The Actual Work and the Implied Work

Comments on Alex Fisher's "Truth in Interactive Fiction"

Alex Fisher, “Truth in Interactive Fiction,” Synthese, Vol. 200, No. 6 (2022), pp. 1-18.

An important, and unavoidable, topic in philosophy of fiction is what it means for something to be true in a fiction, or fictionally true. In order to give an adequate characterization of any kind of fiction, we have to have an account of what the content is, and this is tied to the question of truth in fiction. Interactive fictions are particularly worth considering in this light because they are set up in such a way that the participant/reader/audience/player is a source, not merely a recipient, of fictional truths. Thus instead of a single set of ‘fictional truths’, we get a branching structure in which the fictional truths depend on the choices of the participants. Fisher’s article is an attempt to get a better grasp of this branching structure itself. In doing this, he will also incidentally shed some light on issues involving branching modalities (such as we sometimes use in talking about time or counterfactuals).

Fisher starts with an attempt to generalize the ideas of Marissa Willis on games vs. playthroughs, which we’ve previously discussed. This provides a very plausible first beginning, although Fisher will end up arguing that it has a number of inadequacies. We can distinguish two kinds of fictional truths:

It is work-true that p iff p is true in all branches of the work.

It is branch-true that p iff p is true in some but not all branches of the work. (p. 3)

Such an approach lets us distinguish what is fixed (work-true) from what is variable (branch-true) in interactive fictions. It also lets us sort out what is the responsibility of the originator of the work from the responsibility of the one engaging in it. And, of course, it captures the aspect of the work that is disjunctive rather than conjunctive, preventing us from having to regard it as a mess of contradictions. However, Fisher argues that, as it stands, it causes problems for thinking about fictional truths about future events.

A common way of thinking about the future is to take propositions about it to be neither true nor false; they are incomplete. This corresponds to thinking of interactive fictions as incomplete things in need of completion. However, this way of thinking would cause problems for the work-true/branch-true distinction:

For instance, in The Witcher III, the player is given various choices, but in every complete branch the protagonist Geralt ends up reaching Novigrad. If we take propositions about future fictional events as neither true nor false, then in a branch in which the player gives up playing after five minutes, it is not true that Geralt reaches Novigrad, therefore it is not work-true, since it is not true in all branches. (pp. 5-6)

After all, one of the possible outcomes to any interactive fiction is the participant not finishing the interaction, and this is especially important for understanding video game fictions. This possibility of incompleteness is going to wreak havoc on a distinction that depends on what is true in all or some but not all branches of an interactive work.

We need among other things to modify the notion of work-true. Fisher considers, and rejects, a further suggestion by Willis of work-true as being true in at least one branch and never contradicted. What we really need to do is handle the incompleteness. One way to do this is by saying that fictional truths about the future can be true or false. Thus propositions are work-true if they are inevitable in the full course of the interactive fiction. This handles the problem, but it also, ironically, has difficulty with non-interactive fiction, since we have some reason to think that fictional truths about the future are indeterminate in non-interactive fiction; if you read A Study in Scarlet, at the beginning Holmes and Watson have not yet met, and it seems reasonable to characterize this by holding that at the beginning of the story there are many possibilities and there are no truths about the meeting yet.

Fisher, however, argues that there is another option, namely, to deny that there are really incomplete branches, at least at the level in which we are drawing the distinction between work-true and branch-true. That is, failures to finish just don’t count. Branches should be seen as something like complete ways the fiction can go. (In this sense, they are analogous to possible worlds.) When we fail to complete the fiction, this isn’t a distinct branch; rather, failures to complete are described by families of branches rather than a single branch, namely, all the ways the fiction can go that are consistent with our interaction up to that point. This lets us make the work-true/branch-true distinction without having to make a decision about whether the future is open or closed.

There is another kind of case, however, that could cause problems for the work-true/branch-true distinction. Interactive fiction (and especially video games) allows for mistakes, inadvertences, glitches, and other unanticipated, and sometimes unanticipatable elements. One might very well be inclined to think that these can generate fictional truths; for instance, that on a playthrough your character got stuck in a wall. There is a bit of oddness to this, since we often just ignore typographical errors in writing.¹ Nonetheless, with video games particularly this does not always seem to be a real option. Glitchy characters can be part of how you actually play the game. Part of this seems tied to the fact that video games are not merely interactive, but self-involving. The player actually interacts with the glitch in the course of the story. Participants in interactive fictions can establish glitches, mistakes, and errors as part of the fiction by treating them as part of the fiction.

If glitches and the like can become part of the interaction, and thus true in the fiction, however, this creates a problem for the work-true/branch-true distinction. That distinction, remember, depended on being able to say that some things are true in all branches, or in some but not all branches. But this requires a manageable grasp on ‘all’, which becomes problematic if even typos and programming errors can cause a branch. It can make things that are surely supposed to be work-true merely branch-true, because the error can deviate from the way the fiction is set up. It need not even be an error – video games can be structured as to include secret and deviant branches, for instance, that are not there as ‘the game’, properly speaking, but are fun alternatives or weird supplements or mere jokes planted for the enjoyment of the players. They are put into the game deliberately in order not to be the standard game. Likewise, ‘cheats’ can be planted to bypass standard portions of the game. If these are counted as branches, then things we would ordinarily think of as work-true become merely branch-true.

On the basis of these considerations, Fisher suggests that we need to recognize that there are actually two works: the actual work and the implied work. When we are talking about fictional truths, we can be doing so with respect to either. Mistakes, glitches, and the like pertain to the actual work, and can contribute to its fictional content; the implied work, however, only captures what is intended by the creators, or, perhaps, our best hypothesis about what they intended. Despite the complications of an authorial-intention account of truth in fiction, this distinction gives reason to think that it is best not to reject such an account entirely, because intention does seem to do real work in our understanding of fictional truths. One area in which this distinction between the actual and the implied work seems to make a difference, Fisher suggests, is in discussions of the ethics of actions in video games, in which it is often relevant whether the player has to do something or not. This is properly often something that belongs to the implied work rather than the actual work.

Fisher also suggests that things like secret joke endings put in by the creators (as opposed to standard alternative endings within the main track) should be seen as parts of the actual work, not the implied work. This seems much less likely to us, and to suggest instead that the implied work is able to be quite a complicated thing. Indeed, it seems plausible that implied works can include other implied works, and that particular implied sub-works can be privileged within the overall implied work. If we think of implied works as broadly deontic in character, we can see these secret joke endings as ‘non-obligatory permissibles’ that take into account the fact that players attempt the game many times; they should perhaps be seen as allowed detours in a broader interaction in which it is understood that the ‘obligatory’ elements of the game are not actually optional for intended playthrough. This still preserves the distinction: A given actual playthrough can involve the player detouring in the actual work, in a way directly allowed by the implied work, without affecting what the implied work requires, because while the ‘deontic structure’ of the actual work is specifically built out of what happens in the playthrough (including glitches), the ‘deontic structure’ of the implied work accommodates indefinitely many actual works, including some that can be be regarded as just entertaining diversions from the overall game.²

The distinction between the actual work and the implied work does seem to be necessary for handling many aesthetic situations. It’s unclear to us, however, that it actually salvages the work-true/branch-true distinction, at least without recognizing ‘work-truths’ as being at least partly deontic in character, and thus by that very fact not definable in terms of ‘all branches of the work’. More work needs to be done on this, however.

1. And, to add to this, such errors are easily seen as a source of jokes rather than a serious feature of the story. In the 1991 movie, Delirious, Jack Gable, played by John Candy, finds himself apparently stuck in the soap opera for which he is the lead writer and discovers that he can re-write the story from inside. In trying to do so, he accidentally makes it so that at an upscale, high society, the butler starts bringing in reindeer (‘cold deer’). To which Gable responds, “Not deer! Beer! It’s a typo! Use your head, for Christ’s sake!”

2. Putting it this way perhaps suggests that the issue of what the actual work requires, as opposed to what the implied work requires, might be related to contrary-to-duty paradoxes in deontic logic. An example is what is known as the gentle murder paradox, or Forrester’s paradox. This paradox arises from the fact that we seem to be able to affirm simultaneously statements of the following structure:

It is obligatory that A not murder B.

It is obligatory that, if A murders B, A murder B gently.

In a situation in which A does murder B, however, we seem to get the conclusion that it is obligatory that A murder B gently, which is inconsistent with the first statement. If we carry over the analogy, the deontic structure of the implied work seems to be like the first statement (e.g., you must do such-and-such to win the game), but the actual work allows ways of viewing the game with deontic structure like the second. You have what the game prescribes as intended, but in actual playthrough, you might have a glitch (like A actually murdering B), and you have a question of what the game prescribes (or allows) in light of the glitch. There seems to be something to this analogy, but it likely requires a great deal more precision about implied works and actual works (and perhaps obligations) to be adequate. But a view with multiple deontic structures seems attractive on both sides. Consider the multi-level view of obligations sketched out by Henry Prakken and Marek Sergot, “Contrary-to-Duty Obligations,” Studia Logica: An International Journal for Symbolic Logic, Vol. 57, No. 1, Papers in Deontic Logic (July 1996), pp. 91-115. But on the other side, confer the comments on the disanalogy between game rules and moral rules in Roderick Chisholm, “Contrary-to-Duty Imperatives and Deontic Logic,” Analysis, Vol. 24, No. 2 (December 1963), p. 35.