[s0apgun]

Pretty interesting video on word frequency and language by Michael! Check it out.

This is one of the most interesting videos I have seen in a long time!

Absolutely worth taking a look at.

I don't usually take much notice of these recommended videos (my loss, to be sure!) but this was really something else.

Very cool find! ...and oddly inspiring. Makes me want to pick up one of my "advanced" books for a good slog to find some hapax legomenon.

Thanks s0apgun Very interesting. Reminds me of a ted talk I saw a few years ago.

It's really not that miraculous. Math is a collection of abstractions created to represent reality. It's not a coincidence that it does the job, and does it well.

Stansfield123 Wrote:It's really not that miraculous. Math is a collection of abstractions created to represent reality. It's not a coincidence that it does the job, and does it well.

Just because math is a very useful tool designed to understand reality, it doesn't follow that one equation should describe hundreds of completely different systems. It also isn't necessary that the equation should be incredibly simple and easy to remember.

Zipf's Law sort of falls into this category I have given in math called "Spooky Math." I've ran across numerous examples after keeping my eyes open for them. But the first one I ever encountered was Ulam's spiral. Prime's are suppose to be a relatively random phenomena, and yet, you get patterned banding in the spiral and clustering.

yogert909 Wrote:Just because math is a very useful tool designed to understand reality, it doesn't follow that one equation should describe hundreds of completely different systems.

I don't understand the math behind this equation completely, but, based on a superficial look, it probably does follow. My best guess, without actually going through and fully understanding the math, is that a statistician could take a pen and paper, and prove that that equation necessarily applies to all systems defined by a set of parameters that define natural languages, no matter what the particulars.

Here's an analogy:

Let's say you have a die. If you throw it n times, where n is a large number, odds are it will land very close to 1/6*n times on each side. Now let's say you have six buckets, that someone keeps moving around at random, under your 10th story window, while you keep throwing coins out your window, without looking where you're throwing. Odds are, after a while, each bucket will contain 1/6th of all the coins that land in buckets. If it's ten buckets, it will be 1/10th. I could go on and on, listing system after system where the same even/proportional distribution shows up. In fact, as long as we are talking about a system where there's not some other, deliberate order, even distribution is the norm.

Why do you think that is? Is it a coincidence? Or is it because the concepts even and proportional have been created by people, to describe an aspect of reality in general?

It seems to me that these numbers, counting words in various languages, are also just another example of large amounts of discrete pieces of unordered information, being distributed evenly. It's the same principle, in a more complex form. But it's not THAT complex. The field of Statistics, built entirely on this very basic observation of reality, has a lot more complex equations than this one, and they can all be logically deduced from the basic principle.

Stansfield123 Wrote:

yogert909 Wrote:Just because math is a very useful tool designed to understand reality, it doesn't follow that one equation should describe hundreds of completely different systems.

I don't understand the math behind this equation completely, but, based on a superficial look, it probably does follow. My best guess, without actually going through and fully understanding the math, is that a statistician could take a pen and paper, and prove that that equation necessarily applies to all systems defined by a set of parameters that define natural languages, no matter what the particulars.

Here's an analogy:

Let's say you have a die. If you throw it n times, where n is a large number, odds are it will land very close to 1/6*n times on each side. Now let's say you have six buckets, that someone keeps moving around at random, under your 10th story window, while you keep throwing coins out your window, without looking where you're throwing. Odds are, after a while, each bucket will contain 1/6th of all the coins that land in buckets. If it's ten buckets, it will be 1/10th. I could go on and on, listing system after system where the same even/proportional distribution shows up. In fact, as long as we are talking about a system where there's not some other, deliberate order, even distribution is the norm.

Why do you think that is? Is it a coincidence? Or is it because the concepts even and proportional have been created by people, to describe an aspect of reality in general?

It seems to me that these numbers, counting words in various languages, are also just another example of large amounts of discrete pieces of unordered information, being distributed evenly. It's the same principle, in a more complex form. But it's not THAT complex. The field of Statistics, built entirely on this very basic observation of reality, has a lot more complex equations than this one, and they can all be logically deduced from the basic principle.

I guess I was lumping in the link I posed about power law distributions. I'm sorry you find math a equation describing everything ever written so meh, but I find it fascinating. It's even more interesting that the same equation describes the power consumption of cities, walking speed, wealth, crime rate, earthquake magnitudes, ect. It's not very complicated math btw, which is a big part of why it's so interesting.

yogert909 Wrote:I'm sorry you find math a equation describing everything ever written so meh, but I find it fascinating.

Stansfield finds most of life pretty meh. It can actually be quite charming once you get used to it.

Vsauce is the best.

tokyostyle Wrote:

yogert909 Wrote:I'm sorry you find math a equation describing everything ever written so meh, but I find it fascinating.

Stansfield finds most of life pretty meh. It can actually be quite charming once you get used to it.

I never found Stansfield charming, but there was a time when I valued him as a contributor. Stansfield likes to be the contrarian and sometimes he brings up some thought provoking original ideas. Usually I didn't totally agree with what he was saying, but there was occasionally something unorthodox in what he was saying that was useful.

Lately, he hasn't had anything useful to say, so he's only been disagreeing with people simply to pretend that he knows something the rest of us are ignorant about and being rude about it.

PMotte Wrote:Vsauce is the best.

Once you realize he came into his prime well into posting YouTube videos and learn to ignore his early work in favor of his free-form insatiable curiosity videos, it's clear he has a unique ability to connect disparate ideas.

(His early videos were putting editing fart noises onto Mario's jumping and the like. I understand his desire not to disown them, but I do wish he would move them to another sub-channel so that a scrape of his posted videos would be the amazing ones and not the college fart humor ones.)

For people liking the intelligent thread-finding type shows, Connections, its sequels, and "The Day the Universe Changed" are excellent similarly fashioned (and not surprisingly better researched, given the BBC heritage.) I could listen to James Burke forever.

The documentary torrent tracker has all these posted and seeded, last I checked.

https://en.wikipedia.org/wiki/Connections_(TV_series)