose fighting stances in a game is sufficient to prevent it from being a railroad.  There is not a great deal of difference, to my mind, between the game just imagined, and the game in which all you get to do are pick what stance you follow the predetermined course in.

Random Wizard says 
I am in agreement with your general underlying idea (if I interpreted correctly). I am of the opinion that railroading and sandbox both refer to the same thing. They are markers on a spectrum that conveys to the players to the DM about "Where does the game that we are playing start and end?"
On the other hand, Telecanter said,
If you think about the original metaphor-- a train that starts at one place and goes to another-- expectations have little to do with it. If we both play a video game and the game makes us meet the same people, watch the same cutscenes, fight the same bosses, we are essentially trains on the same rails.
Your expectations come into play in whether you can still enjoy sitting on a train or would rather take a car and stop and different places and at different times. But whether you enjoy an adventure path or not or are surprised by it doesn't make it any less a railroad if everyone that plays it is expected to to the same things, in the same order, and end up the same place.
I think railroad is an important descriptive term and shouldn't be discarded because people have become familiar with railroads or enjoy them.
We may find –C’s response to Telecanter instructive.
I think this is post-rationalization.
In case it is not clear, “post-rationalization of an action is the act of giving meaning and purpose to actions after they have been conducted”, (, and it is unclear why –C would think this to be so.  Again, we can look toward others to see how the term is generally used:,,  When we do so, we see that Telecanter’s usage follows common usage.

(As an aside, I wonder how this gibes with "The issue comes because the way you start parsing out the motives of the person is that you ask them. When you construct a motive out of thin air it is at best rude."?  Is -C claiming that he is, at best, rude here?  Or is he claiming that there is a substantive difference between his questioning of Telecanter's motives, and my questioning of his?)
The term originated because it was about misread expectations. Players believed they were driving a car. Illusionism told them, they were actually 'on rails'. They tried to drive the car off the tracks and found out that their vehicle was actually a train.
 The term arose from conflicts of expectations of agency.
It is actually pretty clear where the metaphor came from, and one can peruse various meanings of the term in various contexts (not simply limited to rpg contexts), in order to gain a better understanding of how the term arose. might help:

"To railroad" means to rig a situation such that events can only play out in a particular manner, or to a particular end. When applied to a trial, it means to manipulate the judicial system such that a defendant is virtually guaranteed a conviction. The metaphor derives from the nature of a railroad track, which does not offer a train the ability to choose its path of travel.
"The lead prosecutor railroaded Eddy into a first-degree murder conviction: she relied heavily on the testimony of unqualified experts."
 "If you are hosting a murder-mystery party, avoid railroading your guests into quickly solving the case. Give them freedom to pursue tangents in the storyline they find interesting."
We will note that, within the general context, Telecanter is far closer to accuracy than –C.

We will also note, I hope, that with all of this material, we are still talking about a specific problem.  We are not talking about simply loss of agency, but loss of agency to influence how a game unfolds.  We are talking about all forms of agency – a game can be a railroad and still allow for other types of agency, and a game can not be a railroad and still restrict other types of agency.   (More on this below.)
The reason why the expectation is important to the metaphor, is that according to the example you've given, every game ever is a railroad, because by following the rules of the game, you are stuck on the train
Again, this conclusion is only valid if all forms of agency, and all levels of inclusion or exclusion, are co-equal in relation to role-playing games.  The value of a term like “railroading” is derived from its relationship to a particular subset of player agency; removing this makes the term less specific and removes its communicative force.
You can't say "Final fantasy is a railroad" because you have agency in a (most, at any rate) final fantasy game (the ability as a player to make choices that affect outcomes). Every game has agency, otherwise it wouldn't be a game.
Railroading is only useful as a definition if it quantifies a specific design issue. And it does, the remove of agency from players within the context of the game.
I could go on, and I do urge you to read through the comments section of the original article. 

I will point out, though, that –C’s statements should lead one to believe that he is aware that he is not using the term as it is commonly used.

One more point about whether or not rules can railroad.  –C writes:
An obstacle (like a door or a trap or a monster) is not agency restriction because it does not affect the ability of the player to take action.
A giant rock may be blocking a door. This rock is indeed blocking the character's path.
It does not block the ability of the player to take action and have a result that matches the intent of their action. They may get dynamite and blow up the rock. They can break it apart with a pickaxe. They can cast passwall or stone shape to bypass the rock. They may ignore it and go find a secret entrance.
A railroad is when the agency of the player is actively removed. They use the dynamite and are told the rock is 'too hard to blow up' or 'there is no dynamite for sale anywhere'. Or their pickaxe keeps breaking. Or their spells don't work in the cavern because of an anti-magic field.
The difference is, an obstacle is affected by player actions, and a railroad nullifies player actions.
 Are we actually to believe that a rock being too hard to blow up, dynamite not appearing on equipment lists, or pickaxes breaking when used in caves cannot be a rules artefact?  Are we seriously to believe that spells not working in the cavern because of an anti-magic field cannot be part of a game, such as a computer game, where the game is, in and of itself, a single scenario?

This is, of course, also true in a game like the original Advanced Dungeons & Dragons, where there was intent that a portion of the rules be hidden from players, to increase their specific agency to make decisions as characters in the fantasy milieu.

I hope that the problem here is as obvious to you as to me.

In fact, it is even obvious to –C.  When Rev. Dane Black says
This dates me, but the classic example of Railroading within a RPG is the original Dragon Quest/Dragon Warrior.
After slaying a dragon within a forgotten cave you happen upon the missing Princess. She asks if you will escort her back to her home. If you refuse, she responds, "But thou must!" and asks again. They only way forward is through the agreeing. It also happens at the end of the game, she asks if you will take you with her in your travels following the resolution of the major plot. Again, insisting that "You must!" if you refuse. 
The Dragon Quest games are filled with these seeming decisions that only allow for one choice that advances the plot.
He replies
Right.Why give you a choice? 
The game acts like it gives you a choice, and then removes the result of your choice. It was a decision made by a programmer, where you were given the expectation of agency and then that agency was removed.
So, it is clear that even –C agrees that a possibility exists where the game itself can remove agency, at least where there is no clear distinction between the game and the scenario.  If we accept propositions (3) and (8), though, the distinction becomes extremely blurry.

Especially if you consider statement (19) If you were being railroaded, you wouldn't be playing a game, because by definition your agency is being invalidated.  That's a pretty black and white statement that doesn't require distinctions between who or what is removing agency.  It states that you simply cannot be railroaded in a game.  IOW, the term "railroad" might as well have no meaning, because a railroad can never happen.

Consider the Sources

-C notes that he includes his sources as evidence [onus probandi, (Making claims without proof, and then claiming that it is my burden to disprove your claims. My sources are in the article.)], so let’s see what they are.

(1) Mateas, 2001.  This is of limited use as a source, because no information is given to follow it up.  But a bit more research ( makes one doubt the relevancy of this source.

(2) Hmm.  This source is himself.  You can and should, of course, point out other things you’ve written that your readers might also be interested in.  But you saying something repeatedly is not proof.

(3) (-C says: The insight that the agency is not always in deciding the direction of the story was noted by Jason Alexander).  But Jason Alexander is not saying that inclusion of any agency prevents railroading.  Far from it, in fact:
I just got done running the most heavily railroaded session in probably my last 15 years of gaming, including heavily forced scene transitions and huge dollops of illusionism. 
(Context: It was a dream sequence being experienced by a comatose PC. They were taken through a highlight reel of their memories — both the ones they’ve experienced and the ones their amnesiac character has forgotten — with the other players jumping in to play current and former versions of themselves in a kaleidoscopic dreamscape.) 
I bring this up because I think it’s given me a fresh appreciation for why combat encounters — particularly those in “delve format” adventures — have become so overwrought in the past 10 years: It’s because, in a culture of “storytelling” GMs with railroaded plots, the combat encounters are the only place where players can actually experience freedom; where their choices actually matter. 
So you get a large class of players who are primarily focused on the combat encounters because that’s where they’re actually allowed to experience the true joy of roleplaying games (and, therefore, that’s where they have fun). And to cater to those desires, adventure design (and then game design) focuses more and more on making those encounters really exciting. 
But then, as that cycle degrades into itself, we end up with a situation where the tail is wagging the dog: Where the railroaded plot that strings together the combat encounters becomes thinner and thinner as more and more effort is put into propping up the combat encounter tent poles.
So, the sources used to prop up the argument either do not do so, or directly contradict the argument –C is making.

A Question of Motive

Contrary to –C’s accusation, I’ve read the article.  I’ve followed the links.  I understand what is being said.  But it doesn’t make logical sense, it flies in the face of common usage and experience, and the links supplied offer nothing supportive of the argument.  I’ve pointed out some of the problems in what I think are fairly gentle ways, and I am getting an increasingly hostile response.

So here is the point where my spidey sense starts to tingle, the same way it does when confronted by a used car salesman/politician/corporate wank.  And I cannot help but notice that –C has given us clues to a possible motive other than clarity of language:

  • I have often seen these terms applied to JRPG's like Final Fantasy or to situations where a player says "Let's run this module or adventure path." These are not railroads.
  • Final Fantasy games aren't railroads, because the agency is in how you level up your party and fight the battles.
  • This means that if you like knowing where the story is going or you enjoy playing in role playing adventure paths, this does not mean you are a fan of railroading.
  • [T]his terminology exists without needing to overload the word railroad.  Module. Sandbox. Adventure Path. Series:Episode:Scene. Campaign.
  • We don't have to worry about reasonableness, because no matter what game we are playing the choices the players are making, I.e. their agency within the game, is explicit and quantifiable, because games are designed.
  • What you say is not true of my megadungeon campaign, for example. They can't 'do whatever a human can do'. They pick from a menu in town, and move according to the rules in the dungeon.
  • I assure you, I'm role-playing more than once weekly, and in no game I play or run in, is this [creating a game around immersion in verisimilitude] a priority, desirable, or anything that anyone I play with is interested in. 
  • That isn't a subjective or an emotional question. It is a quantifiable one that is answerable in every case given enough information.
  • The term is not subjective. It refers to a specific quantifiable action that can be taken. If people misuse the term, that is certainly something that occurs often, but it doesn't change or remove the actual meaning of the word, or its usefulness in discussing design.

 (If I need to, I will discuss in another post that the subjectivity of all language is extremely well known, with many papers being written on the subject.)

These things made me suspect that –C was trying to “prove” that the games he was interested in were not railroads, because the term “railroad” has a pejorative context.  If you simultaneously believe that what you enjoy are railroads, and railroads are bad, then something’s gotta give.

This is an extremely common circumstance, and is called cognitive dissonance

Hence the two other unstated assertions:  (5) How the author is going to define these terms is correct definition, and (6) You will look like an idiot and inhibit communication if you disagree.  The goal is not to discuss what the terms mean, but to tell you what they mean and reinforce your acceptance of these definitions.

To my mind, -C’s comment to Part I of this series reinforces this conclusion: 
The 'motive' is I was having a discussion with people who were claiming that they 'loved railroading'. They then said they really liked games where they didn't have narrative control. As a designer this is nonsense. There was no reduction in their agency. So how were they getting railroaded?
It is interesting to me that -C accused me of fallacies he was engaged in, while accusing Telecanter of post-rationalization.
In psychology and logic, rationalization (also known as making excuses) is an unconscious defense mechanism in which perceived controversial behaviors or feelings are logically justified and explained in a rational or logical manner in order to avoid any true explanation, and are made consciously tolerable – or even admirable and superior – by plausible means. 
Now, it is obvious that one could easily claim that what I am also doing is rationalization of my "rude behaviour" in strenuously objecting to the redefinition of these terms on -C's blog.  Okay, that would be fair. I have tried to demonstrate, on a surface level, some of the problems with -C's construction, and I hope that what I have done will stand up to rational scrutiny.

I know, from -C, that I have parsed his argument extremely well (his words were, "This is an extremely accurate presentation of my points.")  I know that I understand the argument, and that understanding the argument does nothing to fill in the gaping holes in the argument.  This is a non-starter.

Why It Matters….Or, the Value of Definitions

Does it matter how we define terms like “railroading” or “player agency”?

-C suggests that it does, and I agree with him.  Having a clear working definition allows “designers to communicate clearly about the structure of a game.”  It also allows players (or potential players) to communicate with designers, and with each other.  –C says that “Communication is about shared meaning”, and this is true to an extent.  But communication is about shared meaning that arises from some sort of consensus; it is not imposed from above.

For example, the dictionary does not define words in the way that they should be used.  Rather, the dictionary defines words based upon their actual use by actual people.  This means that languages drift and evolve over time.  Common words become less common, or die, and sometimes words change meaning.  Words have both a connotative and a denotative meaning, and both of these can and do change over time.  Attempting to conflate one meaning with another, in order to transfer the properties of one meaning with the other, is the equivocation fallacy we talked about in the last part.

It should be clear that where terminology is concerned, we can attempt to add terms, and we can attempt to define terms.  If we are attempting to add terms, we do not want to use an existing term, unless the new term is a clear expansion of, or has a clear relationship to the old term, so that the new term does not increase confusion.  If we are attempting to define a term, we want to get at an understanding of how the term is actually used, so that we are not accidentally (or purposefully) actually creating a new term which we are then conflating with the old.

Because if we are attempting to redefine a term, and then equate the new definition with the original meaning, we are clearly and knowingly engaged in the equivocation fallacy.  We have experience with politicians, corporate mouthpieces, and others who are involved in this type of equivocation on a daily basis.  The purpose is to control a conversation, by redefining terms in such a way as to prevent an opposing view from being clearly articulated.

And we do not want that.

Noting that many players seem to dislike railroading, we do not want to inhibit their ability to discuss what they dislike.  Rather, we want to expand the discussion so that we can better understand what is meant.  Knowing the above, we are very careful to avoid suggesting that common usage is instead mis-usage.  We are very careful not to suggest that those who are not satisfied by our attempt to get at a working definition look like idiots.  We are, instead, open to their criticism.

We believe that definitions matter.


Now, there are a host of other spurious claims in the statements I listed, such as the idea that no one likes their agency removed.  And I could scour psychological journals and web sites to find links that demonstrate that, no, some people do like their agency removed.

Or I could leave you with this excellent post, which is as well (or better) thought-out than anything I could write anyway:

Or I could post once more, talking about what railroading is, and how to avoid it.  But I’ve already done that in this series of posts:,, and

Unless you really, really, really want me to keep beating this dead horse, I’ll do one more post, and then call it a done.  I thought that I made a record of the posts on Hack & Slash prior to their deletion, but the file was overwritten by a later version in which the comments were already deleted.  So, just some concluding thoughts and on to more interesting things.

(We hope.)


  1. Nice analysis. You remind me of some programmers I am friends with.
    Couple of comments on the cosmetic appearance of the article. A large percentage of males in the world are red, green colour blind. The red text on a black background is hard to read.
    It appears that there is a duplicate link in the last three links you present. Part 1, Part 2, and then the last one is Part 1 again.

  2. Thank you for your comments. I fixed the link. I wasn't aware of the red/green problem.....I was using the colour codes to make it easier to read, not harder. Have to give that some thought.