Bugs: When the Rules are Wrong

Count Timothy von Icarus

Summary:

People can think rules are are "wrong." This isn't just, "the rules produce bad outcomes," but rather the rules seem "broken" themselves.

"Bugs" in games are a good example of this. I would argue that "bugs" can show up in social practices as well, or even natural languages, and that the desire to "fix bugs" and have rules align with our expectations of what they should be drives rule (and language) evolution. In some ways, this puts intentionality back in the driver's seat in determining meaning, but not in a straightforward way that recovers the full intuitions classical theories of meaning.

Full post:

I wanted to look at cases where communities think their rules are "wrong," (i.e. broken, "bugged," as opposed to simply "producing bad outcomes.") This seems to have implications for Kripke's version of Wittgenstein, where an assertion that a rule that is being followed is justified by the expectations of other rule followers.

Particularly, I wanted to look at the case of "bugs" in video games, because they leave out a lot of the political and ethical baggage that comes in when we consider "bugs in the tax code," etc. Also, video games have the benefit of having canonical rules that are faithfully executed by a computer. The bugs of interest are those that result in unintended, counter-intuitive processes, not those which cause programs to crash or obvious glitches.

Example of Such a Bug:

In the 1996 classic, Diablo, there is a mechanic whereby, if your character gets hit particularly hard by a monster, you go into a stun animation during which you can't move or attack. This is disadvantageous. It can lead to be surrounded by monsters and being "stunlocked," unable to act until you are killed.

Due to how the decision tree works for calculating if the stun animation should play, there is a way to avoid the possibility of being stunned. If a hit would have killed a player by reducing their health to zero, the animation won't play. But there is also a "mana shield" spell that causes damage to be subtracted from magic points before health points. Thus, by lowering their maximum life permanently to 1 (possible), characters can be made immune from stuns.

Discussion:

This is a case where we have a canonical set of rules and rule following is ensured by computers. How can it be that we have a rule that the community and the developers think is "wrong?" Not bad, mind you, but "working incorrectly," although the bug does not involve any mistakes in the execution of the code. It's a good example because it is relatively simple, based simply on where a check is made in a decision-tree.

IDK, maybe it's trivial, but it seems to get at something about how rules are developed, defined, and refined that they can be seen as "wrong" in this way. When it comes to the realm of human rule following, there is always ambiguity about if a rule is being applied correctly, so the computer game is a nice example because it can make us sure that it is truly the rules, not how they are applied, that is "bugged."

What is a "bug" then? It seems to be cases where the rules don't "make sense." The interactions between rules fail to meet our expectations. For example, there is no sensible reason for the ability to get around the stun function in Diablo; nothing about how the rest of the game works makes us think it should work this way. It's an unintended interaction that seems to work against the goals for which the rules were created. It counterintuitively makes it beneficial to have your character's life permanently lowered. One of the game's most severe penalties becomes an asset. Stats that are supposed to guide you on "what to do" and "what is good" turn out to be misleading because of the bug. Less life ends up being better than more. There is a lack of coherence here. This is different from a bad design choice, those perfectly coherent decisions that make a game worse.

Board games have a lot of this too. Deck building games might ban certain card combinations from tournaments because following the rules correctly seems "wrong." It defeats the purpose of the game as a game of skill, or destroys its pacing, etc.

This phenomena has implications for social practice. It seems like developers are more apt to change something if it is seen as a bug, because then it isn't an issue of fixing the rules for balance (often controversial), but of "making the rules work correctly."

---

Now, if rules can be widely considered "wrong" in this way, even when people keep following them (or are forced to by computers), what does that say about rules being established by the weight of social repetition? Do we have similar phenomena in natural languages where people disagree with standing practice? Would this be akin to how some established words end up getting intentionally removed from everyday use, e.g. "retarded," or "oriental?" The move from "Mrs. and Miss" to "Ms."? The move to make "they" and "their" singular pronouns? The organized drive to change "Kiev" to "Kyiv?"

More obviously, there seems to be "bugs" in laws at times, ways for laws to interact in unintended and ludicrous ways. "Division by zero should be infinite," might be another example.

The relationship seems similar in some cases. For example, using Mrs. and Miss was proper rule following, but in the face of changing gender roles it began to challenge expectations about what the rules should be. If a rule becomes seen as a hindrance to communication, or stupid, clashing with people's expectations, people stop using it - they "patch the bug." The expectations of other language users doesn't just help to fix meaning then, it also seems like a likely engine of language evolution. After all, the rules governing the meanings of individual words don't stand on their own, they're are part of a system, and old rules can come to seem "broken" due to changes in other rules.

Obviously, we can try to change social practice re word usage ("Kiev" to "Kyiv" was one coordinated effort), rejecting current social practice and starting our own form of practice. But people will only know what we are talking about if it "catches on."

This ends up circular, but not in a bad way. Words have their meanings because of how people use them, but then how people use words ends up being driven (in at least some cases) by what people want/intend them to mean.

I also think the phenomena of "wrong" rules is a reminder that rules ultimately are also defined and refined in terms of their purposes, which is easy to recall with games, but harder with natural language and mathematics.

Count Timothy von Icarus

Passive voice might be another example. It's perfectly grammatical, and should be used in some cases. It can also be used in order to make things ambiguous on purpose, often to avoid assigning blame, e.g., "mistakes were made."

But, because it often cuts against the purpose of language vis-á-vis clear communication, many people want to essentially remove it as an option in most cases. That is, it is fine according to the rules, but people see something wrong with it despite this, and this wrongness relates to the function of the system. Not a perfect example, but a common one. Natural language bugs will necessarily be hard to define due to the lack of canonical rules and faithful execution.

Ludwig V

I also think the phenomena of "wrong" rules is a reminder that rules ultimately are also defined and refined in terms of their purposes, which is easy to recall with games, but harder with natural language and mathematics. — Count Timothy von Icarus

"Purposes" has to be interpreted liberally here. It will work in some cases, but not in all. For example, I don't think it works very well in the context of a game. (An "exercise" in the context of teaching or training or practising is different.) I prefer to think of the "point" of the rules. The problem with the work-around you identify is that if only some people know about it, it is an unfair advantage, and if everyone knows about it, it make the "stun" feature pointless.

This ends up circular, but not in a bad way. Words have their meanings because of how people use them, but then how people use words ends up being driven (in at least some cases) by what people want/intend them to mean. — Count Timothy von Icarus

This isn't circularity. It's feedback.

Also, video games have the benefit of having canonical rules that are faithfully executed by a computer. — Count Timothy von Icarus

In a sense, that's true. But both the software and hardware are designed and built to produce certain results which are meaningful in the context of human life and practices. So the ultimate foundation that Wittgenstein arrives at "This is what I do" does apply.

Natural language bugs will necessarily be hard to define due to the lack of canonical rules and faithful execution. — Count Timothy von Icarus

I like your example here. As a matter of style, there's a lot to be said for avoiding the passive unless it is unavoidable just because it is vague. It's to be expected that natural language will be messy and complicated, and also that people will find and adopt ways of using it in problematic ways. Although one might argue that ambiguity is often useful or desirable in pragmatic ways. Of course, whether they are desirable or not will depend on your point of view.

The most profound consequence of all of this is that it tells against the approach to language as a complete consistent structure with its own metaphysical existence. That model is not wrong, because we all learn language as a pre-existing practice and it can be useful, for example in logic or linguistics, but it cannot be fully achieved. The view that language is a practice amongst human beings and part of the human way of life is more helpful in many ways.

I'm not a fan of systematic analyses, but perhaps we could distinguish between three different kinds of problem here.
1. One is issues caused when a difficult or anomalous case turns up in the world. The discovery of black swans or of platypuses.
2. Another is the kind of discovery that has been so much evident in mathematics - irrational numbers, etc. The problem of what to do about "0" is perhaps not quite the same, but shares the feature that the standard rule don't apply. But it is the rules themselves (given the standard interpretation of them) that produce the result.
3. A third is where people take advantage of (misuse) the rules to achieve some thing that is not strictly relevant to them. The passive voice is one example, and the "fix" for your bug seems to me to be another.

Ludwig V

Our brains seem bugged when it comes to communication. Or perhaps it's both language and brain? — Vaskane

Maybe the expectation that either is a coherent system is a mistake. Evolution only requires that the systems work in normal circumstances. So quick and dirty ways of arriving at answers will survive so long as they work for the creature on which they are running.

I sometimes wonder whether "This sentence is false" or "the set of all sets that are not members of themselves" are bugs in logic, or language or in our brains. Or possibly whether they are not bugs in the sense that they need to be fixed - more like "0" and divisibility, we just need a rule to give us (or maybe refuse to give us) the answer.

Ludwig V

I suppose my next question is what is flawless without bugs? Even concepts of the most perfect thing "God," are riddled with bugs. "everything being a quick and dirty fix," I like that for a little comedy routine I've been toying with, it Starts off with "Y'all are some Naaaassty motherfuckas..." Ty for the inspiration. — Vaskane

Well, perhaps, "without bugs" is an ideal, a target that may well not be achievable.

Wittgenstein identifies what could be considered a bug with that idea - that we cannot think in such a world. "Back to the rough ground".

Count Timothy von Icarus

↪Ludwig V

This isn't circularity. It's feedback.

I agree that its feedback, but isn't feedback a sort of circular causality?

The most profound consequence of all of this is that it tells against the approach to language as a complete consistent structure with its own metaphysical existence.

Right, it's protean and dynamic, in the same way that organic life is.

The view that language is a practice amongst human beings and part of the human way of life is more helpful in many ways.

More helpful than what? A systems view of language? But then it's always been obvious that language is a social practice and this alone doesn't really elucidate any of the big questions in philosophy of language.

Attempts to unpack what "social practices" are seem to lead to more questions. E.g., if rule following is just based on "the expectations of others," what are we to make of apparent rule following in animal behavior, biology, and "law-like" behavior in nature? Are these different sorts of rules?

One example might be how Asian fireflies all blink in unison due to the rules males follow for deciding when to blink. These don't seem to be based on "expectations," but are rather instinctual, the result of each male trying to "blink first." But then the mathematics used to describe this synchrony ends up also describing how heart cells synchronize, even in petridishes, and here "expectations" are even harder to find — and then the same principles show up in earthquakes. Stogratz' book Synch has a bunch of great examples on this. For the doctor or biologist, defective heart cells inability to "follow the rules," ends up being defined in terms of function. The "bug" issue in games is interesting because these also seem to be defined in terms of function as well.

I do agree that attempts to formalize language in simple ways are not time well spent. E.g. Carnap-Hillel information. You're better off looking at the aggregate and emergent. Trying to pin down how the meaning of individual utterances works in formal terms is hopeless, like trying to give a theory of discrete actions in economics. Economics deals with complex human practices, but it works because it sticks to fairly high levels of analysis.

I'm not a fan of systematic analyses, but perhaps we could distinguish between three different kinds of problem here.
1. One is issues caused when a difficult or anomalous case turns up in the world. The discovery of black swans or of platypuses.
2. Another is the kind of discovery that has been so much evident in mathematics - irrational numbers, etc. The problem of what to do about "0" is perhaps not quite the same, but shares the feature that the standard rule don't apply. But it is the rules themselves (given the standard interpretation of them) that produce the result.
3. A third is where people take advantage of (misuse) the rules to achieve some thing that is not strictly relevant to them. The passive voice is one example, and the "fix" for your bug seems to me to be another.

:up:

I think this is a good classification. Although, they can also blend together a bit. E.g., the black swan causes us to discover the Type 2 problem, or the Type 2 problem opens up the possibility of exploiting incoherencies in a system.

sime

I think 'Kripke's Wittgenstein' was tainted by Kripke's semantic foundationalism. Kripke was correct IMO to conclude that Wittgenstein's concept of rule-following involves appealing to semantic criteria that are independent of the psychological facts of the rule-grasper (roughly speaking), due to the trivial implication that "to grasp" something entails a distinction of the grasper and the grasped thing. However, Kripke was wrong to conclude that the external criteria referred to assertibility conditions laid out by social convention. I think Kripke's incorrect conclusion about the later Wittgenstein's views was due to the fact that Kripke, unlike the later Wittgenstein, could not accept the non-existence of a universal and shared semantic foundation.
For Wittgenstein, any assertibility criteria can be used for defining the meaning of 'grasping' a rule, and not necessarily the same criteria on each and every occasion that the rule is said to be 'used'. And a speaker is in his rights to provide his own assertibility criteria for decoding what he says, even if his listeners insist on using different assertibility criteria when trying to understanding the speaker's words.

Ludwig V

I agree that its feedback, but isn't feedback a sort of circular causality? — Count Timothy von Icarus

Yes. It's like the difference between parasitism and symbiosis. To put it this way, they are the same phenomenon, except that parasitism damages the host, and symbiosis benefits (or at least does not damage) the host. Or think of the difference between murder and execution, which both mean killing, but in different contexts, or with different evaluations.

More helpful than what? A systems view of language? But then it's always been obvious that language is a social practice and this alone doesn't really elucidate any of the big questions in philosophy of language. — Count Timothy von Icarus

Yes, more helpful than a systems view of language. I'm not sure what was obvious before Philosophical Investigations and How to do things with words. Part of the point of such views is that they encourage us to consider the possibility that (some of) the big questions in the philosophy of language are the result of the systems view of language. We could call them bugs.

Attempts to unpack what "social practices" are seem to lead to more questions. E.g., if rule following is just based on "the expectations of others," what are we to make of apparent rule following in animal behavior, biology, and "law-like" behavior in nature? Are these different sorts of rules? — Count Timothy von Icarus

Wittgenstein took us only so far, leaving us to take the idea further. These are really interesting questions. I'm not dogmatic about the answers. But surely that the "higher" mammals, at least, are capable of responding to the expectations of others, because of the way they interact with us. Whether the same applies to, for example, the social insects or schools of fish is another question. Lots of difference cases, no expectation of a tidy distinction.
The short answer to your last question is Yes, of course.

I think this is a good classification. Although, they can also blend together a bit. E.g., the black swan causes us to discover the Type 2 problem, or the Type 2 problem opens up the possibility of exploiting incoherencies in a system. — Count Timothy von Icarus

:smile: Of course they blend and interact. I regard that as a feature, not a bug.

Kripke, unlike the later Wittgenstein, could not accept the non-existence of a universal and shared semantic foundation. — sime

I have the impression that Wittgenstein did think that "way of life" and "human practices" gave a shared context. If they don't, how could he think they explained how come we agree about the interpretation (application) of a rule?

For Wittgenstein, any assertibility criteria can be used for defining the meaning of 'grasping' a rule, and not necessarily the same criteria on each and every occasion that the rule is said to be 'used'. And a speaker is in his rights to provide his own assertibility criteria for decoding what he says, even if his listeners insist on using different assertibility criteria when trying to understanding the speaker's words. — sime

On the other hand, you are quite right that human life is as much the stage for divergence and disagreement as a shared basis of consensus. The importance of the idea is that human life is both a basis for agreement and the common ground that is necessary for divergence and disagreement to develop.

013zen

Attempts to unpack what "social practices" are seem to lead to more questions. E.g., if rule following is just based on "the expectations of others," what are we to make of apparent rule following in animal behavior, biology, and "law-like" behavior in nature? Are these different sorts of rules?

One example might be how Asian fireflies all blink in unison due to the rules males follow for deciding when to blink. These don't seem to be based on "expectations," but are rather instinctual, the result of each male trying to "blink first." — Count Timothy von Icarus

As you pointed out, there seems to be a clear distinction here regarding on the one hand 1. acts governed by instinct and 2. acts governed by expectation.

For example, when I produce a treat for my dog Otis, he involuntarily starts to drool - his body is simply following preprogrammed rules. When, however, I fail to immediately give him the treat, he raises his paw to "shake" since historically, that's gotten him a treat. He has come to expect it.

For the doctor or biologist, defective heart cells inability to "follow the rules," ends up being defined in terms of function. The "bug" issue in games is interesting because these also seem to be defined in terms of function as well. — Count Timothy von Icarus

I would say that these are not the same. A heart cell has a certain set of rules it follows, and when it fails to execute those rules it is faulty. When a video game, like the one you mentioned in the OP, Diablo, is facing a bug, its following the rules just not the expectations involved in its design. We can talk of expectations of a fully functioning heart cell, but this is an equivocation of expectation since one is the result of a creator making something and it simply not functioning in the intended manner due to what the code is telling it to do and the other is our expectation which is based on how normal heart cells typically function. One case is the result of our expectations and the other is the result of our observations.

I'm interested in your original post, but I am afraid I don't entirely understand your intent or what exactly you're getting at, but it sounds very interesting.

Ludwig V

A heart cell has a certain set of rules it follows, and when it fails to execute those rules it is faulty. When a video game, like the one you mentioned in the OP, Diablo, is facing a bug, its following the rules just not the expectations involved in its design. — 013zen

Yes. But I find it very hard to state the point clearly. I think we have to distinguish them this way. When my heart fails to fulfil it purpose - what it is "designed" to do, the fault is not in the design, but in the execution of them. When a bug arises in a program, there is a fault in the design of the game/program, not in the execution of the rules.

I see the basic issue of the thread as something like this:- How is it possible for the rules that describe a game to be wrong? and the underlying issue is about the way that a set of rules can produce surprising (counter-intuitive) results.

flannel jesus

How is it possible for the rules that describe a game to be wrong? — Ludwig V

Because the people implementing them didn't think about all the cases. There are usually edge cases they didn't think of, or interactions between features they didn't plan for.

Count Timothy von Icarus

↪Ludwig V

↪013zen

I would say that these are not the same. A heart cell has a certain set of rules it follows, and when it fails to execute those rules it is faulty. When a video game, like the one you mentioned in the OP, Diablo, is facing a bug, its following the rules just not the expectations involved in its design

Yes. But I find it very hard to state the point clearly. I think we have to distinguish them this way. When my heart fails to fulfil it purpose - what it is "designed" to do, the fault is not in the design, but in the execution of them. When a bug arises in a program, there is a fault in the design of the game/program, not in the execution of the rules.

Well, in the case of some genetic disorders, e.g. pseudo-exfoliation glaucoma, the problem seems to be with the rules, not the execution. With PEG, the problem stems from a single nucleotide polymorphism that codes for the elastin protein. Essentially, what happens is that this gene instructs the cell to produce this protein in a way that differs from "normal" elastin. This results in a mostly functional elastin that nonetheless clogs up the passage of fluids between different parts of the eye, resulting in higher interocular pressures, which in turn kills neurons in the optic nerve, eventually resulting in blindness.

If we think of the genes as rules, as the instructions for building the organism, the problem here is that the rules are wrong. The cells are producing the protein as instructed, but the slight variation in the protein leads to unintended consequences vis-a-vis function.

Of course, here we run into a problem. As pointed out in the "What Is Logic?" thread, it has become common to think of logic or any sort of rules as being the sui generis product of minds. They only exist "in here" not "out there." A theory of rules as grounded in human social practice sort of goes along with this tendency.

So, in such a view, we can't really say the rules are bugged in PEG the same sense they might be in a game. Physics doesn't make mistakes, teleology is suspect even if impossible to fully root out, and so we shouldn't put too much emphasis in terms of defining conditions in terms of "function."

My objection here is that this seems to create a rather artificial barrier between our understanding of rules and function/purpose, and the world. Where do we get out idea of rules from? Why are we equipped to understand such things if nothing like them exists outside our minds? Why do cells seem to follow rules and "execute programs?"

I guess this sort of gets at some of the controversy surrounding the Extended Evolutionary Synthesis. How strict should the commandments "thou shalt not invoke teleology" and "thou shalt not externalize rules," be? ("Thou shalt not have intentionality play a role in evolution," is sort of another.)

I'm interested in your original post, but I am afraid I don't entirely understand your intent or what exactly you're getting at, but it sounds very interesting.

I didn't really have a point, it just occurred to me as an interesting case. I was thinking that bugs like this might be the sort of thing that drives some aspects of rule evolution.

Ludwig V

Because the people implementing them didn't think about all the cases. There are usually edge cases they didn't think of, or interactions between features they didn't plan for. — flannel jesus

Yes. But how realistic is it to set out to think about all the edge cases and/or all the interactions between all the features?

flannel jesus

↪Ludwig V

Usually not realistic, that's why there are bugs.

013zen

As pointed out in the "What Is Logic?" thread, it has become common to think of logic or any sort of rules as being the sui generis product of minds. They only exist "in here" not "out there." A theory of rules as grounded in human social practice sort of goes along with this tendency. — Count Timothy von Icarus

This is a strange view, and not one that I subscribe to. Logic, seems to me, in some sense to be mind independent. The logical statement, that say, "A=A", or the argument of the logical form: "If A then B, B, therefore A", seems wholly independent upon any social practice.

The application of logic, however, is a social practice. Namely, what counts as "A" and so forth in the practice.

If we think of the genes as rules, as the instructions for building the organism, the problem here is that the rules are wrong. The cells are producing the protein as instructed, but the slight variation in the protein leads to unintended consequences vis-a-vis function. — Count Timothy von Icarus

So, here is, I think the important point.

As you note in the OP:

Deck building games might ban certain card combinations from tournaments because following the rules correctly seems "wrong." It defeats the purpose of the game as a game of skill, or destroys its pacing, etc. — Count Timothy von Icarus

I think it would be helpful to talk about it this way:

1. Any set of rules will have consequences.

2. These consequences can be beneficial or detrimental relative to accomplishing some function.

3. In the case of a game, the function is the developers' intention

4. In the case of nature, the function can be taken to be survivability.

So, you give an example where a mutation causes a variance in the manner in which a cell produces a particular protein strand which results in clumping of those strands within the eye leading to blindness.

Clearly, the rules are being followed and have consequences; insofar as those consequences are detrimental towards survivability, we can construe them as "bad" or "wrong"... but, remember that survivability is a function of the environment; we can easily imagine an environment where, perhaps, such a mutation isn't detrimental. We could call it benign. Or even an environment in which its somehow beneficial, who knows.

This is why mutations drive evolution relative to some environment.

But in order to determine whether or not the rules being followed are beneficial or detrimental, we need to speak relative to some function that we supply in our interpreting them as such.

Bugs: When the Rules are Wrong

Welcome to The Philosophy Forum!

Categories

More Discussions