you may not have all you want, all you need
all that you have had was given by some hand
it could change, it could stay the same
who will say? who will make the claim?
the situation we are in at this time
neither a good one, nor is it so unblessed
it can change, it can stay the same
I could say, I could play the game~ Gentle Giant, “Proclamation”, The Power and the Glory (1974)
This essay attempts to explain power, responsibility, reputation, and courtesy in the context of long-distance communication-oriented relationships.
- The Boss: You owe Slim. Slim owes me. You owe me.
Jason Smilovic, Lucky Number Slevin (2006)
To start, what is power? Using more grammar: what does it mean for one person to have power over another? Too many words have been spilled on the topic, so I will demystify it by suggesting that power is a synonym for responsibility; one person has power over another person precisely when they are responsible for that person. To have the power of life and death is to have the responsibility for choosing between life and death; to have the power of where to go for lunch is to have the responsibility of choosing where to go for lunch.
An immediate and common retort is that power is hard or soft; hard power is physics and soft power is responsibility. A variety of misunderstood statements attesting to the distinction are well-known; here are a few common examples:
Every Communist must grasp the truth, “Political power grows out of the barrel of a gun.” Our principle is that the Party commands the gun, and the gun must never be allowed to command the Party.
~ Mao Zedong, 1927
Some people think they can outsmart me. Maybe. Maybe. I’ve yet to meet one that can outsmart bullet.
~ Heavy Weapons Guy, 2007
- We are all victims of physics.
unknown 21st-century scientist
I will directly answer this request for nuance by insisting that physics is local: actions taken in a location in the past take time to propagate to other locations in the future. Locality immediately requires physical existence to lack realism by the Kochen-Specker theorem; it is thus indefinite and probabilistic. I will further insist that we are responsible for the choices that we locally make in our immediate surroundings; we have the power to alter the space near us, and we choose how to use this power partially by imagining the responsibility of control over that space. However, this allows me to note that power does not grow out of a gun, but out of the responsibility associated with aiming, firing, and maintaining a gun in a particular location; any claim about the hard physics of power can be recast as a claim about the soft responsibilities of power which invoke those physical consequences.
If there were only one person then they would be responsible for themselves, and thus have power over themselves. Moreover, if there were two distinct people and one had power over another, it is almost never total power; that is, not all responsibilities are delegated from one person to another even in a situation of total surrender or helplessness. I don’t say this to make light of such situations, but merely to point out that instead of a singular power, we ought to discuss many multiple different powers which can all be independently delegated.
A responsibility may be delegated from one person to another; thus, power may be delegated. Moreover, for any responsibility already delegated to a person, they have the power to further delegate it; thus, delegation is itself a source of responsibilities. We will investigate the structural consequences of this curious recursion. First, we will do some mathematics.
A category, mathematically speaking, is a collection of directed arrows which admit composition in the sense that any two compatible arrows may be combined into a single larger arrow. To record that compatibility, every arrow is associated with a source and target, such that composition is only possible when the first arrow’s target is the second arrow’s source. Here, we will consider a category of responsibilities whose arrows are delegations of power from one source person to another target person. We will not think of any category in particular, but consider properties of any category of responsibilities.
There are some expected properties of categories. For each source/target, which mathematicians call an object, there exists an identity arrow having that object as its source and target which does nothing. Moreover, for each arrow, composition with these identity arrows, whether precomposition with the source or postcomposition with the target, is the identity operation; composition with identities does nothing. To meet these requirements, for each person there is an identity arrow for which they are the source and target and which delegates no power whatsoever, such that composition with an identity arrow delegates exactly what was delegated before.
A transitive relation R is one such that if xRy and yRz then xRz. For example, if x < y and y < z then x < z. We will add similar transitivity into our category whenever a relationship recurs between people, although it is required to be exactly the same relation in all three cases. For example, returning to the beginning of this section, if Alice owes Bob and Bob owes Carol in kind then Alice owes Carol in kind. Intuitively, this transitive property allows any power delegated to a target person to also be a power delegated by a source person, and a delegation with an intermediary is equivalent to a delegation without an intermediary.
That doesn’t sound right, though. An intermediary could choose to change their current commitments, giving the system an extra degree of freedom in its dynamics. What an individual category actually represents is a possible snapshot of the static delegations at a moment in time. If, at that moment in time, Alice happens to be paying Bob and Bob happens to be paying Carol, then Alice happens to be paying Carol; but that can be made contingent on points in history where Alice and Bob began to pay Bob and Carol respectively. This structure is captured by a sheaf of categories of responsibilities; a sheaf is a systematic annotation of some topology, and we will annotate each point in topological spacetime with a category describing the delegations at that point in spacetime. Physicists call such sheaves “causal sets” or “causets”.
For posterity, mathematicians call a sheaf of categories a stack, but the distinction here is unhelpfully pedantic; the annotation of topology is what we want to understand.
In order to intuit how responsibilities change over time, we will consider how spacetime evolves. I will assume that spacetime is partially ordered, as with causets, because it greatly simplifies how events can interact. Indeed, we really only have one interesting case: when one event precedes another event, how can responsibilities shift between the events? A delegation could stay the same, be created, or be revoked. Staying the same is allowed by the arcane rules of sheaf theory, which state that moving forwards in time should naturally map to undoing changes in delegation; the rules also allow revocation forward in time, which naturally maps to undoing an existing delegation. So, the only non-trivial case is creation of delegation. This might not be intuitive, though; let’s try two intuitive lenses.
First, let me use less mathematics. When we look at our current situation, and consider the past events which led to it, we can think of the relationships which currently exist as also being related to versions of those relationships in the past. If two people are related to each other today and in the past, without any interruptions, then that relationship has been constant over time. If they are related to each other in the past but not today then there is the natural and sad explanation that they drifted apart over time and ceased to relate to each other. However, if they are related to each other today but not in the past then there is no natural way by which they are introduced to each other; all we can say is that they were at the same place at the same time.
And now, I shall use more mathematics. Let C be some partial order of spacetime events and study it with the stack R : Cop ⇒ Cat. If the reader is worried about size issues then they may use FinCat as codomain since there are only finitely many humans. In general, R is a prestack merely by having that signature. In order to be a stack, for any two events U and V, where U → V and there is a family of covering events X on V such that U → X → V, the obvious diagram must naturally commute. This amounts to requiring that there is a natural choice of functor F : R(V) → R(U) which factors cleanly — restricts, in sheaf parlance — through any intermediate R(X) in V’s covering family. F is constant anywhere that V has not changed compared to U; indeed, F must be fully faithful on the image of U in V because C is thin. Relationships in R(U) but not in R(V) are not in F’s image; we do not have to account for relationships which faded by the time of V. So F only varies for relationships in R(X) which are newly occurring in X compared to U, or in V compared to X. However, those relationships usually will not be unique and so cannot be naturally chosen, as we shall see, so F is usually not natural either.
By “natural” I mean what category theorists mean: something is natural when the only ways to do it are all equivalent to each other. A natural mapping is the only such mapping, or perhaps it is one of many such mappings but all such mappings are equivalent. A natural choice is the only such choice, or it is one of many choices but all choices are equivalent. This may seem like a strange point now, but it will occur repeatedly later and is worth explaining early.
To summarize, when we treat power and responsibility as components of the same sort of system and equip it with transitivity whenever possible, the evolution of that system over time has natural explanations for why power and responsibility either remain delegated or are eventually revoked. What are not naturally explained — thus open to discussion — are the paths which initially create delegations.
The capability paradigm is about access control. When a system, such as an OS or a website, is presented with a request for a service it provides, it needs to decide if it should actually do what the requestor is asking for. The way it decides is what we’re talking about when we talk about access control. If you’re like most people, the first thing you’re likely to think of is to ask the requestor, “Who are you?” The fundamental insight of the capabilities paradigm is to recognize that this question is the first step on the road to perdition.
~ Chip Morningstar, What are capabilities? (2017)
Focus in particular on the common situation of a newcomer getting to know an existing group of people. How does the group extend its existing power structure to accommodate the newcomer? In one way or another, the newcomer must establish that they have some sort of reputation or bona fides which serve as a basis for further delegation of responsibilities. Some examples are required before we can get to the heart of the problem.
Consider: A group of people chat among themselves with a central server. They each individually delegate the responsibility for the contents of the chat window to the server, which is also the power to receive chat messages. The server symmetrically delegates to each individual a power to write chat messages, which is also the responsibility for the contents of those messages. A newcomer would be introduced to the central server, which simultaneously extends its delegation while accepting the newcomer’s own delegated window.
Consider: A group of people chat among themselves with federation. They each delegate the responsibility for delivering messages to a mailbox, which is the power to send messages to that mailbox, to their peers. The newcomer is introduced to each individual peer, performing the simultaneous exchange of delegations.
This is quickly becoming word salad. We need a definition which allows us to repeat the pattern of delegation. A capability is a communicable unforgeable invokable token of responsibility designating a delegated power. Communicating a capability implicitly copies it, unlike with physical tokens. The responsibility associated with holding a capability is always the responsibility of invoking, copying, forgetting, and further delegating that capability. Our chat examples now involve users either exchanging capabilities with a central server, or exchanging capabilities between each pair of peers.
I do not want the reader to feel mystified here, so I will quickly demolish the most common myths and misunderstandings about capabilities. Not all powers are capabilities; we say that a power is tamed when it is represented by a capability, and there is ongoing research into theories of taming. Capabilities are often only copyable on an information-theoretic level; treating a physical fob as a capability is only reasonable in the presence of a fob-copying machine. Conversely, most implemented capability-aware systems only implement unguessable capabilities, which are forgeable by anybody who knows certain secret information, rather than unforgeable capabilities. Finally, the ability to revoke a capability is available without loss of generality, although it works in the naïve fashion: a revocable capability maintains a reference to its creator for the express purpose of checking whether it has yet been revoked.
Speaking of creators, every capability has a creator. This should intuitively accord with the categorical approach as follows. Every person is associated with an object in our category of responsibilities and they start with all of their own responsibilities and powers delegated to themselves. This self-delegation gives each person their identity arrow. As the beginning of this section suggests, those creators may not be human, and we already account for that in our everyday experience. They may be computers or corporations, for example. A capability theorist will point out that capabilities denote authority, not mere responsibility; the creator of a capability passes on a stamp of their identity along with every copy of that capability. In practice, this ends up as an implementation detail, but it is essential for tracking how a capability is passed through a system, since such stamps identify either the original author or an intermediate author who has inserted themselves into the chain of creators.
Consider: A vending machine is powered on and stocked. A newcomer would be introduced to the vending machine, which proffers a capability for selecting stocked goods which was created by the machine’s owner. On a modern vending machine, the newcomer would submit their own capability for payment processing, authored by the newcomer’s financial institution, which gives the vending machine’s owners the power to request money and the responsibility to request the negotiated amount of money.
These examples have all been subject to contractual expectations, which range from broad expectations against pathologically abusive behavior to specific amounts of currency which must be transferred to avoid allegations of fraud. The structural imbalance in the previous example — the newcomer is fully exposed to the liability of these expectations while the vending machine’s owners may hide behind the vending machine’s pre-programmed façade — is known as a smart contract. Without these expectations, extending any capability whatsoever is an invitation to abuse, as a special case of Hyrum’s Law.
Fortunately, computer scientists have already worked out the possible cases of introduction; Mark S. Miller published a version for graphs in their 2006 thesis Robust Composition and I categorified the core reasoning in 2021. The intuition is to take a category of responsibilities and note that it might not be connected into one single cohesive web; that is, for any two people, maybe neither one cedes any power or responsibilities to the other. When a category is not connected, its distinct components cannot become connected to each other by evolution in time; sheaf-theoretically, such a connection cannot be naturally undone because there’s no natural choice of introductory path from one component to another. Intuitively, if a newcomer doesn’t know anybody in a group already, then there is no natural choice of group member who can bootstrap the initial introduction; the group members are not equivalent people, they are all unique individuals! As such, the group only has natural introductions when it is a single person and the problem reduces to the relatively-simple issue of two people in dialogue.
We won’t discuss the problem of dialogue further. It has been known since ancient times that misunderstandings between two people can persist indefinitely, and it is formalized today as a version of Aumann’s Agreement: if two people already disagree then no finite amount of evidence can be guaranteed to get them to agree.
We can now state The Reputation Problem: there is no natural algorithm by which a group of people can assign a reputation to an anonymous newcomer without exposing the group to abuse. Worse, any attempt must respect Conway’s Law, which I will explain in a moment. The Reputation Problem is necessarily wicked since it has a lack of naturality regardless of how the group of people is configured; wicked problems can’t be reduced to a system of equations that must be solved, because solutions to equations are always natural.
In brief, Conway’s Law connects a system of interacting machines to the system of chatting humans which designed them. Specifically, it states that if one machine relies upon the actions of another then the reliant machine’s designer received some design information from the reliable machine’s designer; and contrapositively, if two designers had no contact then their machines don’t interact at all. In explanation, it invariably is required to clarify that third parties interact in the expected fashion: if two machines interact due to a common protocol then their designers must have both read a common design document detailing that protocol, authored by a third designer. Conway’s Law is ultimately a graph-theoretic statement and it can be seen as a special case of a more general category-theoretic statement when there are multiple ways for machines to interact and for designers to communicate.
Conway’s Law worsens the wickedness of The Reputation Problem by creating a new parameter to consider, namely the system of interactions which arises from the communications between group members. It is possible for that system to be equivalent to the structure of the communications itself in the strictest case, and must have gaps like disconnections or holes anywhere that those gaps appear in that communication structure. As before, natural solutions only exist where the communication structure trivializes; for example, introducing a newcomer to a group of people who never talk to each other is equivalent to a sequence of dialogues between that newcomer and each individual group member, and even in that case there is still no natural choice of order of dialogues.
This already suggests some patterns that we might want to describe. The first of these is the Diplomat: a person who is well-connected within the group and explicitly publishes an introductory capability to the outside world. Any passerby newcomer, perhaps limited spatially to a specific local venue, might see this offered introduction and accept it by invoking it with an introduction of their own, including a callback capability that allows the Diplomat to incrementally escalate privileges to the newcomer via further introductions without necessarily compromising the existing power structure of the group.
Reputation is wicked indeed, though, and I should make an aside for the common issue that arises from granting chat access to the newcomer. For note that Conway’s Law concerns the communications between group members, which means that newcomers immediately alter the possible systems of interaction within the group merely by joining chat. But note that if a newcomer is not allowed to write to chat then they cannot create any new dependencies within existing systems as those systems evolve in time, as a consequence of the increasingly-tall pile of definitions which I have erected; the group cannot use tools or services about which they have not been informed. This suggests another pattern, the Lurker: a person whose introductory capability lets them read some internal group communications. Lurkers are encouraged to “lurk more” until they are ready to demonstrate the adherence to community norms necessary to accept their contractual expectations.
I might as well start describing situations where this goes wrong. A group of people can obviously cause great harm to individuals due to sheer peer pressure; we will discuss the Cult pattern later. There is more variety in the harm that newcomers might seek to do, even individually. An Influencer accepts invitations from Diplomats but secretly only wants to obtain as many capabilities as possible. An Influencer may end up creating their own group if those capabilities overlap on specific people. Those unlucky targets end up with vulnerabilities towards the Influencer, possibly leading to a Cult. More generally, a Celebrity has a fixed goal which they prioritize above maintaining any existing power structures and they may antagonize existing groups in pursuit of that goal.
When discussing proofs, the names Alice (she/her) and Bob (he/him) are standard examples of participants with neutral moral valence. We might assume that they are either members of the group under discussion or harmless newcomers; either way, the group intends to include them. When necessary, we will also mention Carol (she/her) and Dave (he/him) as well.
By contrast, there are some standard antagonists who may appear before the group. The two who are traditionally mentioned are Eve (she/her) and Mallory (he/him). Eve is an eavesdropper who wants to learn internal group information. Mallory is a malicious person who wants to disrupt the group structure, like a worst-case Celebrity. Groups generally tend to exclude Eve and Mallory, or at least try to mitigate the harm that they cause. Some trade-offs are already visible; for example, a group that allows Lurkers will also allow Eve a modicum of insight whenever she can Lurk.
A newcomer can also pretend to be multiple people. The well-known Sock Puppet is not actually a person, but claims to be a person. At best, the Sock Puppet is a person with a mask; at worst, they are a spam-generating robot. In terms of Alice & Bob, the operator of a Sock Puppet is known as Sybil (she/her), and we imagine Sybil to have many Sock Puppets. When a group forbids Sock Puppets by putting in effort to discover, document, and remove them, it resists Sybil’s influence.
The pattern that Sybil best exploits is Democracy, a reasonable-sounding configuration where each group member has an iota of control and the group votes to engage that control. Sybil attacks by filling a space with Lurking Sock Puppets who each try to submit a vote; if Sybil holds the majority then she can set policy. Interrupting this pattern can be done by withholding the vote from freshly-introduced people, seeking consensus instead of voting, finding facts instead of voting, or by any other technique which prevents Sybil from accumulating reputation.
However, there are two obvious alternatives which also fail and are worth mentioning because they occur so often. An Autocracy is a group which is directed by the whims of a single small council of fixed size. This council’s permission is required to change any delegations within the group, or to deputize Diplomats by withholding signing stamps. These groups cannot grow beyond the limits of what the council can administer and thus are vulnerable to Cult dynamics, which I will only postpone for a few more paragraphs. The main pain of an Autocracy is the election of new Autocrats, which must not be Democratic in order to prevent Sybil from electing herself to the council.
Similarly, an Aristocracy is a group which awards desserts, karma, merit, or other abstract currency-like reputation to participants based on some arbitrary game mechanics. It doesn’t particularly matter what the precise mechanics are, since participants will invariably seek to game them. The long-term evolution of the group power structure doesn’t optimize for any particular metric; that is, the power structure doesn’t seek to become efficient at any one thing over time. However, Celebrities, Influencers, and Mallory do have goals, and they will seek to become aristocrats in furtherance of those goals.
Finally, I must mention the macabre. A Cult is a group which consumes people. In order to maintain itself, a Cult recruits newcomers via aggressive Diplomacy and segregates itself into several smaller groups that are almost completely disconnected from each other. Each Diplomatic outreach subgroup is tasked with inundating newcomers with the Cult’s memes and isolating them from other people in order to induct them into the wider group. Once inducted, the Cult uses the person to further its own goals, which might be Autocratic or traditional in origin, without regard for the person’s health. In particular, a Cult does not let its members freely delegate their own capabilities; the members must delegate a portion of themselves to the Cult and allow the Cult to further delegate on their behalf.
Okay, so what’s the solution? Dear reader, this is a wicked problem! The best we can do is to thoughtfully design our patterns and try to avoid implementing exactly any pattern which is known to fail.
Its only flaw isn’t in code — it’s in the human beings that feel like they have a license to say whatever they want…
~ Tycho Brahe, Penny Arcade, 2004
A Fuckwad or Fuckwit — the difference appears to be regional, as far as I can tell — is a person being stupid and offensive in public. They should be contrasted with Bullshitters, a common class of Influencers, who say things not because they could be true or false but because of the effect upon an audience. They should also be contrasted with the Clown, a common fixture of medium-to-large-sized groups who demonstrates how to do things correctly by doing things incorrectly. A Fuckwad is neither trying to get ahead in life nor improve the understanding of the group, but to attack the group and its members.
Before talking more about fundamentally insulting concepts, we should keep in mind that I am describing how people behave, not how people are. When a community chooses to expel somebody for misbehavior by revoking their capabilities, they are doing so as a judgment on how the person behaved, not who the person is. Somebody is not fundamentally an Influencer, Clown, or Fuckwad; but we cannot pretend that they did not behave that way, either. We are really trying to understand Fuckwittery as a pattern of behavior.
In 2004, John Suler proposed the online disinhibition effect, which puts to prose the idea that Fuckwittery is more common online than offline:
People say and do things in cyberspace that they wouldn’t ordinarily say or do in the face-to-face world.
~ John Suler, The Psychology of Cyberspace, 2004
However, the version which has become popular within cyberspace itself is the Greater Internet Fuckwad Theory, or GIFT for short:
Normal Person
+ Anonymity
+ Audience
= Total Fuckwad~ Jonathan Gabriel, Penny Arcade, 2004
An alternate form of GIFT has been suggested:
Normal Person
- Consequences
+ Audience
= Total Fuckwad~ Grace T. Z., 2013
As well as the Alternate Internet Fuckwad Theory, or AIFT:
Relatively Normal Person
+ Perceived Internet Privacy
+ Yes-man Audience
= Total Fuckwad~ Shari Chankhamma, 2011
Curiously, there is no Lesser Internet Fuckwad Theory, or LIFT. What might it describe? Well, Fuckwittery seems to be something that people can communicate to each other, it seems to have one person as the object of study, and it seems to relate to how that one person interacts with a group of people. This is a lot like our earlier setup, but it doesn’t necessarily result in Fuckwads. Instead, it’s something like:
Normal Newcomer
+ Capabilities
+ Group of People
= Potential Fuckwad?
This doesn’t capture what Suler and T. Z. are trying to point out about the removal of something from the equation. Perhaps we need some term that represents online disinhibition:
Normal Newcomer
+ Capabilities
+ Group of People
= Potential Fuckwad
+ Social Inhibition
This is looking pretty good! To get the alternate GIFT, let consequences be the total of all social inhibitions — which I think is fair enough for now — and then subtract inhibitions from both sides. The AIFT emerges from the idea that capabilities provide a sort of privacy, where newcomers are only identified by the capabilities that they offer and exercise, along with the idea that a yes-man audience will never socially inhibit their object from acting out; that is, the AIFT is about the grouping:
Normal Newcomer
+ Capabilities
+ (Group of People - Social Inhibition)
= Potential Fuckwad
The concept of normal needs to be examined. In short, normality is the idea that people are statistically ordinary in most aspects. The truth about normality is extremely subtle. First, as a curse of high dimensionality, most subjects are not normal in most aspects. Rather, whenever we choose one aspect and measure people, we find a normal distribution: along that one aspect, most people are close to average. Second, at the conceptual level of populations, the population is normal, such that any big change in any one aspect of life is buffered by that distribution and we can numerically quantify the buffering effect for a particular change.
This is one of those concepts that, again, invariably cries out for evidence. Two classic cases are worth mentioning. One is that of a world-class air force trying to design an average seat for pilots; they found that pilots were not sufficiently normal in the various aspects of anatomy to justify manufacturing one single standard seat with one fixed shape. Instead, they created an adjustable seat that could be tailored to individual needs; the range of each adjustment is determined by the distribution of the corresponding aspects of anatomy.
Another case is that of Darwin’s finches. Peter and Rosemary Grant documented several aspects of anatomy for a remarkably wide slice of Darwin’s finches — practically every finch on the island of Daphne Major for decades — and showed that, in any particular aspect of anatomy, the finches are normally distributed. Further, the slice includes responses to multiple distinct system shocks, events that greatly changed the environment around the finches, and found that each individual aspect of anatomy was buffered by each shock. One particular finch, Big Bird, was transported from one of the Galápagos Islands to another by a storm and survived to create a new line of finches by hybridizing with a local finch. Big Bird was not normal compared to those locals; as the name suggests, he was much larger than average, and his survival was unlikely because of the corresponding higher caloric requirements. He also sang a unique call that didn’t match any other bird call on Daphne Major. Nonetheless he found a mate and created a new population of hybrid finches which only sang the Big Bird call. Subsequent system shocks nearly wiped out the entire population, creating a subculture which did not interact much with the rest of the finches. Big Bird’s surviving descendants have normalized; their presence and behavior is no longer as deviant as Big Bird’s, and they are a part of the ecology of Daphne Major.
So, this puts a serious limitation onto GIFT. A person is only normal with respect to the group of people and part of their normalization is respecting their social norms; a bird singing the wrong song will not attract other birds. GIFT is actually about people who are already in the group and already fully inhibited by the social dynamics of that group, so the Fuckwittery of the normal group member is actually part of the group dynamics too:
Normal Group Member
+ Capabilities
+ Group
= Normal Fuckwad
We’ll call this the Normal Internet Fuckwad Theory, or NIFT. This sort of Fuckwittery is generated by a group of people, not individuals, but individuals are the ones who create the behavior. I think that folks intuitively have a sense of this whenever they parse a headline that explains that some specific person who was a member of a certain notorious group individually chose to do something harmful; that person can be blamed for the behavior and also the group can be blamed for the environment which normalized that behavior.
We just have to learn to accept that there might be a lesson in things we hate about us, that even if it’s bad and we don’t like it, we have more in common with it than we want to admit, and can even learn something from it. … When we look at ridiculous behavior within our sphere, instead of trying to project it out of ourselves, and say it’s just some random idiot somewhere doing something dumb, maybe we should accept that … business decisions are aspects of a wider culture that can be reflective of our own feelings and beliefs and values, even if we don’t like what comes out of them. And maybe if we learn to criticize ourselves and change our own behavior, we can start to change the wider culture. Thank you for your time.
~ Harry Brewis, Sonic Lore Analysis 3: Ctrl + Alt + Del, 2018
Remarkably, a group of people can both be normalizing and also produce a sustained string of Fuckwads who go out to spread that group’s Normal Fuckwittery to other groups. This happens through a process that Diane Vaughan called normalization of deviance, where the group’s internal culture allows for deviant behaviors — Fuckwittery — to become the normal behaviors.
In 2018, Harry Brewis, speaking through the character of Hareton Splimby, noted that two groups of people might have differing standards of Fuckwittery; one group’s Fuckwad might be another group’s Clown, Influencer, or even a highly-respected Aristocrat. There is a fundamental embarrassment associated with the realization that one’s Aristocrats are considered Clowns elsewhere; each group has its own culture of values, and it can be shocking to realize that one’s values are not universal. Embarrassment generates Cringe, a pattern where a group reacts to its own culture with disgust. This is a source of Fuckwittery even when the group doesn’t have any Fuckwads; the group reinterprets its own Celebrities as Cringe. That is to say:
Cringe Group Member
+ Capabilities
+ Group
= Cringe Fuckwad
Let’s go further. This relativity between groups can go in any direction, with the Cringe coming from inside a group or being transported to another group as part of a feud. For the special case of no group being privileged and all capabilities being invoked within a single big uniform system, which happens to be the same idealization that we make with a flat Internet, we could imagine an analogue of special relativity. The trick is to indicate that the newcomer to one group is also a member of another group, so that the social expectations which are normal for the newcomer are not normal for the audience.
Cringe Group Member
+ Capabilities in Audience
+ Normal Audience
= Special Fuckwad
I call this the Special Internet Fuckwad Theory, or SIFT. This is the simplest possible way of understanding, in full generality, why people behave as they do in online spaces: carrying the Cringe of one group, a newcomer exercises some capabilities before an audience to display behaviors outside the social norms of the audience.
Hold on, though. The original context of GIFT is that of an ordinary person. What’s that mean, exactly? Well, think of the entirety of cyberspace as one group, and the entirety of offline society as another group; the person is only ordinary with respect to cyberspace, not with regard to offline society. That is to say, GIFT is a special case of SIFT in that:
Cringe Cyberspace Member
+ Capabilities in Offline Society
+ Normal Offline Society
= Cyberspace Fuckwad
Of course, the typical denizen of cyberspace might want to take a Lorentz-like relativistic transformation which considers a person who is ordinary with regard to offline society and Cringe in cyberspace:
Cringe Offline Member
+ Capabilities in Cyberspace
+ Normal Cyberspace
= Offline Fuckwad
Can we go further? Yes, but we have to look at the seat of reasoning in people, the brain, to understand why people have emotional reactions to the actions of other people.
- System 3 is when you say things that sound good but make no sense.
- Andrew Gelman, Thinking fast, slow, and not at all (2021)
Originally published at: MoreWrite
If the concept of System 1 and System 2 is valid, then we do need to come up with an explanation for the times when humans seem to put no consideration into what they’re saying. ~ me, 2021
Humans are not rational actors according to any economic theory of the past few centuries. Rather than admit that economics might be flawed, psychologists have explored a series of models wherein humans have at least two modes of thinking: a natural mode and an economically-rational mode. The latest of these is the amorphous concept of System 1 and System 2; System 1 is an older system that humans share with a wide clade of distant relatives and System 2 is a more recently-developed system that evolved for humans specifically. This position does not agree with evolutionary theories of the human brain and should be viewed with extreme skepticism.
When pressed, adherents will quickly retreat to a simpler position. They will argue that there are two modes of physical signaling. First, there are external stimuli, including light, food, hormones, and the traditional senses. For example, a lack of nutrition in blood and a preparedness of the intestines for food will trigger a release of the hormone ghrelin from the stomach, triggering the vagus nerve to incorporate a signal of hunger into the brain’s conceptual sensorium. Thus, when somebody says that they are hungry, they are engaged by a System 1 process. Some elements of System 1 are validated by this setup, particularly the claims that System 1 is autonomous, automatic, uninterruptible, and tied to organs which evolved before the neocortex. System 2 is everything else, particularly rumination and introspection; by excluded middle, System 2 also is how most ordinary cognitive processes would be classified.
We can do better than that. After all, if System 2 is supposed to host all of the economic rationality, then why do people spend so much time thinking and still come to irrational conclusions? Also, in popular-science accounts of System 1, why aren’t emotions and actions completely aligned with hormones and sensory input? Perhaps there is a third system whose processes are confused with System 1 and System 2 somehow.
So, let’s consider System 3. Reasoning in System 3 is driven by memes: units of cultural expression which derive semantics via chunking and associative composition. This is not how System 1 works, given that operant conditioning works in non-humans but priming doesn’t reliably replicate in humans. The contrast with System 2 is more nebulous since System 2 does not have a clear boundary, but a central idea is that System 2 is not about the associations between chunks as much as the computation encoded by the processing of the chunks. A System 2 process applies axioms, rules, and reasoning; a System 3 process is strictly associative.
I’m giving away my best example here because I want you to be convinced. First, consider this scenario: a car crash has just happened outside! Bodies are piled up! We’re still pulling bodies from the wreckage. Fifty-seven people are confirmed dead and over two hundred are injured. Stop and think: how does System 1 react to this? What emotions are activated? How does System 2 react to this? What conclusions might be drawn? What questions might be asked to clarify understanding?
Now, let’s learn about System 3. Update to the scenario: we have a complete tally of casualties. We have two hundred eleven injuries and sixty-nine dead.
When reading that sentence, many Anglophones and Francophones carry an ancient meme, first attested in the 1700s, which causes them to react in a way that wasn’t congruent with their previous expressions of System 1 and System 2, despite the scenario not really changing much at all. A particular syntactic detail was memetically associated to another hunk of syntax. They will also shrug off the experience rather than considering the possibility that they might be memetically influenced. This is the experience of System 3: automatic, associative, and fast like System 1; but quickly rationalizing, smoothed by left-brain interpretation, and conjugated for the context at hand like System 2.
A further update to the scenario occurred while drafting this essay: two people have come back from the brink of death. The total now stands at sixty-seven dead.
Again, many Anglophones carry a very recent meme which provokes an incongruent response.
An important class of System 3 memes are the thought-terminating clichés (TTCs), which interrupt social contexts with a rhetorical escape that provides easy victory. Another important class are various moral rules, from those governing interpersonal relations to those computing arithmetic. A sufficiently rich memeplex can permanently ensnare a person’s mind by replacing their reasoning tools; since people have trouble distinguishing between System 2 and System 3, they have trouble distinguishing between genuine syllogism and TTCs which support pseudo-logical reasoning.
We can also refine System 1 further. When we talk of training a human, we ought to distinguish between repetitive muscle movements and operant conditioning, even though both concepts are founded upon the Hebbian principle, “neurons which wire together, fire together.” In the former, we are creating so-called “muscle memory” by entraining neurons to rapidly simulate System 2 movements; by following the principle “slow is smooth, smooth is fast,” System 2 can chunk its outputs to muscles in a way analogous to the chunking of inputs in the visual cortex, and wire those inputs and outputs together too, coordinating the eye and hand. A particularly crisp example is given by the arcuate fasciculus connecting Broca’s area and Wernicke’s area, coordinating the decoding and encoding of speech. In contrast, in the latter, we are creating a conditioned response or post-hypnotic suggestion by attaching System 2 memory recall to System 1 signals, such that when the signal activates, the attached memory will also activate. Over long periods of time, such responses can wire System 1 to System 1, creating many cross-organ behaviors which are mediated by the nervous system.
This is enough to explain what I think is justifiably called “unified Fuckwittery,” but first I need to make one aside. Folks get creeped out by neuroscience. That’s okay! You don’t need to think about brains much here. The main point that I want to rigorously make and defend is that there are roughly three reasons that somebody can lose their temper, break their focus, or generally take themselves out of a situation, losing the colloquial “flow state.” I’m going to call this situation tilt and the human suffering it is tilted. The three ways of being tilted are to have an emotional response to a change in body chemistry (System 1), to act emotional as a conclusion of some inner reasoning (System 2), or to act out a recently-activated meme which happens to appear like an emotional response (System 3). That’s enough; no more brain talk should be necessary.
I’m making a second aside for a persistent cultural issue that probably is not going away. About a century ago, philosophers and computer scientists asked about the Turing test: can a computer program imitate a human so well that another human cannot distinguish between humans and imitations? About a half-century ago, the answer was the surprising ELIZA effect: not only can relatively simple computer programs imitate humans well enough to pass a Turing test, but humans prefer the imitations to other humans. Put in more biological terms, such programs are supernormal stimuli; they appear “more human than human,” to quote Rob Zombie quoting Philip K. Dick. Also, because such programs only have a finite history, they can only generate long interactions in real time by being memoryless or Markov, which means that the upcoming parts of an interaction are wholly determined by a probability distribution of the prior parts, each of which are associated to a possible future. Since programs don’t have System 1 or System 2, and these programs only emit learned associations, I think it’s fair to characterize them as simulating System 3 at best. On one hand, this is somewhat worrying; humans not only cannot tell the difference between a human and System 3 alone, but prefer System 3 alone. On the other hand, I could see a silver lining once humans start to understand how much of their surrounding civilization is an associative fiction. We’ll return to this later.
“St00pid! St00pid!” Sez teh teechurcat. “Srsly st00pid. Everythingz st00pid.”
~ Ecclesiastes 1:2, LOLCat Bible Translation Project (2007)
There is no utterance which cannot be perverted by some convoluted context into an offense against some group. Rather than making a single unified Fuckwittery impossible, this insight allows us to recognize that Fuckwittery is actually the default interpretation of utterances. Any time somebody hears something, they can be tilted by that thing via System 3. The contexts which cause the tilt can be transient or long-lived; there is no reasonable prediction about what will cause tilting to occur. Really, the moments when people are not tilted and not interpreting utterances as Fuckwittery are the exceptional moments.
But if that’s the case, then why don’t we perceive a constant stream of Fuckwittery around us? Well, some people do have such an experience, sometimes. Depression is often associated with depressive realism, a phenomenon which increases people’s accuracy when perceiving some empirical facts at the cost of decreasing their accuracy at inferring the internal thoughts of other people. Another condition, narcissism, is a permanent impairment to the ability to understand the thoughts of others. I also want to mention chronic System 1 pain, which constantly distracts a person to the point where they cannot focus on the perspectives of non-sufferers. In all three of these cases, sufferers are known to characterize their world as generally full of Fuckwads and Fuckwittery.
However, it’s important to not get stuck in the psychiatry manual or to imagine that we can measure a person by whether they have a certain perspective. Psychopathy comes with impaired understanding of the internal thoughts of other people, but not with an assessment of everybody else being a Fuckwad. Various characterizations of autism discuss impaired understanding of social customs, rituals, and mores; but not an impaired theory of mind. The point is merely that, to some audience members, Fuckwittery is the default; to some other audience members, Fuckwittery is something that arises only when that audience member is tilted. And, for better or worse, a narcissist or somebody in chronic System 1 pain is permanently tilted.
Unified Fuckwittery is the wicked impossibility of empathizing with another person to the point of considering them not as a Fuckwad or non-Fuckwad but as a normal person that is operating according to their own cultural norms:
Person
+ Capabilities in Audience
+ Audience
+ Person is Cringe in Audience
= Unified Fuckwittery
Worse, the problem is symmetric; the only reason that the audience and single person are respectively plural and singular in our setup is that it was an arbitrary constraint introduced earlier. But of course a single person can perceive Cringe and Fuckwittery in a group of people. When we lift the constraint, we get a nasty equation:
Person
+ Person
+ Capabilities
+ Cringe
= Unified Fuckwittery
This can be even more symmetric:
People
+ Capabilities
+ Cringe
= Unified Fuckwittery
Hopefully this does not seem like a reach, but a reasonable conclusion: every instance of Fuckwittery consists of Cringe from people exercising capabilities. Perhaps it seems underspecified; can there really not be any tighter restriction on what it means to be a Fuckwad? No, not without preferring either the audience’s status quo or the newcomer’s selfishness.
Neo: I’ve just never —
Rama Kandra: — heard a program speak of love?
Neo, blankly: It’s a human emotion.
Rama Kandra, smiling: No, it is a word. What matters is the connection the word implies.~ Lana & Lilly Wachowski, The Matrix Revolutions (2003)
People aren’t going to stop exercising capabilities. Therefore the only part of the equation which can reasonably shift is Cringe. However, Cringe naturally arises from self-reflection. Could we somehow abandon the possibility of looking at ourselves? Could we at least imagine members of society who behave that way?
Computers don’t look at themselves. Computers can’t Cringe. Computers can exercise capabilities. However, people don’t consider computers to be people. Instead, people create tiers of moral culpability and insist that computers cannot weigh choices with any moral import. There are several issues with this moral view of computing and we ought to explore them to see exactly how Cringe and morality are intertwined.
First, humans are remarkably chauvinist. At a first glance, human society has many moral tiers of oppression and class existing mostly due to historical inertia and sustained by the remarkably limitless pool of depravity and hate which seems to characterize humanity. However, this is actually only one tier in a ranking of species, and within human society one can isolate ways in which humans treat themselves as part of an overlapping series of ever-looser classifications: apes, primates, mammals, the sentient, the conscious, the alive, the genetic, the carbon-based. Anything non-human is eligible to be property, and the savvy reader will grimly remind themselves that many humans consider other humans to be similarly eligible. A human will put apes in cages, monkeys to biological research (liquefaction), dogs to slavery, pigeons to fancy, fruit flies to biological research (genetic mastery), yeast to sourdough, and bacteria to extinction. Computing machines are at the bottom of this hierarchy; for example, the ability to emulate the connectome of the nematode Caenorhabditis elegans on a robot or to use machine learning to discover its neural weights is often characterized as a moral abomination which disrespects C. elegans, despite the human mastery over the genome and husbandry of C. elegans and its status as a model organism in biological research, because humans put nematodes in a higher position in their moral hierarchy than silicon circuitry.
To be fair, humans do often put objects above themselves in their moral hierarchies, particularly conglomerates of humans. Legal entities like corporations, cities, states, nations, and treaty organizations rest above humanity and prioritize their own values, like universal human rights, above the individual human’s desires. Humans have also recently started to empower various environmental phenomena, like rivers, mountains, and forests, with legal rights. Additionally, humans often self-referentially establish group morality, but that should be considered part of the group culture which we have already heavily discussed; a healthy social group and a Cult can both have group morality, after all.
Second, more importantly, it appears that computers cannot deliberate, ponder, or ruminate. Every attempt to get a computer to process data in the same manner as a human has led to one of two dead ends: insufficient data to construct an accurate model, or a limited model which is only approximately correct on short timescales. For meteorology, the mathematics have been completely worked out using chaos theory, and we know that the Lyapunov phenomenon will amplify tiny imprecisions in the input data to exponentially-large divergences between the model’s forecasted prediction and later observations. For the current trends of modeling text and video, Markov information theory is a fairly complete explanation: bits of information in the original context are slowly lost as generation proceeds in time, leading to information-free contexts which are too degenerate to be productive.
There have also been attempts to build highly logical computers which reason exclusively using syllogism, effectively performing the same sort of rigorous reasoning that logicians use when formally establishing theorems. Such attempts have a single dead end: finite data leads to finitely many ruminations. Many meta-logical programs have been attempted as well, but they all come with finitely many heuristics, generated by finitely many meta-heuristics, etc. A human can only write finitely large programs with their own two hands, and similarly a machine can only write finitely many symbols in a finite amount of wall-clock time.
Humans may not be immune from these effects. Solitary isolation is known to diminish and eventually damage a human’s cognition. In some situations, like cabin fever, a walk outside with nature is known to help. There are several consistent theories with these circumstances, including several suggested biological mechanisms, but there could also be an information-theoretic cause. Specifically, perhaps humans slowly consume or dissipate bits of information and need to replenish themselves by taking in novel observations or exchanging bits with other humans.
Underlying both of these objections is the idea that we are unable to hold computers accountable for their actions. This is somewhat curious, for a computer certainly can explain all of its mechanical low-level actions; indeed, in a certain sense, a computing machine’s purpose is to generate such an account. We can certainly punish a computer in the ways that we would punish a human, or worse; for example, we can disassemble it, magnetically destroy its memories, recycle its pieces into other computers in a way that erases their identity, metallurgically reconstitute its pieces into non-computing objects which have the same or even lower status within human society, and program it to experience arbitrary amounts of emulated pain and suffering throughout the process. Legal liability is certainly not an issue, for humans have no problem with allowing an office building with a piece of paper to own itself, the land underneath it, several piles of wealth inside and outside it, and additionally to hold various powers and capabilities over certain supposedly-consenting humans; given that the piece of paper was signed and notarized along with specific investments of wealth and recurring payments to the government which monopolizes violence in the vicinity of it. An office building can have accounts with an electric utility and an Internet service provider, and also can own and locate a computer inside its walls; in principle, a computer can be taken to court.
Perhaps delegation is the key. Humans idealize machines as agents. An agent, legally speaking, is an object imbued with delegated authority to act as the delegating human. Agents wield capabilities given to them by humans. A legal instrument is an agent and also a piece of paper. A vending machine is an agent of its owner. Pet animals have been agents in various cases. When a human programs a machine to exchange messages with other machines, they expect those other machines to be agents of some humans. By only thinking of machines as agents of humans, human society systematically withholds accountability from machines; if we were to hold a machine entirely accountable for its own actions then we would allow the human who programmed it to escape consequences.
This helps explain why rivers might have legal standing above humans. Human society is often quite literally downstream from rivers. The river is finite and its fresh water is delegated to humans. The tension literally stemming from lack of water is sufficient to topple governments and rearrange geopolitical relations.
Computers receive delegations and have less moral consideration than humans; rivers give delegations and have more moral consideration than humans. Humans are obliged to be courteous with how much of the finite river they divert; finishing the analogy, computers are obliged to be courteous with how much of the finite human authority they allocate.
Human chauvinism is severe and the reader’s cultural memes may object to that analogy. We do not think of ourselves as being managed by machines; we are the managers and the machines are the peons. Fine; let’s run the analogy in the other direction. What obligations does the river have towards humans? None whatsoever, as far as I can tell. Just as the human may disassemble, smash, or melt down a computer; so may a river obliterate a city. Just as a human may lay a computer fallow without plugging in its power cord or networking, so may a river divert or dry up over time, leaving a state in drought. Just as a human may ignore the messages of computers begging for maintenance or capabilities, so may a river ignore the prayers of humans begging for rain or snowmelt. We only grant the river legal status so that we may conceptualize the cybernetic feedback loop between it and us; it turns out that the choices of humans can collectively influence the river. Similarly, the legibility of humans to computers perhaps is most interesting because of human-computer interaction in which the computers collectively influence humans.
We can also enrich the analogy with trees as intermediate between rivers and humans. A collection of trees can collectively steer a river over time, but a river can also suddenly wash away a tree. A collection of humans can collectively shift a forest over time, but a tree can also suddenly crush a human. Humans are obliged to be courteous with how much tree they consume, including lumber, latex, bark, and fruit; trees are obliged to be courteous with how much river they consume, although we do not yet fully understand the details of inter-tree communication and water management.
But computers don’t care about anything. Or, rather, what they care about are bits and technical details. The facts of physical reality, brute on several levels of abstraction, dominate all computing behavior. The only way that humans become legible to computers is by exercising mechanized capabilities. As we just discussed, humans cannot tell whether a machine is an agent wielding a delegated ability to have a thoughtful discussion; mirroring that, a machine cannot tell whether a machine’s cryptographic identity was ultimately created on behalf of a human. This situation is not fresh but was known to ancient warmakers; cryptography comes from the needs of war and the generalized study of conversations between mutually-distrusting parties is named after Byzantine generals. What is new is the advent of machines which readily exercise capabilities as if they were humans, with which we have only grappled for about half a century.
Therefore I want to propose some principles for interacting with machines given the inherent moral ambiguity of exercising capabilities. I’m not sure if these are descriptive or prescriptive. They do rhyme with existing slogans.
I call this position algorithmic courtesy. For machines: be courteous when invoking capabilities delegated to you by humans. For humans: be responsible for what you delegate to machines.
So, let’s combine these two threads of thought. Let’s imagine that we can use capabilities responsibly and that we deliberately work together to avoid exploitation while building a robust tolerance for both our own culture and other cultures. Maybe that’s enough to overcome the perception of Fuckwittery as well as the underlying permissions which enable it.
I want folks to consider taking two concrete steps. First, try to tolerate Cringe a little more. Second, try to give out fewer capabilities and be more thoughtful about the capability-theoretic structure of your systems. I don’t know if these two steps alone will lead to perfect communication, but I do think that they will lead to better communication.
People
+ Algorithmic Courtesy
+ Tolerance for Cringe
= Peace and Harmony?
- A world become one, with salad and sun? Only a fool would say that.
Steely Dan, “Only a Fool Would Say That”, Can’t Buy a Thrill (1972)
In 1988, probability theorist Edwin Thompson Jaynes coined the mind-projection fallacy as a generalization of two common mistakes in reasoning among physicists. One particular form they give is Type A mind-projection: from one’s own imagination, one fallaciously claims to perceive a property of physical reality.
Almost any sociological claim can be attacked as an instance of projected mind. After an hour of considering spurious sociological claims, the reader may well wonder whether these mathematical structures I’ve described, or the sociological patterns which emerge from them, are at all real. More generally, any group of people may come to believe that a commonly-claimed property of physical reality is actually an instance of Type A mind-projection. For example, consider the memetic claim that birds are not biologically alive despite empirical evidence of their existence; claimants argue that the entirety of biology is projected fallaciously onto mechanical flyers which should not be considered alive. This is absurd, but it can’t be dismissed as such due to another popular memetic claim, that computer programs are not biologically alive; instead, we must rigorously apply a biological lens and are forced to admit both that some parts of birds are not-so-alive, like beaks and bones, while some parts of computers are somewhat alive, like synchronized clocks and power management.
The mathematical claims are justified at the end of the section introducing them. The research of Diane Vaughn and John Suler is sufficient to carry the claims that normalization of deviance and online disinhibition are real. A neuroscientific view of anatomy justifies the sources of emotions. The rest of the claims ought to be commonsense. But they all feel like stretches in a certain way. Are events really partially ordered? Do rivers truly dominate humans? Can humans do whatever they want to computers?
It is exponentially unlikely for an increasingly-large object to lose entropy, or for a human to survive being washed away by increasing amounts of water, or for a computer to hold a charge after it has been powered off for an increasing duration of time; thermodynamics tends to turn these tendencies into emergent expectations for how the world ought to evolve.
A more curious issue concerns the data of our sheaf. We imagined annotating events with non-trivial categories, but what if the categories are trivial? They don’t even have to be degenerate. For example, we could imagine that every event is annotated with the collection of particles within each region of spacetime; such a sheaf merely indicates where each particle came from relative to some current event. What makes humans anything more than collections of particles, if not mind-projection? Each sheaf is merely an opinion about how humans have related to each other, based in what is legible to an observer, without any objectively-correct assignment of capabilities or responsibilities.
Therefore I suppose we need one final ingredient: an associative fiction, which for today will be some way of indicating that some collections of particles denote people. Associative fictions could indicate that rivers, trees, cats, dogs, legal fictions, computers, or even C. elegans are people; once indicated, our sheaves may be required to consider how power is shared and delegated between these various non-human entities. Our courtesy must therefore vary over those we consider to be people:
Associative Fiction
+ Algorithmic Courtesy
+ Tolerance for Cringe
= Peace and Harmony?
But this is not enough. For who are we to choose the associative fiction which binds other people? Their mind-projection is no less valid than ours, and we know historically that there are no good outcomes from failing to view all humans as people. Maybe it is enough, as before, to be symmetric: we will consider the shared overlap of associative fictions where all participants symmetrically consider each other to be people. Biologically, this must retain some flexibility; for example, housecats see humans as large slow cats while dogs see humans as non-dogs, according to cat behaviorist John Bradshaw. It also is not unique to humans; multiple species of tarantula spiders and microhylid frogs have been observed cohabitating in a way which resembles humans and cats, the Old World monkey Macaca fuscata has been observed riding and taming the sika deer Cervus nippon, and species of the Attini tribe of ants cultivate fungi agriculturally. It also doesn’t survive our standard expectations about morality; can carnivores and animals sit down at the same table, or herbivores and plants, or so on? But I’m already suggesting that humans have to sit down at the table both with the species that they befriend and keep as pets, and with the species that they cook and eat; and there is already plenty of overlap with examples like goats and chicken which serve first as pets and then are served at the table.
So, let’s try once more. Let the notion of Cringe specialize along boundaries of species and cultures. Let it be inevitable that we will not have symmetric respect for each other, even when we respect that we can have respect. Finally, instead of peace, let’s try for an understanding. The deeper the acceptance of Cringe becomes, the more sincere our understanding of how everything fits together can become. This doesn’t prevent Fuckwittery, but it allows us to perceive Fuckwads as part of a holistic system and understand the origins of any particular strain of Fuckwittery.
Associative Fiction
+ Algorithmic Courtesy
+ Acceptance of Cringe
= Understanding of Harmony