Difference between revisions of "Duolingo"
(6 intermediate revisions by the same user not shown) | |||
Line 8: | Line 8: | ||
* I've noticed that sometimes, if I get a sentence almost correct but miss one phrase, that phrase will show up isolated as a phrase in a following screen. I'm not sure if this is done on purpose or if I just happen to get the most difficult phrases wrong. | * I've noticed that sometimes, if I get a sentence almost correct but miss one phrase, that phrase will show up isolated as a phrase in a following screen. I'm not sure if this is done on purpose or if I just happen to get the most difficult phrases wrong. | ||
* Another pattern I've noticed is that if I get a Russian -> English translation problem wrong, then that Russian sentence will often show up in a bit as an audio -> Russian text problem. I guess this is to build familiarity with the problem I missed. | * Another pattern I've noticed is that if I get a Russian -> English translation problem wrong, then that Russian sentence will often show up in a bit as an audio -> Russian text problem. I guess this is to build familiarity with the problem I missed. | ||
+ | * one clever phrase i encountered was "одно окно и две двери" (one window and two doors). the number and object are phonetically similar, and it seems like they are doing this to combat interference. once you learn that window is "okno", then you later learn the one is "odno", you might start confusing window and one, but by presenting them together in the same phrase, you are forced to distinguish the two, and it might even highlight for you that "huh, there are two words that sound very similar". | ||
+ | ** another similar one was "Стол стоял около окна", which means "The table was standing near the window". | ||
+ | * on problems where I need a phrase like "grandmother and grandfather", there are sometimes two endings, like -a and -u. i've noticed that duolingo actually does the -a and -u endings on ''both'' the grandmother word and grandfather word, so that you can't just guess the answer by selecting the ending that appears twice. | ||
+ | * something i like about the reminder emails is that they get sent around the time you did your lessons on the previous day. so it automatically figures out your routine and tries to fit in with it. | ||
Above, I am not including "gamification"/achievements stuff that Duolingo does (e.g. badges, scores, leader board). I think these are dumb and I don't think I benefit from them (other than the streaks counter, which does motivate me to never miss a day). | Above, I am not including "gamification"/achievements stuff that Duolingo does (e.g. badges, scores, leader board). I think these are dumb and I don't think I benefit from them (other than the streaks counter, which does motivate me to never miss a day). | ||
Line 13: | Line 17: | ||
I don't like how the later lessons in duolingo are much longer/have way more components but count for the same amount of experience. when i am feeling tired i feel tempted to just do some easy lessons to keep my streak, instead of doing more of the new content (which i feel is a better use of my time). | I don't like how the later lessons in duolingo are much longer/have way more components but count for the same amount of experience. when i am feeling tired i feel tempted to just do some easy lessons to keep my streak, instead of doing more of the new content (which i feel is a better use of my time). | ||
− | I think it's weird that duolingo penalizes mistakes so much (you get less XP if you get questions wrong). i often feel that i am learning the most when i make mistakes and have to internalize a new subtlety of the grammar or word choice or whatever. i feel like if i get something wrong, then get it right the next time (after some other questions in between) that's good evidence i have learned something new, and that behavior should be rewarded. | + | I think it's weird that duolingo penalizes mistakes so much (you get less XP if you get questions wrong). i often feel that i am learning the most when i make mistakes and have to internalize a new subtlety of the grammar or word choice or whatever. i feel like if i get something wrong, then get it right the next time (after some other questions in between) that's good evidence i have learned something new, and that behavior should be rewarded ([[learning vs competence]]). |
− | [[Duolingo does repetition at the lesson level]] | + | ==See also== |
+ | |||
+ | * [[Duolingo does repetition at the lesson level]] | ||
+ | * [[Duolingo for math]] | ||
+ | |||
+ | ==What links here== | ||
+ | |||
+ | {{Special:WhatLinksHere/{{FULLPAGENAME}} | hideredirects=1}} | ||
[[Category:Learning]] | [[Category:Learning]] |
Latest revision as of 23:48, 13 August 2021
Some learning tricks/techniques that Duolingo uses:
- the initial levels of a lesson are mainly recognition-based, so cognitively easier; they build up confidence and familiarity with the material without being too difficult
- later on, you need to do stuff like type in the words using cyrillic letters. This is harder, but by this time you've encountered the word multiple times so it's not too hard.
- in many contexts, you can hover over words to get their definitions, so you're never totally lost.
- email notifs
- there is no explicit instruction of grammar rules -- you just pick it up by trying to understand more and more complicated sentences. This makes it similar to "tutorial modes" in some video games, or puzzle games without explicit instruction (like The Witness).
- I've noticed that sometimes, if I get a sentence almost correct but miss one phrase, that phrase will show up isolated as a phrase in a following screen. I'm not sure if this is done on purpose or if I just happen to get the most difficult phrases wrong.
- Another pattern I've noticed is that if I get a Russian -> English translation problem wrong, then that Russian sentence will often show up in a bit as an audio -> Russian text problem. I guess this is to build familiarity with the problem I missed.
- one clever phrase i encountered was "одно окно и две двери" (one window and two doors). the number and object are phonetically similar, and it seems like they are doing this to combat interference. once you learn that window is "okno", then you later learn the one is "odno", you might start confusing window and one, but by presenting them together in the same phrase, you are forced to distinguish the two, and it might even highlight for you that "huh, there are two words that sound very similar".
- another similar one was "Стол стоял около окна", which means "The table was standing near the window".
- on problems where I need a phrase like "grandmother and grandfather", there are sometimes two endings, like -a and -u. i've noticed that duolingo actually does the -a and -u endings on both the grandmother word and grandfather word, so that you can't just guess the answer by selecting the ending that appears twice.
- something i like about the reminder emails is that they get sent around the time you did your lessons on the previous day. so it automatically figures out your routine and tries to fit in with it.
Above, I am not including "gamification"/achievements stuff that Duolingo does (e.g. badges, scores, leader board). I think these are dumb and I don't think I benefit from them (other than the streaks counter, which does motivate me to never miss a day).
I don't like how the later lessons in duolingo are much longer/have way more components but count for the same amount of experience. when i am feeling tired i feel tempted to just do some easy lessons to keep my streak, instead of doing more of the new content (which i feel is a better use of my time).
I think it's weird that duolingo penalizes mistakes so much (you get less XP if you get questions wrong). i often feel that i am learning the most when i make mistakes and have to internalize a new subtlety of the grammar or word choice or whatever. i feel like if i get something wrong, then get it right the next time (after some other questions in between) that's good evidence i have learned something new, and that behavior should be rewarded (learning vs competence).
See also
What links here
- List of techniques for managing working memory in explanations (← links)
- List of AI safety projects I could work on (← links)
- Duolingo does repetition at the lesson level (← links)
- Repetition granularity (← links)
- Duolingo for math (← links)
- Tinkering in math requires loading the situation into working memory (← links)