Lakoff.md

Lakoff.md

With all the talk about what constitutes a chair lately I thought folks might be interested in a primer on the limits of classical categorization (or taxonomy if you like) and how Prototype Theory can help you avoid spiraling into a nihilistic hellhole where words have no meaning. Don't believe the hype: You know what a chair is.

This post is based primarily on the textbook Women, Fire and Dangerous Things by George Lakoff, which was assigned reading at Evergreen, former home of everyone's second least favorite Weinstein, Brett.

It's important to remember that categories are just tools, and the imperfect taxonomies we see in classical categorization work just fine until they don't. As an analogy if you are measuring something in physics and the object is moving slower than 1/10th the speed of light you don't need Special Relativity; Classical Mechanics will work just fine. It's the point when things don't fit cleanly into our categories that we need the extra tools from Prototype Theory.

A Working Definition Classical Categorization

Let's start with some properties about categories:

1. Categories are abstract containers for things (a Ford Focus is a kind of Car)

2. Categories have a set of properties:

• A Car has an engine, four wheels, and cupholders for Styrofoam protein vessels.
• A car moves people around.
• Cars drive on roads

3. Things that satisfy a category's properties are in that category. Things that don't satisfy those properties don't go in the category:

• A Bus doesn't have cupholders for your Styrofoam protein vessel
• A Semi Truck moves goods around, not people
• A four wheeler drives off road.

4. Categories are ordered in hierarchies/taxonomy

• A Ford Focus is a Car
• A Car is an automobile
• An Automobile is a piece of technology

Categories are axiomatic in Math, Logic, and Science. We don't have to prove that categories exist. Aristotle handled that shit for us two thousand years ago.

How Do Categories Break down?

People from California break them

Psychologist Eleanor Rosch in the 1970s identified two problems with categories hanging out with the Dani people of Papua New Guinea:

1. If Categories are defined only by properties that all members share, then no members should be better examples than others There should be no "prototypical" (ie best) example of a category

Which of these is a better example of a car?

Which of these is a better example of a chair?

2. If categories are defined only by properties that all members share, then they should be independent of the person doing the categorizing

Which one of these is a better example of "Good President"

It's impossible to separate the category from the person doing the categorization. Lakoff writes that we're all categorizing things automatically all the time, that we do so with groups that have fuzzy boarders, and that we even do it within the framework of our language and cognition. We need more rules to properly describe what it means to be a category

Prototype.md

Prototype Theory

Unfortunately, once philosopher dorks figured out shortfalls in system they used their new system to self referentially describe itself. Here's a classical definition of what a Prototype Theory might be:

A theory of categorization in which there is a graded degree of belonging to a conceptual category, and some members are more central than others. We start with a "Best Example" of a category (the Prototype.) We then come up with a set of rules for adding things to the group (the Generator) which describe the similarity between our prototype and other types. Then as we go about our normal offline lives we add or remove things from both our category and our Generator to build our collection.

With that in mind let's bootstrap our definition of Prototype Theory:

The MetaPrototype

Prototype theory is a categorization system that address the shortfall in classical categorization:

The Generator: Some Properties of PT

• Family Resemblances: Members of a category can be related without satisfying every property of that category.

These are all games. They seem to all at least have rules, but those rules can differ

• Centrality: some members of a category are "better examples" of a category Ehh. It's a pool -Dave Chappelle

• Gradience: Some categories have degrees of membership. Some Categories have degrees of centrality

Which one of these nerds is the best example of a Socialist?

Which one of these Shiba are more masculine?

• Basic-Level Categorization: Categories do not naturally form a top to bottom hierarchy. They instead form a web (or graph) where basic categories form in the middle, generalization grows "upward" and specialization grows "downwards"

If you think of the category "animals that could eat me" you might find yourself focusing on the category Chordata (A shark could eat me, an alligator could eat me, a giant eagle could fly out of the sky and eat a baby, a bear could eat me.) Chordata is in the middle of the taxonomy. We 'naturally' focus on what is the center, instead of the most general (all kind of animals could eat me) or more specific (think of all the dogs that could eat me)

• Reference Point Reasoning: Part of a category can stand in for the whole category when we reason about things, without losing the sense of the whole category. That stand in need not be the prototype for the category

People Really like SpongeBob Characters in their memes on Reddit

• Independent Categorical Evaluation: Categories that appear related are not nescarrily dichotomies Just become something is included a category doesn't mean it's excluded from another category, and vis versa

Is the El Camino a Car or a Truck

Reasoning.md

Reasoning with Prototype Theory

To maximize the chance of this post causing a shitstorm, Let's construct a definition of gender using Prototype Theory

Prototype Phase:

Gut reaction, without trying to out-liberal-arts me, which of these things is masculine and which is feminine?

If you're a paleoconservative, you're done! Have a nice day.

If you're thinking "I'm a man but I'm not He-Man" lets fuzz this category a bit

Generative Phase

Applying Centrality

Oh Look it's He-Man and his secret identity prince Adam. Adam is less central to the prototype of masculinity than He-Man, but he's obviously still He-Man. Our category expands!

Applying Gradience

Uh oh it's BTS! They're cis men but their style fits closer to your prototype of femininity. However since you're not a homophobe it didn't shatter your category of the idea of a man, you didn't go make boomer "real men do x" memes on facebook. You go about your day.

Applying Reference Point Reasoning

It's Steven. He'd like an expanded definition of what it means to be masculine or a man. He has traits that are masculine and feminine, and others that have no gender component. You watch the stream and don't have to consciously move your definition of masculinity while you're watching.

Conclusion

When our cognitive models fail to correctly predict or classify it is important that we tune them, rather than throwing them out entirely. Our natural tendency when a system fails is to throw the baby out with the bathwater.

Expanding our categories consciously to have softer boundaries has the tangible benefits in decreasing cognitive dissonance. This can be done without falling into the "everything is a construct" trap.

Stay chill and based.

PS
Conclude these nutz on your chin