WILearned
patreon
Bonus for Patrons: "How to Create a Language" video comments
🕑 Added 2019-09-23 11:59:38 +0000 UTCHey everyone,
This video took a lot out of me. As usual, I waaay over-researched for this as usual - have about 50 pages of unused notes left so I'll for sure need to make another video on this topic from a different perspective at some point.
There's a lot of interesting neuroscience that I wanted to include, but just giving the index, icon, symbol progression story and explaining the implications of Daniel Everett's work with the Piraha ended up taking much longer than expected. For those of you interested in what neuroscience says about language, this lecture is very good: https://youtu.be/0aFMzANQd7A
I looked at four other books:
Steven Pinker's "The Language Instinct"
Noam Chomsky & Robert Berwick's "Why Only Us"
Michael Corballis' "The Recursive Mind"
Eduardo Kohn's "How Forests Think"
The most intriguing was "How Forests Think," but it gets excessively dense at points. He talks a lot about the sign progression theory of language / semiotics and gives an interesting picture of the interspecies communication that occurs in the Amazon forest(s).
I think most people would enjoy Corballis' book more than Pinker's (in terms of readability, that is - they have differing arguments) and would enjoy Chomsky & Berwick's the least.
Here's a quote from the book: "There is a workspace, which has access to the lexicon of atomic elements and contains any new object that is constructed. To carry a computation forward, an element X is selected from the workspace, and then a second element Y is selected. X and Y can be two distinct elements in the workspace, as when read and books are merged to form the syntactic object underlying the phrase read books."
I think Thom Scott-Philips was right when he said "linguists have physics envy." You can't codify language because it's a gestalt. He points out in a lecture that in the "Oxford Handbook of Language Evolutin," there are 213 entries for syntax, 145 entries for semantics, and only 8 entries for pragmatics.
To that point, here's a clip from my Skype talk with Daniel Everett about why you can't learn a language like some other academic discipline, you have to learn the culture: https://www.youtube.com/watch?v=kR48Req8iCQ
▪︎One point that got cut out of the final video was the one made in Deb Roy's TED talk ( https://youtu.be/RE4ce4mexrU ) He points out that, based on 3 years of audio visual data of his son's upbringing, caregiver's utterances get more and more simple as the child starts to learn a word. Then, after learning the word, complexity goes back up.
I thought this was a good point to show that language learning is a very difficult process, but a coordinated effort. This would act as a counterpoint to the "poverty of the stimulus theory." ( Essentially the theory that linguistic input received by children is insufficient to explain their knowledge of a first language, suggesting language is innate) It takes two years of being bathed in language to be able to say sentences, so it's surprising to me that this theory got any traction.
Anyhow, I cut this out because it interrupted the flow as I don't get into "poverty of the stimulus" in this video.
▪︎Lately I've been putting more effort into the thumbnails because, for the most part, if the thumbnail isn't good, the video doesn't get near as many views. I tried several different thumbnails and finally ended up with the one you see on the video. For those of you interested in youtube or just my posting process, here's a quick breakdown
-First I started with the Piraha kid "How Start Talk" (tried two different fonts). The thinking was to entice people to think "Why did he phrase it like that?" (And would add to my point about being able to say a lot with minimal/bad syntax)
Also I thought the emoji sequence was a good way to communicate the same phrase. "No speak" to "Speak," "Why/How?" Then again the initial reaction would probably just be "OK so he's talking about emojis in this video?"
I ended up scrapping this one as the balance didn't feel quite right and I thought using the Piraha kid would miscommunicate that the video was about learning English as a second language or something like that.
-Then I tried the baby, which I thought was cute and kinda funny as he (she?) looks pissed off. But, thought that would give the impression that I was going to talk about how babies learn to speak.
-I actually initially posted with the one with Daniel Everett looking to the right with "How Begin Talk?" I thought it looked pretty decent, posted the video and finally went to sleep.
Then I woke up and looked at my phone, and realized when I look at the thumbnail when it's very small, it sorta looks ...bible-ish. Something about the Tratello Font I was using made it look (to me) like the video was going to be more about the Tower of Babel.
- I started working on some more ideas then I thought that while the bad grammar title might make people curious, it's usually the clearest titles that are most grabbing. At least for me, when I'm browsing around there's plenty of videos on a topic, I end up going for the one with the clearest title that seems to be in line with what I'm looking for. I might think "what does he mean by that?" if there's a peculiar title, but I won't think "I should watch that 20 minute video to see what he's thinking..." When it's more clear I just think "That's what I'm looking for." *Click* So, I went with "How to Create a Language" (which may be confused with how to create a conlang, but oh welll).
I also wanted to fit in some example of icons in the thumbnail, and thought it looked well balanced to have emoji panel there. Initially I had 1. 2. on there (as in "Step 1, Step 2") but thought that looked busy. I thought the arrow would communicate that I'm talking about some sort of progression ...but then again it sorta looks like I'm saying "people started to speak and then developed icons" but oh well. This is the consequence of overthinking everything all the time.
Bad news is the video isn't performing as well compared to the usual:
But the good news is there's a 26% increase in view duration which perhaps means the editing/script was more engaging. I'm still not entirely sure if the great performance on the Doublespeak and Sleep Position videos was due to youtube algorithm god's blessing or reddit. In any case, Daniel Everett said I represented his work quite well so I'm very happy with that. And of course, I hope you guys like the video 😀
P.S. I just came across this clip today: https://youtu.be/v7Spzjh9QgA
It demonstrates more clearly how a lack of recursion works in the language.
At 00:47 Everett points at two different figurines and says:
"He killed a jaguar." (pointing at the first figure)
"This one killed a bird." (pointing at the second figure)
"Afterwards, he fell down." (pointing at first figure again)
Then he takes the figures away and says "Who fell down?"
The Piraha man says "Cat, fell down."
Everett then explains to another English speaking person there: "I said 'Who fell?' He said 'Cat, he fell.' - The guy who killed the jaguar."
