Common vs Expert Jargon

post by Raemon · 2017-09-21T03:05:31.203Z · LW · GW · 17 comments

Contents

  I. Lessons from Game Design
    1. Common Cards
    2. Keywords
    3. Evergreen keywords
  II. Building a high level conversation
    Further Reading
None
17 comments

tldr: Jargon always has a complexity cost, but you can put effort into making a concept more accessible, and it's especially valuable to put that effort in for terms that you'd like to be used by layfolk, or that you expect to be used a lot in spaces where you expect lots of layfolk to be reading/participating.

I. Lessons from Game Design

Magic the Gathering deals a lot with complexity. Each year, new abilities and rules are added to the game. This gives experienced players the chance to constantly discover new things, but it comes with some issues.

First, it makes the game harder for new players (the game kept growing more complex over time, raising the amount of information a new player had to process at once)

And second, even for experienced players: each instance of complexity is a cost. Players (both new and old) can only handle so much, and some forms of complexity are less fun than others. (For example, forcing players to do a lot of book-keeping, rather than letting them make interesting strategic decisions)

Six years ago, their creative director wrote about a new paradigm of Magic design. One of their solutions was to pay careful attention to how they spent complexity points in ways that affected new players.

Three examples:

1. Common Cards

In Magic, when you buy a new pack, 11 cards are "common", 3 are "uncommon" and one is "rare". Experienced players buy lots of cards and can have access to lots of rares, but new players generally just buy a few cards, so most of their cards are common. Therefore, the complexity of the cards at common determines how much complexity newcomers have to deal with.

2. Keywords

One way to reduce "effective complexity" is to bundle concepts together in a keyword. Instead of saying "this creature deals damage to each of the creatures blocking it and then deals the remainder of its damage to the player", it just says "Trample". There's an initial cost of learning what Trample means, but afterwards, every time you see the word "Trample" on a creature it works the same way.

Trample has some neat things going for it: it sounds evocative, and gets to build off of existing ideas in your brain. You already know what a big animal looks like. You can imagine a small creature getting in the way of the elephant, and it slowing the elephant down slightly but not really stopping it, and the elephant continuing on, trampling over it, and then going on to attack some bigger target.

This imagery is helpful for intuiting what the rules mean, even if the wording is somewhat confusing.

The problem comes when you introduce too many keywords at once. It gets overwhelming. Which brings to a final concept:

3. Evergreen keywords

Every 3 months, new magic cards are released to keep things fresh. New keywords are introduced (usually 3-5).

But there are some keywords (like Trample) that are *always* in season. There are about 16 evergreen keywords. Many of them are pretty intuitive (such as flying creatures only be able to be blocked by other flying creatures) so they aren't hard to learn.

A new player has an implicit goal of "learn all the evergreen keywords", which is a manageable task.

II. Building a high level conversation

I think some of this applies to the rationalsphere, where a lot of important concepts have been built up, or, combined together from neighboring disciplines. (See Anna Salmon's Single Conversational Locus)

Jargon is *useful*. They let you summarize a complex concept in a single word, and then have deeper conversations where each word packs a lot more meaning.

I have a lot of thoughts about how to do jargon *right*, which are beyond the scope of this post. But to summarize, I think good jargon:

  • encapusulates an idea that's important to build off of

  • lets you distinguish between *similar* concepts that have importantly-different-nuances. (viral infection vs bacterial infection)

  • provides some context clues that help you learn it (the way Trample does), while...

  • ...not *also* resulting in people confusing what it means (a bad example perhaps being "negative reinforcement", which is not actually the same thing as "punishment")

Some considerations:

1) Sometimes you want a 101 space where you're either introducing ideas to a broader audience. Sometimes you want a 201 space where you're building on those ideas (either helping somewhat-less-newcomers build up a more advanced understanding, or literally developing new content at the cutting edge)

2) Different venues of conversation can have both different expectations of who-is-participating, and different social norms of what kind of participation is encouraged. (i.e an academic journal, a semi-formal internet forum, a facebook post)

3) Some concepts are pretty standalone: layfolk can learn them and use them immediately without having to fit them into a big edifice of theory

4) Furthermore, some concepts make good "gateway" terminology. They're useful standalone, but then they open up a world of ideas to you that you can then further explore.

So my thought is basically: if you are developing jargon, pay extra attention to whether this is Common or Expert Level jargon. There's not a clear dividing line between them, but roughly:

Common Jargon means you're expecting it to be a useful enough idea for layfolk to use regularly (or, you'd like to be able to have conversations with layfolk, or write popularization articles, that rely on the term already percolating into the mainstream, or, use it as a gateway term)

Consequently, it's much more important to put a lot of effort into choosing a term that:

  • resonates easily, is memorable...

  • ...but avoids people latching onto the wrong aspect of it and misinterpreting it

  • doesn't sound like a weird insider term...

  • ... but maybe ideally hints at a broader ecosystem of ideas

Expert Jargon is only really useful if you're buying into a broader ecosystem of ideas that build on each other. Accessibility and avoiding misunderstanding is still important if possible but being precise and build-on-able is more valuable.

Further Reading

This post was inspired by and builds upon:

Complexity is bad (Zvi Mowshowitz)
The Purple Sparkly Ball Thing (Malcolm Ocean)
New World Order (Mark Rosewater)

17 comments

Comments sorted by top scores.

comment by ozymandias · 2017-09-21T14:35:39.815Z · LW(p) · GW(p)

One problem I've noticed with common jargon is that jargon that starts off with a single, precise meaning often evolves to have a broader and less useful meaning. Since I tend to do a lot of arguing about social justice, the examples that occur to me are "gaslighting," "intersectionality" and "demisexual." It seems like providing context clues might help: none of those words have particularly obvious meanings, and "demisexual" in particular sounds like it means "somewhat sexually attracted to people" rather than "only sexually attracted to close friends." But it seems to me that that's not 100% of the problem. When jargon is used by a 101-level audience, their misunderstandings can build on each other to the point where even experts cannot use the term without being misunderstood.

Replies from: Kaj_Sotala, ingres, Raemon, Toggle, ozymandias
comment by Kaj_Sotala · 2017-09-22T12:44:38.035Z · LW(p) · GW(p)

It feels like for political concepts, they are more likely to drift because people have an incentive to make them shift. For instance, once it gets established that "gaslighting" is something bad, then people have an incentive to shift the definition of "gaslightning" so that it covers things-that-they-do-not-like. That way they can avoid the need to *actually* establish those things that bad: it's already been established that gaslightning is bad, and it's easier to shift an existing concept than it is to create an entirely new concept and establish why it is a bad thing. (It's kind of a free riding on the work of the people who paid the initial cost of establishing the badness.) I would guess that less loaded terms would be less susceptible to it.

Replies from: Raemon
comment by Raemon · 2017-09-22T23:22:00.877Z · LW(p) · GW(p)

Huh, yeah. In retrospect this is obvious but putting it that clearly makes it easier to incorporate into broader plans or strategies.

Musing over Ozy's three examples:

- gaslighting is a negative-affect term. Now that it's pointed out, I expect affect-loaded terms (positive or negative) to be particularly vulnerable to people fitting onto things they like or don't like.

- "intersectional" isn't precisely affect-loaded, but it carves social reality into pieces that are then intended to be politicized, which makes it easy for second order distortions to happen. (I guess it's valence-loaded insofar as it says "intersectionality is a good way to think about things", but I think at least half of the issue is that it's referring to a complicated subject and there are other vaguely-similar-seeming subjects one could reasonably have also named "intersectionality")

- "demisexual" is less loaded, and I think is a more straightforward example of "a confusing term."

Replies from: Raemon, lmn
comment by Raemon · 2017-09-22T23:24:17.933Z · LW(p) · GW(p)

My current vague takeaway is something like "when you're forming plans that involve changing the social world via terminology, pause to reflect on the fact that your terminology is going to get distorted to benefit people."

comment by lmn · 2017-09-23T03:44:59.912Z · LW(p) · GW(p)

"intersectional" strikes me as an example of an intentionally confusing term, at least I've never been able to figure out a meaning beyond "a word people throw into arguments to make it a norm violation to notice that said arguments make no sense".

comment by namespace (ingres) · 2017-09-21T22:32:01.105Z · LW(p) · GW(p)

Perhaps one thing you could do is just accept this drift is going to happen, and in the first usage of a jargon term in a document provide some kind of identifier which points to a clear definition somewhere. This would be a bit like citations and provide exactly the meaning the author intends. Punish people who do this without actually meaning it that way, or doing the citation and then using the word contrary to the cited definition.

If SJ people did this as standard and accepted that some people are old and going to use terms like "transexual" because that is what was standard when they learned it that would alleviate a lot of my issues with that particular corner of discourse.

Replies from: lmn
comment by lmn · 2017-09-23T03:47:45.590Z · LW(p) · GW(p)

Um, the whole point of SJ usage is to win arguments through intimidation, thus this would be conterproductive.

comment by Raemon · 2017-09-21T23:37:43.570Z · LW(p) · GW(p)

Yeah - I think there's multiple posts-worth of "how to think about jargon" (some of which I feel qualified to write, and some of which are more of "open problems"), and the problem of "how to avoid dilution" is important.

A useful first-step (when writing a blogpost attempting to coin a jargon term, or popularize it) might be to check "does the term I'm creating seem similar to another concept that people will want to use" (in particular, a concept they'll want to use more often than the the one I'm pointing at), and then specifically highlight those similar concepts (perhaps giving them a name people can use so they don't accidentally use mine)

comment by Toggle · 2017-09-22T03:30:12.480Z · LW(p) · GW(p)

There are definitely domains where this isn't a problem at all- for example, geology terms like 'tufa' or 'shale' seem basically static on the relevant timescales. So it's probably possible to completely solve the dilution problem, it at least some cases.

There are at least a few relevant structural differences between social justice and geology, but I'm not sure which ones are the most important. The main three advantages for geology's stability that I can think of are A) Rocks are boring, and not emotionally charged by tribes and sex and so forth. People are rarely motivated to stretch definitions to cover their preferred cases B) There's a structured process of learning and most of the jargon occurs within similarly structured professional environments, without a whole lot of self-educated geologists talking about rocks on the internet. C) Rocks are well-understood down to the level of thermodynamics, so every term of art can in principle be dissolved down to some precisely defined configuration of atoms, rather than bottoming out in human psychology.

Some of those are more hopeful for rationalist jargon than others, I guess.

comment by ozymandias · 2017-09-21T23:40:31.736Z · LW(p) · GW(p)

Hypothesis: The problem with your proposal is that, as a practical matter, no one ever clicks on the damn links and reads the definition. It makes sense from the reader's perspective-- they assume they know what you mean, why would they bother reading a link? Even if you define the word in the post, many people will skim over the definition. I also think definitions are sort of 'sticky': for example, if someone wrote a post where 'chair' was defined to mean 'table', you would probably end up confused when they suggest putting food on the chair, even if they'd put in a sentence explaining that they mean a chair to be a flat thing with four legs.

I also dislike calls to punish people for things. It seems like figuring out how to use jargon well is hard, and I think rewarding people who get it right is going to be a better course.

Replies from: Raemon
comment by Raemon · 2017-09-21T23:41:48.592Z · LW(p) · GW(p)

Meta - is there a reason this isn't a reply to Hypothesis? (it shows up earlier than his comment in my default-sorting)

Replies from: ozymandias
comment by ozymandias · 2017-09-21T23:43:04.132Z · LW(p) · GW(p)

Because I am confused about how to use the website, I think. :)

Replies from: Raemon
comment by Raemon · 2017-09-21T23:52:40.483Z · LW(p) · GW(p)

Coolio. :)

Do you know if the issue is that the 'reply' button is not easy-to-read enough (or more generally, does this seem like a place where the UI should be different, or just a place where you were in a rush or whatever?)

Replies from: ozymandias
comment by ozymandias · 2017-09-22T00:08:51.736Z · LW(p) · GW(p)

I think I am used to commenting in places where the comments are a single line and not ranked by upvotes/downvotes, so if you want to respond to someone you go to the "reply" button in the original comment. It seems hard to come up with a UI that would direct me away from this default behavior (other than having my comments be misthreaded until I get it, I guess :) ).

Replies from: Raemon
comment by Raemon · 2017-09-22T02:14:06.418Z · LW(p) · GW(p)

Gotcha.

comment by dogiv · 2017-09-21T13:35:36.297Z · LW(p) · GW(p)

I think this is an excellent approach to jargon and I appreciate the examples you've given. There is too much tendency, I think, for experts in a field to develop whatever terminology makes their lives easiest (or even in some cases makes them "sound smart") without worrying about accessibility to newcomers.

... but maybe ideally hints at a broader ecosystem of ideas

This sounds useful, but very hard to do in practice... do you know of a case where it's successful?

Replies from: Raemon
comment by Raemon · 2017-09-21T19:04:36.463Z · LW(p) · GW(p)

I'm not sure if there are great examples (part of the problem is that jargon is hard), but I think "epistemic vs instrumental rationality" are sort of in the right direction.

They're not common-jargon (you'd only use them frequently if you were buying into the entire ecosystem of rationality-thinking), but they are relatively easy to explain, I don't think I've ever heard anyone misuse them, and they highlight that there's a lot more rationality worth learning.