Notes on the NBA finals
Steph, having caught the gaze of a teammate, raises his hand in the air to say “my bad”—attribution of blame, like saying “I’ll pick up this check.” The structure of gesture accompanies taking credit.
A well-balanced basketball team is a system of complementary heuristics—two guards, two forwards, a center.
Teams are groups of individuals whose fortunes are aligned.
What is a probability here? How often a “type” of event occurs, in the wake of another type of event, given a specific “type” of context.
A player goes up for a layup, and is blocked. He and his defender fall onto the court. Both are at the end of the court, and the rest of the two teams’ players have begun running back to the opposite side. The defender gets up first and jumps over the fallen offensive player. The fallen offensive player puts his hands up above his head as the defender jumps over, and the defender falls to the floor again. A technical foul is called for endangering the defender in possible injury. A coach uses a challenge, and the referees review footage of the play. During replay, the commentators cannot decide where the intent was. “I can see the argument where he was just protecting his head from a guy jumping over him.” “And from where I stand I can say, hey, you tripped me.” How could there be a right answer here that is not in the realm of ethics? In the realm of ought? How could there be an ontological “fact” about where the responsibility lies?
In ball sports and video games, we say someone made a “read” of the situation. In that vein, it seems true to say that the other players actions have “written” to the reader, whether they “intended” or not.
“Intention” is an abstraction over determining whether the action “is” or “isn’t” a reliable part of the organism’s pattern, in the way that the action is taken as, to anticipate its future occurrence. In this sense, an “accident,” a logistical or implementation mistake, byproduct effects, and cues are all of the same class. What is this class except about (1) regularity and (2) assigning blame and regulating and managing future interaction?
In other words, if I tell you everything that happened externally, what is missing?
There is a divide in officiating culture between “by the book” and “considering the full context of the game and the penalties and the consequences.” If a player has a technical, and a second small technical-worthy even happens, is it worth ejecting him, even though the rule is two technical worthy fouls lead to an ejection? And should the situation be different if it is a playoff game? If it is a deciding game in a playoff? These rules (objects, categories, heuristics) are constructions of the referees, and there isn’t a right answer about how they ought to enforce it, there are only the effects that the letter and the letter’s enforcement have on player behavior, whether it’s incentivizing or disincentivizing, whether it’s providing a predictable or unpredictable structure of reward, what is trying to be maximized in the game and equilibria, etc.
“Can’t decide if Warriors dominant or if Celtics trash”: what would this distinction mean except with some comparison to a notion of average replacement, or expected performance? The question of blame: like asking whether a failure to read a prescription is the result of sloppy handwriting by the doc, or poor deciphering by the pharmacist. And the “average replacement” of basketball concept is similar to a standard of “reasonable personhood” in U.S. courts.
We have a sense that shapes the feeling that filters what is said and what is not said, and the “feeling” that causes aversion or attraction, that leads an utterance to be selected or censored, underlies the subconscious strategizing.
basketball sports competition communication strategic interaction typification
Seedsmen against eugenics (a draft)
“Has anyone got good literature on species/cultural diversity as terminal or instrumental ethical value as in e.g. climate/environmental discourse?” asked Spendy.
“Oh! My man Luther Burbank does!”
Burbank (to whom we owe the Russet Burbank potato, the free-stone peach, the shasta daisy, the paradox walnut, the thornless prickly pear, nearly all of my favorite plum and rhubarb varieties, and hundreds of other great plants) is my favorite plant breeder. He worked from the 1870s to 1920s tirelessly and fanatically creating hybrids and crosses that enabled the California prune industry, countless orchard and garden favorites, and the modern french fry.
To the embarrassment those who espouse progressive ideals in 2022, eugenics was generally considered to be a progressive ideal in 1922. You’ll notice Burbank’s active career spans a time when eugenics societies were proliferating, and researchers were committing unconscionable crimes against usually unwitting subjects.
Though he’s only remembered by agronomy nerds today, Luther Burbank was a household name at the height of his productivity. In order to draw attendees to a eugenics symposium*, a society invited Lute as a keynote speaker. It would go without saying that the world-renowned breeder, who ruthlessly selects only the best parent lines for his hybrids, would support eugenics, right? So they didn’t bother to check.
Burbank in fact despised eugenics with the passion that only a plant breeder—champion of diversity—could. He accepted the invitation to keynote, and eviscerated the eugenicists: calling eugenics wrong-headed, the opposite of good sense, telling them they should go home and change careers—or at least practice on plants (implying that they would immediately see the outcome of their folly) before committing any irreversible procedures upon humans.
*I’d intended (and do intend) to do a full writeup of Burbank’s excoriation of the eugenics society for my main blog—but I temporarily can’t find my copy of Jane Smith’s lovingly researched Burbank intro Garden of Invention , and I’d like to find that and use its index to orient myself in the original sources. But I didn’t want to withhold the hammer when the iron was hot, so I would welcome any feedback, questions, or things you’d like to see in a full-scale treatment of this anecdote, or the larger question. Til then, I apologize that I’ve forgotten the name of the symposium where this occurred.
plants strategic interaction science alignment eugenics Luther Burbank diversity
Downvoting into the void
If you downvote a tweet, but that isn’t shared with anyone, is it still manipulation?
I was surprised to discover that Twitter recently introduced a new feature to a Select (probably random) group of users: downvoting. Rather than just having buttons for replying, retweeting, liking and sharing, there is now a fifth icon: the universal symbol for “this sucks”.
Upon trying out my newfound power for the first time, I was greeted with an almost-but-not-quite helpful message:
Although I appreciate Twitter trying to be transparent about the effects of user actions, (far too many platforms make it way too hard to figure out how their systems work, even for simple things like online status indicators), there is a strange ambiguity in this informational message.
It’s certainly useful to know that downvotes are private, and not public, which presumably means that no standard Twitter user can look up what tweets I have downvoted. That is useful information (if it is true). But the suggestion that my vote won’t be shared with the author or anyone else? Knowing that tweet authors won’t see downvotes is similarly useful, but “anyone else” seems surprisingly expansive.
I guess the question is, who “owns” my downvote, such that they might be in a position to share it? (If ownership is in fact required for sharing?) Clearly I do, to some extent. If I wanted to, I could tell you about all the great and trashy tweets I have downvoted, though you would unfortunately have to take my word for it. I could even comment directly on a tweet, to let the author know that I have expressed my displeasure in this way.
But when Twitter says that my downvotes won’t be shared with anyone, does that include Twitter employees? Or do they already claim ownership of that information, such that sharing is not required for use? Or perhaps they will only be processed algorithmically, with no human inspection of the individual results?
In either case, two things seem to be almost certainly true:
Twitter will use the downvotes in some ways, especially algorithmically, such that downvoting is still a kind of manipulation, even if it’s one whose consequences I cannot predict, and which may never be seen directly.
Even if #1 were not true, there would still be effects from Twitter having introduced this feature. In downvoting a tweet, I manipulate myself. In some small way, I commit myself to a negative judgment of a subpar tweet, even if it’s not a public commitment, nor something that the author of the original tweet will ever know about.
A more honest way to argue for meanings of words
by Hazard
As the Chicken and the Reason often say, “The meaning of words and language is whatever you can get people to agree to.” I think that’s importantly and fundamentally true. I also see some cucks out there who use such logic to claim that no-one can resist them trying to make words mean whatever they want in some discursive community.
The technical and specific jargon that CS people have developed to talk about data-structures is an incredibly useful resource. It’s an interpretive equilibrium that has taken a lot of work to arrive at which allows for a lot of dense compressed communication of very specific things in very useful ways to people working in the field. This shared vocab is a common good.
Sometimes I see people defending some use of a word, phrase, or bit-of-language, and it’s clear to me that they’re defending it because they see the current interpretive equilibrium as very useful. But sometimes this defense expresses itself as “dude, that’s just what that word means!”
My guess is that the appeal of this rhetorical strategy involves a dash of philosophical simplicity, and more importantly it offers a route to offload one’s own personal desire to the Objective. It’s not that I and my peers care about this particular set of meanings, that’s just the way it is, there’s nothing to argue with dude! It’s reality! Get a dictionary bro.
There’s a frame a lot of people live in where personal desire, or appealing to the desire of specific people, is rhetorically invalid, and everything must be framed in terms of some “higher Objective rightness” for action to be taken. I vaguely recall an anecdote of a university student who couldn’t get their personal complaints about a bad situation taken seriously until they re-framed it away from “I don’t like this and this is harming me” to “As a women, this is a problem.” The university wasn’t capable of taking an actual person seriously, but they could take A Women TM seriously, because now responsibility for the desire has been shifted to a legibilized Objective form that the university has more advanced rhetorical structures to legitimate.
Often when interacting with bureaucracies and institutions, you won’t be able to dismantle their power in any short timescale, so sure, do whatever you gotta do to get the system to work for you. But don’t forget that you can also have conversations and arguments with people that aren’t entirely enmeshed in such structures, and don’t forgot that the honest and grounded-in-what-you-personally-care-about convos you have with your peers have a non-trivial effect on shaping the next generation of institutions.
The fact that meaning is “what you can get away with” does not mean that everyone must let you get away with bullshit. It’s possible to say, “No, fuck you, this word already has some structured specific meaning that my discursive-community has coordinated around, I and others get a lot of use out of it, and I am going to resist you trying to shift the interpretive equilibrium to one I think is worse.” It’s possible to hold that front without falling back to the philosophically sloppy stance of “you can’t use the word differently because that’s not what it means.”
Kettle logic
by Neil
There’s a poem I see periodically on social media, usually called the Narcissist’s Prayer:
That didn’t happen.
And if it did, it wasn’t that bad.
And if it was, that’s not a big deal.
And if it is, that’s not my fault.
And if it was, I didn’t mean it.
And if I did…
You deserved it.
You may find it interesting to know that there’s a term for the kind of argumentation depicted in the poem, where multiple inconsistent lines of reasoning are deployed in parallel. Derrida called it “kettle logic,” after a story related in Freud’s Interpretation of Dreams:
The whole plea […] reminded one vividly of the defence put forward by the man who was charged by one of his neighbours with having given him back a borrowed kettle in a damaged condition. The defendant asserted first, that he had given it back undamaged; secondly, that the kettle had a hole in it when he borrowed it; and thirdly, that he had never borrowed a kettle from his neighbour at all.
This mode of reasoning is taken by Freud to be a characteristic of thought in the dream state; but I think it can be taken more broadly as a strand of self-deception.
Progress studies, pt 1
From a Discord near you:
progress studies seems kinda fake. likely to ignore/not engage with existing scholarship if I had to guess. I guess I just mean that it’s kinda weird to me to claim you’ve elucidated a new field of study when it’s about questions of human flourishing and governance and political economy which people have been talking about for thousands od years.
This kind of critique has been leveled against John Nerst’s “erisology” as well, so let’s get this out of the way: all fields are fake, and all fields ignore pre-existing literature and literature from outside fields. To what extent has NLP properly reconciled with the millennia of discourse about linguistic meaning? How much overlap is there between sociology, anthropology, ethnomethodology, social psychology, and economics? Let a hundred fields bloom.
Now for progress studies. I think for a while the reigning cultural mood, especially among the intellectual classes, has been pessimistic and nostalgic, both culturally and economically. Moreover, this techno-pessisimism and criticality towards progress’s fruits have been mainstays of arts and humanities for a while now, rivals as they are with engineers and scientists. (And also, being political rivals with capitalism, which defends itself on the basis of material and technological progress. To cede the value of material and technological progress is to provide ammunition for the enemy.)
Progress studies seens at least partly to be aiming for a political and cultural intervention on intellectual and popular mood. The other half of the project is taking the stagnation thesis, the 1972 thesis, the cultural nostalgia diagnoses seriously and trying to figure what went wrong. When you have David Graeber and Peter Thiel—ostensibly as politically opposite one another as possible—coming together in discussions and finding remarkable common ground in their social diagnosis, you should probably pay attention and take that diagnosis seriously, even if it proves to be wrong.
I don’t think any existing field is working on specifically the kinds of problems they are interested in, and just because there are vague fields like economics or history of technology fields that sometimes talk about these problems doesn’t mean it’s a priori silly to try to get a group of people together who focus on these problems for pragmatic reasons. There also aren’t humanities fields whose mood and worldview includes techno-optimism, and that shift alone unlocks possibilities for new theoretical discoveries.
Again, fields are constructs of academic institutions, and the name on the tin isn’t actually what those disciplines study. (7+ fields study language with different emphases and interests and models, and only one of them calls itself linguistics.)
You can look at something like ethnomethodology or strategic interactionism and say they’re just reinventing sociology but sociology never looked at social interaction the way that, or with the frameworks that, EM and SI do.
I definitely agree progress studies has its fair share of LARPing and cringe. But I think any new scene with low barriers for entry and identification will. And also that any strategies or efforts to open up new questions or approaches or outlooks, under a banner, that has energy behind it, is a good thing to be encouraged with academic humanities in the gridlocked position they’re in.
I see Progress Studies as looking at existing media studies, digital humanities, social theory fields in academia and saying “fuck that, there’s zero place for how I see technology, I’m gonna sew my own banner and rally people who share my mood because they’re the people I wanna talk to.”
I agree that I would love for them to actually directly rebut the technopessimism, or deal with some of the foundational critiques of the limits of material improvements, rather than cruise off vibes—but in my experience, it’s very very rare that intellectual movements do this—transcend mood-resonance and truly engage with theoretical opponents.
In part because to pull off a new paradigm you have to more or less reinvent your language, and start using a totally different frame for valuing the world, and it’s hard to both do that and also engage with/translate to the existing paradigm.
You’re in the process of figuring out the reasons why your (scorned/ignored) mood is worth defending, and that is a tough ongoing process of verbalization and conceptual refactoring. Turning around and offering a systematic rebuttal is tough and this is true even if your opponents are flat earthers—the existing paradigm had a long time to build an enormous set of plausible sounding arguments that all interconnect, and traversing that graph takes a lot of time. Usually never gets done systematically until long after the dust has settled
progress studies fields boundaries progress knowledge logistics institution-building NLP erisology mood affiliation field-founding John Nerst
The Chomskyite and Wittgenstenian
“Chomskyite” or “Wittgenstenian” become signifiers in a similar way that “Lynchian” or “Kafkaesque” become stand-ins: for ways of seeing, for ontologies of the world which are less a set of explicit principles or category declarations and more like ways of paying attention in the world, of filtering the world; more like a set of examples which seem especially pertinent, of parts which are taken as especially emblematic or metonymically meaningful, of moods and modes, approaches and problem-solving instincts. They become figureheads for individuals’ positions and points-of-view in the discourse, so that Lynch stands for a kind of American filmmaking that can be opposed to other schools of American avant filmmaking, and Kafka for a kind of literary modernism. And they are exaggerated or misrecognized in exaggerated directions, either as strawmen for opponents, or because credulous fans have simplified and distilled and appropriated and over-extended and generalized and projected-upon their views with views of their own.
Ambiguous nonconformity
In a recent post, Suspended Reason laid out a set of ideas related to the function of communication and the game-like nature of human interactions. To summarize briefly, statistical regularities connecting signals with results provide a foundation upon which a legible coordination can be built, which can be to the benefit of cooperative players.1
To quote Suspended, “From these premises, I believe we can assert that public life will be dominated by an inclination toward–nay, necessity of–conformity, which is to say, the self-alignment between communicative acts selected by an agent, and the ‘reputation’ or signification so to speak of the act itself.” As an example, he points to the fact that others will predictably interpret certain facial expressions as anger, which means that you must avoid making such expressions, unless you want, or are willing, to be perceived as being angry.
There is an important corollary to this, which the post hints at, but does not fully develop, which is that there will be greater risk but also greater potential reward for deviating from a more well established norm. If your community has made it very clear that some particular behavior is unacceptable, then you know, and can fully expect, that there may be severe consequences for you if you do it. At the same time, however, the fact there is something in (many? all?) of us that abhors conformity, means that there will always be some subset of people who are seeking to differentiate themselves from the mainstream, and will be in search of a Schelling point around which to coordinate their dissensus.
Even more interesting and important than this, however, is the ambiguity inherent in nonconformity.2 When people recognize that their norms are parochial, they are likely to be very forgiving of violations. (“Oh, you’re sitting in my Dad’s chair, but of course you didn’t know that!”) As norms become more common, there is a greater expectation that everyone should be aware of them, and easier to read nonconformity as rejection, especially once an obvious subculture has taken shape (“The kids these days!”).
But once a norm becomes so widespread that violation would seem to be unimaginable to its adherents, it becomes harder and harder to make sense of someone’s deviation from it. And because the expectations of rational action and self-preservation are so widespread, any attempt to compel an explanation from a heretic only provides a more powerful game in which the iconoclast can confound the conformist. When Bartleby prefers not to, there is essentially nothing that his boss can do to change or even make sense of his purported subordinate’s behavior.
The difficulty of interpretation comes not just from an imbalance in the cost-benefit analysis (which, in extreme cases, might seem to some to be impossible to work out in a way that would be to the benefit of the dissenter), but from our reluctant recognition, at the core of being, that something about our inclinations remains incalculable. Consider Dostoevsky’s underground man:
“Oh tell me, who first announced, who was the first to proclaim that man does dirty only because he doesn’t know his real interests, and that were he to be enlightened, were his eyes to be opened to his real, normal interests, man would immediately stop doing dirty, would immediately become good and noble, because, being enlightened and understanding his real profit, he would see his real profit precisely in the good, and it’s common knowledge that no man can act knowling against his own profit, consequently, out of necessity, so to speak, he would start doing good? Oh, the babe! oh, the pure, innocent child! and when was it, to begin with, in all these thousands of years, that man acted solely for his own profit? What is to be done with the millions of facts testifying to how people knowingly, that is, fully understanding their real profit, would put it in second place and throw themselves onto another path, a risk, a perchance, not compelled by anyone or anything, but precisely as if they simply did not want the designated path, and stubbornly, willfully pushed off onto another one, difficult, absurd, searching for it all but in the dark. So, then, this stubbornness and willfulness were really more agreeable to them than any profit . . . Profit! What is profit? And will you take it upon yourself to define with perfect exactitude precisely what man’s profit consists in? And what if it so happens that on occasion man’s profit not only may but precisely must consist in sometimes wishing what is bad for himself, and not what is profitable? And if so, if there can be such a case, then the whole rule goes up in smoke.”3
Or more simply, remember the poet’s words “Do I contradict myself? Very well then I contradict myself”.4
I do think Suspended is on the right track, but I also think there is something like a singularity at the bottom of a chain of reasoning by which there is both calculable and incalculable value in refusing to conform.
It’s not a long post; I recommend reading it in full!↩︎
I would like to find a better word than “nonconformity”, but terms like “recalcitrance” and “rejection” imply too much of a definite attitude toward what is being deviated from, the lack of which is exactly the point I’m trying to make.↩︎
Fyodor Dostoevsky, Notes from Underground, First Vintage Classics Edition, translated by Richard Pevear and Larissa Volokhonsky (1993), chapter VII, pp. 20–21.↩︎
Walt Whitman, Song of Myself (1855).↩︎
games communication cooperation conformity strategic interaction generalized reading legibility ambiguity literature Herman Melville Fyodor Dostoevsky Walt Whitman
Salvaging Austin
by RIPDCB
Does Austin need salvaging? He very well might not.
I originally came to Austin’s work provisionally, finding myself on the doorstep of Philosophy of Language by way of the Language Poets. Back then, I found the LangPhil stuff interesting but I wasn’t sure what to do with it. I thought to apply it as a lens through which to read the aesthetics of the very poets who led me to Austin, but that ultimately felt more performative than insightful. And, with the back-beaten beast of formal logic lurking behind much of what I was reading at the time (Kripke & Grice, along with Austin), the attempts I made at cohering these texts into a workable thesis on the mechanics of language seemed, to me at least, to be naive and underinformed.
After joining TIS, I found myself once more at Austin’s doorstep, again provisionally, leafing through my copy of his Philosophical Papers on a sunny bus ride. Revisiting Austin in this new light, the prominent place he affords to formal logic in his pragmatic invectives against farty philosophizing no longer struck me as a barrier to comprehension. Instead, it became an obstacle to ram through, a weight to toss off. When he’s at his most lucid, Austin makes one insightful point after another about the linguistic and philosophical inefficiencies holding us back from coherently sketching out the contours of ordinary language and its usage. In reading him again, I’ve realized that he does this despite always wending his path, scorched earth-style, through the jungles of formal logic.
So, if Austin does need salvaging, it’s from the heavy garters of his technical language. The point of this piece (and perhaps others that will follow) will be to cut through the Logic straight to what might be useful in his writing, if even as a single thread, in developing a pragmatic framework for meaning & language.
—
What better place to start than with his essay “Meaning of a Word”. It’s a shining exemplar of all that is so fun and so frustrating about Austin’s writing: its seductively simple title & its plain-spoken delivery; its no-bullshit perspective & close proximity to getting at the root of the problem (in this case, meaning and ‘what we talk about when we talk about meaning’); but also its maddening ability to take the complicated for granted and beat the points that feel the best explored so far into the ground you question whether you ever understood what the essay was after.
In “Meaning of a Word”, Austin uncovers and then enumerates the ambiguities that separate meaningful usages of the phrase ‘the meaning of a word’ from nonsensical ones. The distinction lies between the local and the universal, the contextual and the general; it’s the difference between asking, “What is the meaning of the word brat?”, and, “What is the meaning of the word brat?” In the former, you’re after a definition that is contextually specific; in the latter, an abstract concept that is both 1) sufficiently discrete enough to be properly definable, and 2) sufficiently general and generalizable to the extent that it can transcend any particular usage. Austin’s conclusion is that the former usage is meaningful because you are inquiring about something while the latter is nonsense because you are inquiring about nothing in particular.
If I could double italicize the ‘nothing’ in ‘nothing in particular’ I would, because the arch-argument across the three parts of this essay is that Austin doesn’t believe there to be a discrete entity knowable as the ‘meaning’ of a word that can be excised from its usage. As much as we might ask after a transcendental signified, it doesn’t mean that there is one. But the fact that we can ask after a transcendental signified might be an integral element of how and why we can get so mixed up in searching for or conceptualizing linguistic meaning as anything other than contextually specific and use-defined. Our language seemingly allows us to move from something-in-particular definable in a given context to nothing-in-particular, such that the nothing-in-particular is presupposed to be something-in-particular in a general context because you can phrase it as something-in-particular. Of course, Austin barely spends a page exploring this point, but it’s to me one of the more insightful ones in the essay. By characterizing the search for transcendental meaning as an example of gratuitous philosophical generalizing, Austin has countered the philosophical claim that there is a unique thing–or unique class of things–specific to every word in linguistic terms, allowing us to, moving forward, comfortably situation the question of meaning in a pragmatic, usage-forward context. Which is helpful, because the question of whether or not there is such a thing as transcendental meaning begs to be addressed in similarly metaphysically-oriented terms, which often leads to only more confusion as ambiguity and linguistic sleights of hand build, one after another, upon each other.
One last point: Austin seems to think it’s pretty straightforward what the something-in-particular of any given word might be:
It may justly be urged that, properly speaking, what alone has meaning is a sentence…it appears that the sense in which a word or a phrase ‘has a meaning’ is derivative from the sense in which a sentence ‘has a meaning’: to say a word or a phrase ‘has a meaning’ is to say that there are sentences in which it occurs which ‘have meanings’: and to know the meaning which the word or phrase has, is to know the meanings of the sentences in which it occurs.
While he doesn’t define for us “the sense in which a sentence ‘has a meaning’”, he does give us two provisional categories through which one might communicate the meaning of a particular word: “explaining the semantics” of a word and “explaining the syntactics” of a word, the former a linguistic exercise, the latter an experiential one. Take the definition of the word ‘pretentious’. I could provide you with examples of 1) synonyms, 2) antonyms, and/or 3) common usages of the word ‘pretentious’ in the hopes of clarifying for you how ‘pretentious’ is used and how to use it in a sentence–this is ‘explaining the semantics’. I could also, either in addition to or instead of, try to get you to experience the meaning of ‘pretentious’, maybe by performing pretentiousness for you, or by getting you to imagine an interaction that exemplifies pretentiousness, or by showing you the opening scene of Whit Stillman’s Metropolitan–and this is ‘explaining the syntactics’ of a word. Austin’s usage of syntactics here is, as far as I can tell, pretty unique. The implication is that the syntactics of a word stitch together our experience of the world with the ways in which we might communicate that experience, or might understand shared experiences, or even the experiences of others. In a sense, it parallels in physical, social, or psychological terms his linguistically-oriented use of semantics: the latter situates the usage of a word in an associatively linguistic realm, while the former does so in an associatively experience-based one, giving us two different ways to skin the ‘usage’ chicken, so to speak.
–
So what did we get from this exercise, from salvaging Austin? That there is no meaning without context, or, to state it positively, within a given context there can be meaning. The conditional is important here, because once you abandon the idea that words carry within them a universal something-in-particular (or a class of somethings-in-particular) that can be called their ‘meaning’, you shift your framework from unconditional existence and type (‘what is meaning?’) to production and usage (‘when/how/why do words have meaning?’)–or from nothing-in-particular to something-in-particular. To put it a little differently: that words do mean something does not necessitate that they must mean something. Words only have meaning if they mean something to someone–and if that meaning can be communicated to another–so to search for an essential property within a word that provides the necessary precondition for it meaning something is to mistake and confuse what a word does–convey meaning within a sufficient context–for what it is: meaningful.
The origins of art
“Let us look at a rose. Between it and the eye, in the dark room of the blank page, a number of authors are congregated, each with his own torch, his style. We have no perception of the rose without them. Each takes his turn: some torches cast a glowing white light which reveals the rose to be nothing more than a stalk, thorns, and a cluster of waxy petals. We are disappointed with the poverty of this impression. some others seem to be shining their torches straight into our own eyes. It is very colourful indeed; it is not unpleasurable; but after the seduction of dazzlement has been in progress for some minutes, we turn away, tired, knowing nothing new about the rose.
One man…his radiantly beautiful light seems to emerge from the heart of the rose, and intensifies as much as clarifies the perception; the rose suddenly seems the essence of all things to us.”
-Michael Ogilvie Imlah
1.0. Successful communication depends on an underlying similarity between the receiving agent and the transmitting agent. That is to say, if a signal is to transmit information, it must be presented in a form in which it can be understood by the receiving agent.
2.0. In any communication-game, there may arise situations where the message communicated is deceptive. That is to say, that the information communicated does not fully align with the underlying reality that is purportedly communicated by the information, often described as ‘signalling’. This makes sense, because agents often have different utility functions regarding the outcomes of any communication-game. Potentially ‘adversarial’ games of this type are common throughout nature. Clearly, such games are ‘won’ via obtaining a position of informational arbitrage - that agent has an advantage when they possess more information about another agent’s ‘real’ position, than they have about theirs.
3.0. Consequently, on the most basic level, there is an incentive to increase/decrease the valence/direction of any signal sent, as per an agent’s assessment of what will affect the receiving agent’s perceptions of the state of the game in an arbitrage increasing direction. This logic is of course equally valid in the reverse — it is advantageous to have a rival agent believe that one has ‘bought’ their signal to a greater extent that one has in reality.As a result of the above dynamics, signals sent between agents tend to move towards a greater ‘objective’ similarity with each other, as there is an incentive to signal as ‘strongly’ as possible in an advantageous direction, so as to create arbitrage situations, and in order to signal as strongly as possible, the signal must be presented in a form that is compatible with the ‘reading’ abilities of the receiving agent, which will in turn reflect some underlying nature of said agent.
3.1. Communication-games can take place across any timespan. For instance, the evolution by a carnivore of a weapon in the form of a claw, and the consequent evolution of an armoured hide by an herbivore, can be described as this process of similarity-creation. That is, a claw and the armour meant to protect against it, are in some sense more ‘similar’ to each other as a result of this arms-race, than two randomly selected features of animals that are not interacting in this way. This process of feedback-response is vastly more accelerated in more intelligent animals, where each agent is able to update their response to other-agent signals in a much faster way than can be done via genetic selection.
3.1.1. To clarify this further: while herbivores and carnivores are involved in a communication game, this communication, to put it colloquially, takes place on the level of ‘action’ rather than signalling. That is to say, the informational ‘level of play’ is lower. Their communication takes the form of non-deceptive actions by the participants, everything is left as it is. Some animals may develop higher level processes than this - for instance, an animal may begin to transmit information that is not ‘strictly true.’ For instance, many non toxic animals have developed colour schemas that mimic those of poisonous animals, although the updating/development of these schemas towards a similarity still takes place via the slow process of gene selection. At yet another level, you can have signals that update more rapidly - for instance, animals may shriek, posture, etc, to intimidate each other. Finally, for instance, you may have meta, human level signalling games, e.g. in a game of poker, one may not only signal falsely to deceive an opponent, but try to induce him to ‘bet deeper’ on a bad hand by attempting to signal to him that you have fallen for his bluff.
4.0. It seems likely that at some point, given some sufficient sensitivity of agents, the similarity between signals sent will become asymptotic. That the signals sent will become so similar as to achieve some kind of informational transparency to each agent i.e. both agents will be able to almost fully recognise the signal sent, and the signal received, as a result of their extreme similarity, i.e. a carrier of mutually agreed-upon information, which we call art.
4.1. To again clarify this statement, this is a definition of art that is somewhat compatible, and somewhat at odds with vernacular definitions. Typically, in everyday life, an attempt to communicate certain types of information will not be regarded as ‘art.’ For instance, if it is raining outside, and I ask a bystander if it is raining, and he answers that it is, this exchange would not be regarded as art. However, what I would like to stress, at the risk of stretching the definition of art too far, is that even this interaction takes place embedded ‘in art.’ The interaction takes place within a system of mutually understood symbols. Language is art — so much seems well known, but since it’s manipulation and use is so typical of humans, in the vernacular we reserve the designation of art for particularly human-rare informationally dense or creative forms of it.
Agents
Within certain parts of AI research, the term “agent” seems to have recently become a kind of shibboleth. The genesis of this likely derives, in part, from the classic Russell and Norvig textbook, which explicitly defines its focus in terms of what they call agents.1 Others might have in mind a specific technical meaning. More generally, however, it increasingly seems to be a kind of rhetorical way of gesturing towards especially powerful AI systems, the kinds of systems that are different from traditional AI or ML programs, perhaps even the kind that is capable of independent, and potentially dangerous, action or thought.
Etymologically speaking, “agent”, meaning “one who acts”, comes from the Latin agentem, meaning “effective, powerful”, present participle of agere: “to set in motion, drive forward; to do, perform; keep in movement”. Via a hypothetical proto-Indo-European root “ag-”, it connects to a whole host of terms like action, agency, assay, exact, mitigate, strategy, etc.2
Perhaps most commonly within AI, the primary components for something to be called an agent with AI are the ability to “perceive and act in some environment.”3 Others place greater emphasis on something like goal-directed behavior, perhaps even the ability to make and execute plans to achieve those goals. Rusell and Norvig push this even farther, suggesting that, unlike other types of computer programs, “agents are expected to do more: operate autonomously, perceive their environment, persist over a prolonged time period, adapt to change, and create and pursue goals”, though it is somewhat unclear if this is meant to be definitional, or whether all these conditions are necessary (or sufficient).
Unsurprisingly, all of this gets into messy philosophical territory very quickly. Perceiving and acting are relatively easy to think about (although ambiguity remains), but what does it mean to have a goal, or to be effective? (Many people are highly ineffective at accomplishing their goals, and yet it seems strange to say that they don’t have agency).
For example, on some level it seems reasonable to think of a thermostat as having a kind of agency. It perceives the world via a temperature sensor, and it can take actions that modify its environment, by turning a heating or cooling system on or off. It’s also fairly natural to think of it as having an implicit goal, namely to keep the temperature of its room within some particular range of values.
Contrast this to something like a chess playing program. Both the thermostat and the chess player are essentially just mathematical functions, implemented in silicon, that when called, will determine a particular course of action and respond accordingly. These actions might be said to be “effective” to the extent that they help the system achieve its “goal”.
For the chess program, it seems fairly obvious that it should have the goal of winning the game, but can this be fully disentangled from effectiveness? What about a really bad chess program that is easy to beat? Does it still have the goal of winning?
To some extent it seems so, in that it will make moves that will allow it to win against a novice opponent. On the other hand, it is more generally making moves that will rarely lead to victory. Perhaps it is more sensible to think of its goal first and foremost as making legal moves, some of which will be winning. (We could also contrast this to a program that tries to win by any means necessary, including breaking the rules in ways it thinks its opponent won’t notice).
One difference between a chess program and a thermostat is that the former only responds when told to move, whereas the latter is constantly vigilant, responding with an action as soon as its conditions are satisfied. This also seems to be mirrored by what we think of as background software agents that make up part of an operating system — waiting for some conditions to be triggered, and then calling some subroutine (such as turning off the display).
But then what about a rock, or a chunk of glass? In just sitting there, it is “perceiving” the environment (sensing vibrations in the ground and transmitting them through its body). Most of the time it will remain stationary, but if pushed on a slope, it will begin to move and slam into something below it. Why don’t we think of the rock as an agent with the goal of remaining still, until it is pushed? (Certainly I would be unable to remain so still, even if that were my goal).
Language models like GPT-3 are currently more like the chess program than the thermostat. They are mathematical functions that carry out a computation when called, returning some output, though one could of course connect them to some kind of mechanism that would trigger under particular conditions. They are certainly “effective” (to some extent), in terms of their ability to produce text that seems coherent, but does it make sense to think of this as their goal?
I understand why people are eager to mark AI programs in some way, to call attention to what seems like their unusual ability to act “appropriately” in varying circumstances. But the use of the term “agent” seems to do much more work rhetorically than practically, a kind of request that you think of something as possibly being somewhat more than it is.
Stuart Russell and Peter Norvig. Artificial Intelligence: A Modern Approach, 4th US ed. Pearson (2021).↩︎
For example, “An Open Letter: Research Priorities for Robust and Beneficial Artificial Intelligence”↩︎
Expressive conformity
In previous posts and essays, we have established the following:
- All real interactions (or “games”) are mixed-motive and variable sum; that is, there is no such thing as a purely cooperative or purely adversarial game outside of toy models.1
- Social life, and in general those situations from which language arises, are and must by necessity be cooperation-dominant.2
- In cooperation-dominant situations—which is to say, in the composite of alignments between projects and interests—legibility (“honest” signaling) in the sense of writing pragmatically true (advantageous) information is a better strategy than illegibility.3
- An interpretation of a communicative act consists in the experienced or perceived (via habitus) distribution of relevant entailments, which correlate with the objective frequency of events in the environment (to the extent that one’s experienced events are a reasonable sample of events in that environmental subspace).4
- All (observed, i.e. ecologically huddled) behavior communicates (“emits information”). This information consists of statistical correlations between perception and pragmatic, predicted entailment.
From these premises, I believe we can assert that public life will be dominated by an inclination toward—nay, necessity of—conformity, which is to say, the self-alignment between communicative acts selected by an agent, and the “reputation” or signification so to speak of the act itself. That is, the interpreted (and thus, to the expressing writer, consequential) entailment(s) consists of the conventional (most frequent) correlations between expressive aspect and implied entailment.
Since this abstract description is difficult to follow, a simple example: Conventional facial and gestural representations of anger will be interpreted as anger. Therefore, to avoid being misunderstood as angry when not angry, or as not angry when angry, an individual must synchronize his behavior to others’ behaviors.5
This conformity is tethered to the “type” of context or frame within which the writing is expressed, for instance, that of a subculture. As people come to know each other and become familiar/familial, this type may consist of a “self” over time Ship of Theseus-style (that a gesture “means” something in the context of a specific, known individual and his history). And we can say that this mimesis would exist without the benefits of imitation and conformity typically attributed by e.g. cultural evolution—even as the intrinsic return on acts is—as marginal value—the anchor which distinguishes and attracts the adoption of new practices in the first place.
Commonly, we reify the conventional, associative, common knowledge structures of reputation and inferred entailment within our culture as an inherent property of the thing itself. We do not think “This style of decor is seen by many in my milieu as trashy” so much as we think “This style is trashy.” This reification is computationally efficient insofar as it gets around expensive and difficult theory of mind, via the assumption that one is talking to a bearer of shared culture. When we are forced to regularly interact with those outside our native culture, this illusion may shatter. The pragmatic underpinning and recipient-indexicality of concepts and carvings are re-bared when their reified forms break on us, become present-at-hand once more.
This is a foundational stance in ethology and animal signaling, and explored in Schelling; in a purely adversarial situation, there is no incentive to listen to to the intentional (i.e. non-leaked) expressions of other players.↩︎
This is explored at length in my “Epistemic Strategies” essay on legibility, illegibility, and pseudo-legibility (i.e. deceptive appearances).↩︎
See “Gaming & entailment” and the “Meaning of Meaning” letter exchange which begins here.↩︎
This is also explored e.g. in “Reference on the Fly”’s discussion of the phrase “I love her amulet!”↩︎
games communication cooperation conformity mimesis meaning entailment strategic interaction language habitus generalized reading reification