Metonym Megathread!

From The Flatland Almanack, on Jacob Clifton's writing for TVwoP:

My favorite Jacob recaps are his recaps of The Apprentice, which I still watch in order to appreciate to the fullest Jacob's recaps. For one thing, Jacob has an awe-inspiring grasp of what one of my friends calls "inferential psychology"---i.e., the careful study of the foot of Hercules in order to infer the rest. For another, he clearly understands the principles of group psychology and group dynamics, meaning that his commentary is instructive and discerning and frequently educational. He also has that one thing you have to be born with if you're going to have it at all: the writer's gift for discerning the telling detail, the defining phrase.

And, perhaps most crucially, our lord & saviour's "Meaning & Pointing". Worth reading in entirety, but I'll excerpt a bit:

Pointing, Goodwin says, is not “a simple way of indicating some prelinguistic ‘thing’ in the surround,” but a complex act of communication that utilizes mutual mental modeling and often language.

Pointers rely on convention, context, grounding (the immediate environment). If any of these things changed, the indexical is untethered!

every seemingly innocent fragment of information is secretly plotting how to escape its context and cause trouble

(@thechickenman I think you'd dig the bit in this essay about "paths," which might help reconcile the narrative & incentive structure frames as we've talked about)

Or, a summary of pointers-as-maps, indexicals as compressions, from "Cartographic Compression":

This is where maps (and trails) come in. Maps are compressions of the aboutness-extraction type. They represent and highlight certain features of the domain, forming a useful model, and ignore or downplay others.

Language itself acts as a store of information, shared among minds, and language tends toward compression. As stories and concepts are shared, they become more compressed, until they reach the final stage: a metonym, a single word that represents a story or concept that conversation partners are expected to understand. A word is the ultimate tl:dr for human communication

(fr Meaning/Pointing)

Mountains last for millions of years. Buildings last for decades and stay in the same place. But nothing is static in information space. So what are the landmarks here, the unchanging (or predictably changing) features that anchor our mental maps? Since nothing is unchanging in this landscape, I think we must create landmarks, or perhaps the illusion of landmarks. Ultimate values (in the sense of Baumeister’s first need for meaning, above) must be carefully maintained through shared social signaling in order to seem unchanging. Identities are a kind of socially maintained landmark, either the identities of existing, living people, or the identities of deceased ancestors, historical figures, narrative characters, or deities. Baumeister’s illusory “fulfillment states” (imagined future states of perfect happiness) can be unchanging landmarks precisely because they do not exist. If unchanging or predictably changing landmarks are necessary for the navigation of information space, and if they cannot really exist, then “seeing through” this healthy illusion poses a risk for explorers.

Our information world is changing much faster than that of our ancestors. In navigating this nauseating landscape, we need fixed points that do not exist.

Intriguingly, Sarah writes that "This points to a salubrious role for both bullshit and absurd conflict." In some sense, I'd argue that this situation of drifting pointers precisely gives way to bullshit; moreover, bullshitters are motivated to uncouple maps and territories adversarially. They benefit from, and help drive, the uncoupling.

From DeLillo's The Names:

We discussed the evolution of letters. The praying-man shape of the Sinai. The ox pictograph. Aleph, alpha. From nature, you see. The ox, the house, the camel, the palm of hand, the water, the fish. From the external world. What men saw, the simplest things. Everyday objects, animals, parts of the body. It's interesting to me, how these marks, these signs that appear so pure and abstract to us, began as objects in the world, living things.

Wondering if the best way to understand metonyms is via information theory and posterior probability. Maybe there's some role for continental "difference" to match up with information-theoretic "distinction."

Goffman 1969's "Expression Games" talks at length about metonyms—the signals and cues, expressions and communications, that we base our larger, abductive assessments of reality on. We regularly use metonymic cues to get at the greater whole (indeed, as the whole can never be fully apprehended, this is the only way...), but this tendency is exploitable; others can design cues and signals in such a way that gives a very different appearance. For example, a plane flying over enemy territory might use the metonym of "the shape of tanks from above" as the premise for later abductions as to the presence of enemy tanks proper. But, this metonym is not the whole, and is exploitable: now, the enemy set up rubber or wooden tanks which are indistinguishable from the air, causing false inference by air surveillance.

One approach is to avoid singular cues, even those which feel most reliable, and instead asses "the meshing of cues." Because "the world is real," Goffman writes (waxing philosophically), and because its "multitude of little events [are understood] in real connection with others," we tend to assume that "when a multitude of independent signs tell the same story, this can be taken for the way things are." (We are close to theories of paranoia and manipulation, here.) That is, when many metonymic implications add up in the same way, that implication is taken as reality. But, Goffman notes, this can and has been exploited in strategic interaction:

just before a major invasion we can expect a diffusion and intensification of the theater of war—the real peak of the dramatic season. For example, just before D-Day the Allies apparently concerted their double agents to feed the Germans the false line that the invasion was to occur at Calais in June. A German reconnaissance plane was allowed to succeed in getting over Dover harbor where it could photograph landing craft that could not make it to Normandy and therefore must be for Calais. In the Dover area badly camouflaged armored divisions could be seen, but not seen well enough to tell that the equipment was made from inflated rubber. Mock-up airfields and naval vessels were employed, and at the same time, real installations were camouflaged to look like barns and outbuildings. Radio messages, interceptable, emanated from a headquarters in southeast England, giving the strong impression that the invasion would not be in the Normandy area; the messages, however, originated in the real headquarters and were telephoned to the false one. A stand-in for Montgomery was in Gibraltar preparing to go to Africa, which argued that the major invasion was unlikely from England. In Geneva "all available copies of Michelin map No. 51 (the Calais-Arras area) were bought up." And the Calais area was bombed twice as much as the Normandy area.

I wanna tie together a couple threads bouncing around in our collective head recently. First, is @hazard's idea of words as decision rules: we care about whether or not X is Y because if it is Y, we will have to act differently than if it is Z (or just "not-Y"). Second is the idea from strategic interaction of "exuded expressions," or the information that we give off naturally going about our daily business—every decision we make is a metonym which is interpretable; an observer can abductively reverse-engineer the larger strategy, motivation, or personality behind the decision, and thereby get a higher-level "key" to the code that is the interpreted subject. Finally, the idea of evolutionary epistemology: an organism knows its environment insofar as it can successfully predict its features. Our knowledge is merely the result of hypotheses which have not yet been falsified (more accurately, from a Kuhnian paradigm-shift perspective, "qualified" or "re-conceptualized") or which have "worked" well enough to be transmitted without necessary revision.

Plants, Popper says in "Towards An Evolutionary Theory of Knowledge," know things about their environment. "Trees know that they may find much-needed water by pushing their roots into deeper layers of the earth... Flowering plants know that warmer days are about to arrive, and they know how and when to open their flowers, and to close them—according to sensed changes in radiation intensity or in temperature."

In other words, knowledge is about fitness, and appropriateness: it is about measuring the environment in some way and producing voluntary acts which help further one's goals in light of the environmental constraints and affordances. An organism must be a model of its environment ("good regulator theorem") if it can weather the dangers, and capitalize on the opportunities, of its environment—an environment which, quite prominently, includes other organisms.

What we are really looking for then, "deep down"—what knowledge is, fundamentally—is "decision rules." If this, then that. If not this, then that other thing. Warm weather to flowering, cold weather to leaves dropping. Or: stimulus --> response. From this perspective, behaviorism errs primarily because it does not understand how contextual these stimulus-response rules are.

Thus, what we are interested in, when making evaluations or abductions about the world (including other organisms) is fundamentally "what kind of thing is this?", where the "kind" of thing is a hidden decision-rule. Pragmatic ontology. Words are one way of making assessment, it is a bit more top-down. "X just is Y, therefore it must have the relevant properties in every situation that are crucial to achieving my goals." This of course errs insofar as language has high generic fit—its distinctions are optimized for non-technical, everyday coordination, and when misapplied in technical domains, or relatively rare scenarios, it loses fit—becomes a maladaptive carving. This is a major problem in legal arenas, particularly when it comes to novel inventions—"is the Internet a utility?" has severe ramifications for Congressional policy, just as "is X utterance hate speech?" has ramifications in a courtroom. This is how Aristotelian syllogism errs: if concept names are handles to a system of family resemblances, there is no single feature that you can rely on every member of the set possessing. If you blindly use category-based decision rules, you will hit a case where X does not have the features that make it amenable to Z course of action, and yet is being used for Z course of action. If you have a rule against games, because you are philosophically opposed to competition, you will end up inappropriately discriminating against pure teamwork-based coordination games.

The second way of making an assessment to the "kind" of thing you are dealing with, and therefore how you ought to proceed in action, is by metonymic abduction. This is not an avocado in the supermarket with the label "avocado" on it, but a blur of dark green color in your visual field. It is not a sign, at a fork in the road, clearly pointing "This way to town" but a set of footprints heading in a direction, which allows one to abduct that most human foot traffic takes a left-turn, and thereby abduct that a large population center (likely "town") is at the path's end. In some sense this feels more "bottom-up": it looks for features and then makes the kind (and, by extension, the action) decision from it.

Now, it seems to me that arguably there is approach to words which treats them as metonyms—a sign pointing "This way to town" does not necessarily mean that town is in the pointed direction; there is no physical, inherent connection between the map and territory. But the sign is a good clue, or piece of evidence, which when brought interpretively into the total context the observer finds himself in, can help advance said observer's goal. I'm not sure what to make of this yet, other than to say that, in practice, it seems we have a habit of reifying our words such that they no longer feel like metonyms but reality itself.

Some examples of metonyms in hobbies (gardening, ceramics, cooking) plus in a plant's biological cycle:

I put an asterisk next to "spring arriving" because, crucially, this is categorical shorthand for what is in reality a pragmatic desire. The plant does not care if "spring" is here, or what spring is, or whether summer follows it. What the plant needs is a certain amount of time to germinate, and for other plants/pollinators to be around so it can be pollinated. In normal, planet Earth circumstances, this is synonymous with spring's arrival, because plants have adapted exactly to fit the regular patterns and cycles of their environment.

A second quick note: the only point of a short-term metonymic test is if the "real" quality you care about can only be decided, definitively, in the long-term. Thus you erect short-term surrogates that correlate, broadly, with desired long-term outcomes. I have a feeling this relates to @hazard's recent ideas on monotonicity vs. "it's not over til the fat lady sings."

From Hotel Concierge's "How To Be Attractive"

"Okay, but in real life we don’t have a search function.” Really? Consider: “I saw her from across the room, and I immediately fell in love.” Fell in love with what? “She had these big thick-rimmed glasses…and an impish smile…and we’re holding hands, and it’s the fall…” Right, part for the whole. She had big glasses, so you typecast her into the story you’ve run through your head a thousand times, the story repetition has lodged in your unconscious Id. She had big glasses, so she was the type of girl you could love. “No, you don’t get it—she looked like the girl of my dreams.” Exactly.