Why parsimony?

[ADigitalArtist]

Well, that's great if you are talking about theories. What about other things in life? Keep in mind I did not intend to limit this topic to some particular human endeavor. If you're a storyteller writing a novel, what makes a story element "unnecessary" because it is complex? Just as an example.

It's worth pointing out that our most endeared and long lasting stories trend towards simpler hero's journey stories. There's even a writing rule/advise called 'kill your darlings' usually said to people who are thinking they've wrote a particularly original, complex or deep character, scene or story. When they're really probably just being self aggrandizing. Probably originating from William Faulkner's quote "Whenever you feel an impulse to perpetrate a piece of exceptionally fine writing, obey it – whole-heartedly – and delete it before sending your manuscripts to press."

 

[beenherebeforeagain]

Do they? My experience learning how to identify various plants makes me wonder about that. Minimalist or simplistic observations won't do it for you, especially if you're trying to identify something from the grass family. A grass will look like a grass.... particularly to non-botanists. Hell, non-botanists call things that are actually sedges or rushes "grasses" even though they are not. That brings up the issue of presumed patterns being erroneous (did you know that birds are reptiles from a phylogenetic and evolutionary standpoint?), mistaken attributions (that moment when that person you identified as a male turned out to be a female), and so on and so forth. But I'm getting rambly now.

And dang you, I wish I lived in the forest. All the forest properties in my area are way over my price range. *sniffles*

Forests are made up of Christmas Trees and All Others...

 

[Revoltingest]

How is it determined that a particular element is "unnecessary complexity?"

If one eliminates an element, & predictive value isn't reduced, then it was unnecessary.

I mean, I don't know about you guys, but I find simplistic storytelling hella boring. Seen more simplistic "good vs evil" morality stories than I want to shake a stick at.

Even in storytelling, tis best to avoid unnecessary complexities.
The Matrix would'a been a better movie had the writers not
introduced cold fusion energy generation using humans as the
reason the machines farmed them in the elaborate artificial reality.
To give no reason is better than giving a stupid one.

 

[beenherebeforeagain]

Why would Occam's Razor apply to storytelling?

Why wouldn't it? Genres like murder mysteries thrive on providing clues from which one can build various models about who committed the murder and why...but there's only one from the storyteller's perspective--and it can be any solution from the simplest or the most complex...and the quality of the writing determines whether or not the reader will be surprised at the conclusion...

 

[Quintessence]

Why would you include factors in a model when you don't know whether they have an effect or what sort of effect that might be?

In the sciences, we have statistical tests to tell us how well the factors we've assumed explain the data we have. If the data is well explained by some set of factors, why add others?

I see parsimony in the sciences as just an extension of the principle that we shouldn't assume things without justification.

This is the last post I really wanted to get to, so pardon the disjunction.

Given I did modeling work in grad school, man, there are tons of reasons to include factors in a model when you don't know whether or not they have an effect - that's how you test if they have an effect... lol! Honestly, getting good, reliable data becomes the bigger limiting factor on what variables you include. There were factors we wanted to include, but didn't, because the data just wasn't available. Even if the data is "well explained" by some set of factors, there is good reason to add more because it can increase the power and accuracy of the model. Granted, I also worked with ecological questions, for which "silver bullet" simplistic explanations are pure fantasy.

 

[Revoltingest]

Why wouldn't it? Genres like murder mysteries thrive on providing clues from which one can build various models about who committed the murder and why...but there's only one from the storyteller's perspective--and it can be any solution from the simplest or the most complex...and the quality of the writing determines whether or not the reader will be surprised at the conclusion...

Regarding writing, I've often heard the advice....
- Don't use big words when a diminutive one would do.
- Don't introduce distracting characters, plots or devices.

Certainly, stories can be complex in structure & rich in language.
But this should serve the purpose of the writer, & not be gratuitous.
Consider use of violence.
"Into The Badlands" is a violent TV show, but the violence is uninteresting & pervasive.
"A Clockwork Orange" is an even more disturbing violent movie, but all the violence
therein is essential to the story. (Great movie, as is the book.)

 

[beenherebeforeagain]

Regarding writing, I've often heard the advice....
- Don't use big words when a diminutive one would do.
- Don't introduce distracting characters, plots or devices.

Certainly, stories can be complex in structure & rich in language.
But this should serve the purpose of the writer, & not be gratuitous.
Consider use of violence.
"Into The Badlands" is a violent TV show, but the violence is uninteresting & pervasive.
"A Clockwork Orange" is an even more disturbing violent movie, but all the violence
therein is essential to the story. (Great movie, as is the book.)

Yes, but again that's the skill of the story-teller. How much, and how depicted, character/violence/etc., is "absolutely necessary?" The author/editor/producer/director/publisher's version is one interpretation. What a reader thinks is another. When the two match up, you've got a hit on your hands.

When young, I read the Lord of the Rings trilogy, which to some is overlong and boring because Tolkien is always off into history and mythology that really doesn't have much to do with the present action. Anyway, because I liked the LOTR, I thought I'd try another fantasy classic, The Worm Ouroboros...which I found to be long, boring and unnecessarily complicated.

Personal preference. Some just love Game of Thrones...to me it's a not very interesting soap opera, as a book series or as a TV series. It was simply written...but here, let's try following roughly 50 main characters, some of whom will check out a few chapters after they are introduced, while others won't be introduced until Book/Season Five...Others will disappear for years, only to suddenly reappear...

Again, what is "necessary and simple" for one may not be for someone else...

 

[Rick O'Shez]

Given I did modeling work in grad school, man, there are tons of reasons to include factors in a model when you don't know whether or not they have an effect - that's how you test if they have an effect... lol! Honestly, getting good, reliable data becomes the bigger limiting factor on what variables you include. There were factors we wanted to include, but didn't, because the data just wasn't available. Even if the data is "well explained" by some set of factors, there is good reason to add more because it can increase the power and accuracy of the model. Granted, I also worked with ecological questions, for which "silver bullet" simplistic explanations are pure fantasy.

Though with simpler models there are fewer variables to control, so it is always a trade-off. I used to do research and audit in the health service, complex systems and many variables!

 

[GoodbyeDave]

I thought I'd try another fantasy classic, The Worm Ouroboros...which I found to be long, boring and unnecessarily complicated.

I prefer Eddison's Zimiamvian trilogy: The Mistress of Mistresses, A Fish Dinner in Memison, and The Mezentian Gate. Sadly, the last volume was never finished, although we do have the plot-outlines for the unfinished chapters and the last chapters were completed. But it is even more complicated! C.S.Lewis and James Stephens loved it, though.

 

[beenherebeforeagain]

I prefer Eddison's Zimiamvian trilogy: The Mistress of Mistresses, A Fish Dinner in Memison, and The Mezentian Gate. Sadly, the last volume was never finished, although we do have the plot-outlines for the unfinished chapters and the last chapters were completed. But it is even more complicated! C.S.Lewis and James Stephens loved it, though.

I don't believe I ever saw those other works; doubt I would have read them if I had. But to each their own!

 

[Revoltingest]

Yes, but again that's the skill of the story-teller. How much, and how depicted, character/violence/etc., is "absolutely necessary?" The author/editor/producer/director/publisher's version is one interpretation. What a reader thinks is another. When the two match up, you've got a hit on your hands.

When young, I read the Lord of the Rings trilogy, which to some is overlong and boring because Tolkien is always off into history and mythology that really doesn't have much to do with the present action. Anyway, because I liked the LOTR, I thought I'd try another fantasy classic, The Worm Ouroboros...which I found to be long, boring and unnecessarily complicated.

Personal preference. Some just love Game of Thrones...to me it's a not very interesting soap opera, as a book series or as a TV series. It was simply written...but here, let's try following roughly 50 main characters, some of whom will check out a few chapters after they are introduced, while others won't be introduced until Book/Season Five...Others will disappear for years, only to suddenly reappear...

Again, what is "necessary and simple" for one may not be for someone else...

I agree.

 

[icehorse]

Do they? My experience learning how to identify various plants makes me wonder about that. Minimalist or simplistic observations won't do it for you, especially if you're trying to identify something from the grass family. A grass will look like a grass.... particularly to non-botanists. Hell, non-botanists call things that are actually sedges or rushes "grasses" even though they are not. That brings up the issue of presumed patterns being erroneous (did you know that birds are reptiles from a phylogenetic and evolutionary standpoint?), mistaken attributions (that moment when that person you identified as a male turned out to be a female), and so on and so forth. But I'm getting rambly now.

And dang you, I wish I lived in the forest. All the forest properties in my area are way over my price range. *sniffles*

(Redwood groves, even "my" small one, are really amazing - they make their own nano-climates.)

I took the OP to be more on the level of subconscious responses than at the level of deep academic studies. But even with deep academic studies it seems that most scientists sort of intuitively look for parsimonious explanations. Think of Einstein's last decades.

 

[icehorse]

Dang you, Icehorse! Dang you for saying most of what I said six times more parsimoniously than I myself said it! For that, you get a "like", but I'd give you ten "likes" if I could.

I was just standing on the shoulders of giants

 

[Quintessence]

Though with simpler models there are fewer variables to control, so it is always a trade-off. I used to do research and audit in the health service, complex systems and many variables!

In the case of the stuff I was doing, the variables were not really things that you could "control" - they were attributes of the species we were looking at. We were looking for what attributes of a species - whether its life history or the habitat it finds itself in - were contributing to what we wanted to predict and then make accurate predictions for species that weren't part of the model design. Fun stuff. Kinda miss it, really.

 

[McBell]

So if your goal is petty inter-religious bickering, parsimony is a useful principle?

It is no fault of mine that so many theists think that god not being needed to explain something means god does not exist.

 

[Sapiens]

Why would you include factors in a model when you don't know whether they have an effect or what sort of effect that might be?

In the sciences, we have statistical tests to tell us how well the factors we've assumed explain the data we have. If the data is well explained by some set of factors, why add others?

I see parsimony in the sciences as just an extension of the principle that we shouldn't assume things without justification.

Often in science the first step is to determine what the important variables are. Usually some variables carry signal and others carry just noise. Parsimony has the added plus of making the signal clearer by removing the noise.