1.

Spotted: a dirty brown bag crushed under a tennis shoe. Looks like someone won’t be eating dinner tonight. XOXO… Gossip Girl. The thing about voices is they are contagious.  I can’t do accents, but spend a week watching TV re-runs and I can GPT-3 an essay in the narrator’s delivery.

from a Discord server somewhere near you.

Most people who play Tetris are familiar with the Tetris Effect, through experience if not name—the blocks streaming down the eyelids before bedtime, gamespace continuing via closed-eye visual. And those who seriously game may have experienced other, similar “game transfer phenomena”—mild closed- or open-eye hallucinations, the framing and transformation of real-life perception by the paradigms of the game. The boundaries on what constitutes game transfer phenomena are hazy—full-blown hallucinations at one end, the mere use of game metaphors (“leveling up,” “scoring points”) at the other. But what members of this set have in common is that sensory inputs in one cognitively and attentionally intensive task become top-down psychological projections onto another task or domain. (In this way, they resemble the waterfall illusion.)

But there is a broader, less-noted class of effects, which, along with my fellow Pfeilstorch member WithoutCourage, I want to call (merely) “transfer phenomena.” The same bleed from media to reality occurs regularly with literature and film: the narrator’s tone and sentence cadence, in a Jane Austen novel, takes over the reader’s self-narration of everyday life; the anxious headspace of a protagonist leaks into the reader’s everyday headspace. 

Research has turned up a handful of personal accounts on Reddit and Twitter threads, excerpted here:

Just a few days from twitter feed to Jane Eyre, my thoughts, even on day to day tasks, bear resemblance to the writing I’m reading. Like I’m still confused.. I would never think to say “bear resemblance”.

If I read something with an Irish or Scottish narrator, I find myself thinking with an accent, haha.

[this is] why i exclusivley [sic] read noir detective novels. life is more inarresting [sic] that way, see.

We’ll call this specific type of transfer style transfer, after the concept in deep neural networks, in which the style from one image is transferred onto another image (a reformulation of its content according to the stylistic form of the origin source)—e.g. the style of a Van Gogh landscape transferred onto a modern landscape photograph.

At its more extreme end, this transfer does not merely re-style one’s inner narration, but one’s entire sense of self:

If I read from the perspective of a scared, paranoid girl, I’ll be shrinking back from things for a while. While other times I’ll be reading from the perspective of a conqueror, and be more confident. I have a lot of empathy and really get into what I’m reading/watching

i’ve always been extremely influenced by some of the books i read and love. and i’ve noticed that it’s not just my internal monologue that changes: its also in the way i might talk to other people, my passions and hobbies and even how i perceive my life. for a while, i subconsciously act as if i’m a character in the book i’m reading, as to live inside of it– romanticizing everything.

We’ll contrast this with frame transfer, which is the porting or projection of an interpretive frame, or hermeneutic, taken from one subject onto another. One interviewee gives the account:

i don’t have a word for this but I have a distinct memory of it happening with art- the first time I went to the Institute of Contemporary Art in Boston (i was young, maybe high school freshman?) I was walking through all the exhibits and having my mind blown because I didn’t realize you could do that sort of stuff with art. Then I turned a corner to what I thought was going to be another exhibit but was actually a place to view the harbor, and everything looked completely different and magical. I probably spent 5 minutes just staring at those floating orange things they use to keep waste runoff contained

Adam Strandberg

In the former case, a style of cognition which is directly modeled in a piece of media, and which the reader must inhabit (“take on”) as their own (e.g. inner voice) bleeds into experience and cognition outside encounters with that media. When I finally resurfaced, blinking and dazzled from [Tolstoy, Poe, Laclos], it took me days to stop thinking in their cool polished, crystalline rhythm.[1] In the latter case, a framework of interpretation which is necessary, or which proves productive, for one set of stimuli is brought to bear on another set of stimuli. That is, style transfer is the structuring of one’s expressive output by the principles of recently experienced expressive inputs. Frame transfer, on the other hand, is the structuring of perceptual set through the lens implicit in, or required of, some previous set of perceptual encounters. You’re walking down the street after a Grand Theft Auto session, and a flickering impulse says, “Smash the window; jack the car.”

2. 

The purpose of language is coordinating. Some linguistic use, of course, is adversarial and deceptive—but, just as fraud requires a high-trust society (otherwise no one would entrust a stranger with their savings), deception is a freerider atop genuine communication, and reputational sabotage freerides atop gossip’s genuine function as a channel for social information. As with parasites and mimics in the natural world, such entities are subject to frequency-dependent selection dynamics. If too many red dart frogs are phonies—are not actually poisonous, then predators begin ignoring their claims to protection (to toxicity). If an information channel is too often manipulated or exploited, rather than providing useful information, then it will quickly be ignored.

How does language work? A mental model of the world is built up, extracting patterns, objects—constructing family resemblances, which we then connect to a handle, a name, a word or phrase. Communication is premised on your and my historical exposures and thus correspondences of these handles to reality being similar enoug that we end up, pragmatically, talking about “the same thing”—you understanding my reference in pragmatically the same way (the same in all relevant ways to the telos of the interaction at hand).

Language and culture together provide a landscape of reference points which allow coordination. If one were attempting to coordinate physical movement, among multiple parties, across a physical landscape, one would naturally end up relying upon the natural referential affordances of that landscape. Thus, one would meet at “the top of the mountain,” or at the top of a waterfall, or by the pond. All parties would have to have roughly similar associations between landmarks and names, such that each ends up at the same spot when “mountain” is meant, or waterfall, or pond. One party might warn another of a mud patch or obstructing bramble, of closed-off roads or private land; might instruct one to hang left at the first fork in the road, and hang right at the second, finally crossing over the pasture behind Bill’s old lot—you know Bill, lived on the outskirts in the 80s?

Here, we have begun veering into culture—a shared cultural memory of residents, and their names, and their farms. And when we are attempting to coordinate more vague and abstract behaviors than physical landscape, our cultural affordances become crucial. We may speak to a certain vibe, by clustering adjectives, to coordinate a social event between different organizers, or a new product among different design and production teams: “a bit of a Pacific Islands thing, you know, with the tiki torches—I mean, not literally tiki torches, which are terribly démodé and tacky, but that kinda atmosphere.” 

When individuals are engaged in long-term coordination, be it a romantic or professional partnership, a long-term friendship, etc, they naturally build up a set of new, commonly known (in the sense that both parties know the other party knows) set of reference points. A family that has lived on the land a long time will have natural, shared memories to different sub-plots and histories, and can forget that others do not naturally share these reference points—a friend invited her SO to a hunt deer, with her father, on their land. The hunting required semi-elaborate tactical maneuvers, where some parties would wait at designated spots while the other members of the group flushed the deer through the area the first group waited. Understandably, the SO was baffled by commands to “stand near the old beekeeper stand,” and ended up standing in the wrong location when the deer were flushed toward him—allowing them to safely escape trophy-dom on the father’s wall.

When people talk about how having experiences together is crucial for building a relationship, it’s because those experiences provide coordination points which can later be referenced. “Oh, it’s like that thing in Havana in ‘08.” Douglas Hofstadter talks about being at the Grand Canyon with his young son Danny, who preferred to play with the ants over gazing into the canyon—a preference for small scale instead of large. He remarked on it with his wife, perhaps discussed it for a few moments, and moved on. Years later, traveling with his family in Egypt years later, he noticed the young child of a friend was more interested in collecting local bottle caps than in the Great Pyramids—and remarked that it was like “Danny at the Grand Canyon.” Suddenly, from what Hofstadter calls a singleton—a single experience or object—typification, or analogic extension, occurs, and scenarios similar to the original can be described using the handle of the singleton. The phrase has become, if it was not already, a concept—and Hofstadter began using it more regularly with family and colleagues.

Why does this happen? The first thing to note, obvious as it may be, is that our language does not fully cover the set of phenomena we care about. Many parts of our experience exist as little more than feelings. Many of the patterns which recur through our lives have no name. Many of the important nuances or repeating patterns of the coordinating group’s interpersonal dynamics, or their potential points of conflict, will not have handles already; the generic fit of the English language can’t possibly accommodate them. At the same time, gaining control over them—being able to notice, and intervene on, or learn from, the pattern—requires being able to think of the pattern as pattern, that is, as a structural regularity with distinct features or properties. And being able to coordinate around it—to intervene on, or learn from, the pattern jointly—requires being able to talk about it, and for a similar-enough concept to exist in all parties’ minds, in connection with whatever verbal “handle” (e.g. “Danny in Grand Canyon,” “Havana in ‘08”) becomes attached to it. Thus, members of these relationships must make non-things “things” all by themselves: a successful relationship, like the inexact sciences, is a series of parallel pipelines from non-thinghood to thinghood.

In this way, individuals in long-term relationships, naturally needing to coordinate, slowly form new, semi-private dialects with one another. A groupchat’s tendency to produce novel language is the same phenomenon, meeting the same needs, as a couple that references “that fight back in ‘08” as a way to prevent future fights. The “type of guy” theories popular on social media, in which young women create semi-niche stereotypes of different men they have dated, and the general varieties of men, testify to an impoverishment in their languages that prevents them from more precisely talking about their dating experiences before these types existed. In order to coordinate with increasing precision, bespoke language is required, which is built from previous experience and—crucially—shared cultural exposure, which can provide new reference points that are often pre-patterned (the author or artist has presented patterns through multiple examples, so that they are not merely singletons). That is, they build “linguistic capital” together—empowering themselves to more accurately, precisely, and efficiently communicate, and thereby, coordinate. The young women writing “type of guy” theories are creating “landmarks” in their social-romantic “landscape” which can then be referenced—“I’m looking for type X, but always end up with type Y”; “Type Z’s are the worst; all my experiences with them have been bad.” 

And a key technology for this dialect formation is linguistic and interpersonal synchronization. Mutual imitation, during linguistic interaction, is a widely recognized linguistic phenomenon which appears to improve the outcomes of interpersonal interactions, and to facilitate language learning. (Typically, significant variation is found in how much a given individual, or interaction partner, synchronizes to the other.)

In thinking through why we might have the cognitive equipment to perform style transfers, one possibility I’ve entertained is that this is an extension of, or connected to, our natural abilities to synchronize with each other, and imitate others’ speech patterns. 

People seem to have tones and registers—dialects, we might call them—that they prefer to operate on, and that when people interact, they varyingly learn the other’s dialect and adopt it. There is a lot of variation in interactants’ flexibility—some people have basically one register, and either you adapt to them or interactions between you two will always feel slightly strained. Others change their own registers more flexibly, so that the more stubborn speaker has more “gravity.” And when you have two relatively flexible people it’s a matter of meeting in the middle, and you develop a particular dialect to your social interaction. This, along with the shared construction of an experientially-based vocabulary, creates the interpersonal dialects discussed.