Is Nuance Overrated?(chronicle.com)
chronicle.com
Is Nuance Overrated?
http://chronicle.com/article/Is-Nuance-Overrated-/232771/
5 comments
lol - "this guy" doesn't like rigor: that's an awesome lack of googling this guy. These comments seem so enamored by the title that they overlook the content of the article where Healy trips over himself repeatedly stating that he's not arguing against nuance.
>a verbal/symbolic model meant to account for the complexity of the world
I think you and I have very different notions of the word "model." Rational spherical cows in a vacuum may not be precise or nuanced, but that doesn't make them inaccurate, and the simplified view is a useful model for learning about big-picture concepts. It's true that it won't catch every corner case, and when you're making decisions about human beings sometimes that's dangerous, but there's value at every level of granularity. Looking at each grain of sand and learning everything about the rock it came from will let you put it in the perfect place, but you'll never build a sandcastle that way, and if you want to reshape the whole beach you'll need yet another level of focus.
Generalization is a good thing. It's powerful. Humans got where we are by spotting patterns - not because everything matches the pattern perfectly all the time, or even ever, but because you can see general commonalities and start to abstract them away into something that can be reasoned about.
If there's one theory that explains more of the detail and variation in the world than another, at a finer granularity, without giving up generality, then yes, it's clearly a superior theory. But throwing out highly general theories just because they aren't perfect is intellectually lazy.
I think you and I have very different notions of the word "model." Rational spherical cows in a vacuum may not be precise or nuanced, but that doesn't make them inaccurate, and the simplified view is a useful model for learning about big-picture concepts. It's true that it won't catch every corner case, and when you're making decisions about human beings sometimes that's dangerous, but there's value at every level of granularity. Looking at each grain of sand and learning everything about the rock it came from will let you put it in the perfect place, but you'll never build a sandcastle that way, and if you want to reshape the whole beach you'll need yet another level of focus.
Generalization is a good thing. It's powerful. Humans got where we are by spotting patterns - not because everything matches the pattern perfectly all the time, or even ever, but because you can see general commonalities and start to abstract them away into something that can be reasoned about.
If there's one theory that explains more of the detail and variation in the world than another, at a finer granularity, without giving up generality, then yes, it's clearly a superior theory. But throwing out highly general theories just because they aren't perfect is intellectually lazy.
Exactly, I thought the original article was brilliantly written, but I'm surprised it didn't mention these related concepts:
https://en.wikipedia.org/wiki/Map%E2%80%93territory_relation
https://en.wikipedia.org/wiki/On_Exactitude_in_Science
https://en.wikipedia.org/wiki/Bonini%27s_paradox
https://en.wikipedia.org/wiki/Map%E2%80%93territory_relation
https://en.wikipedia.org/wiki/On_Exactitude_in_Science
https://en.wikipedia.org/wiki/Bonini%27s_paradox
Sociology is full of chaotic systems, so slight deviations in the premises can have wide effects on the outcome of the system you are describing.
I would replace your cow analogy with a 2-pendulum analogy. Depending on the nuances of what is being observed, a small change in the inputs of your system can create dramatically different behaviors. It's important to account for how you define your initial conditions and variables so that you don't end up with a model that matches early in a simulation but ultimately diverges into something that looks nothing like what can be observed in the world.
I would replace your cow analogy with a 2-pendulum analogy. Depending on the nuances of what is being observed, a small change in the inputs of your system can create dramatically different behaviors. It's important to account for how you define your initial conditions and variables so that you don't end up with a model that matches early in a simulation but ultimately diverges into something that looks nothing like what can be observed in the world.
I don't buy the 2-pendulum analogy. Nuance, as I read in the article, is about detection of ever decreasing observed quantities ad nauseam. It would be fine-grained discussion of the non-chaotic ranges of said pendulum. Basically a pissing contest, which is as good as any, but still just that. Once it goes beyond any practical use, has no more value.
I think this "discovery" is less about nuance in general and more about nuance in sociology. But that, again, would require some nuance.
Relevant XKCD: https://xkcd.com/451/
Relevant XKCD: https://xkcd.com/451/
The paper underpins the political philosophy of President Dwayne Elizondo Mountain Dew Herbert Camacho.
When you make something a metric, people will game the system. Nothing new. Nuance is a tool, and should not be a goal.
AKA: bikeshedding is a waste of time.
> It is not the job of theory to verbally reproduce the complexity of the world.
Yes, it is. That's what a theory is -- a verbal/symbolic model meant to account for the complexity of the world. I think this guy just doesn't like rigor and would rather apply blanket statements about the world without having to back them up with actual data.