Postby Cainntear » Tue Jan 15, 2019 12:12 am
I've said a lot about Duolingo in the past, so there's a fair chance I'll be repeating myself here.
First up, it's great that Duolingo encourages people to keep at it, but it's a squandered opportunity. Once you've got the habit, you should be upping the complexity.
Von Ahn says that upping the learning means upping the drop-out rates... in that case he's doing it wrong -- both teaching and gamification.
Teaching:
"Fun" is the name we give to the sensation of having our brain stimulated. Learning stimulates the brain. Therefore learning is fun -- what is not fun is not learning. Duolingo could use their massive user base to experiment and identify the "fun" point, identifying the really good learning.
Gamification:
The single biggest mistake Duolingo has suffered since day one is that they've totally misunderstood XP. In Duolingo, every correct answer is worth 1 xp, from matching milk to Milch in a multiple choice question to translating My two cats' four kittens have taken the aubergine to the dentist (not a real DL sentence!!) in a heavily case-marked language. Of course people aren't going to appreciate difficult tasks if they don't get any reward or recognition for completing them!
In a role-playing game (RPG), you'll get 1 XP for swatting a fly, and 1000 XP for killing a dragon, which motivates the players to take on harder and harder challenges. Would you risk your life fighting a dragon for 1 XP when you could get them the easy way?
A related problem with XP is that they've copied the idea from RPGs that "levels" need progressively more XP the higher you get... but that's done explicitly to encourage taking on harder, higher-risk fights, which is not an option in Duolingo.
They compound the problem by bundling questions together, and only giving you XP when you get all questions right (previously all but 3, but the hearts have been gone for a while). This means that the difficult questions not only aren't getting a fair reward, but end up being an obstacle that stops you getting rewarded for what you did do right. It triggers problems of "loss aversion", so yeah, of course people are going to dislike difficult tasks (you know, the ones that lead to actual learning).
It wouldn't be that hard to do right. Reward players for what they get right. Reward players for taking on bigger challenges. Present things like stories and dialogues as a big challenge that you're training for -- a goal, a target; not something that just gets in the way. Hell, even make them boss fights to get you the final crown or something.
But no. It's easier to just Skinner-box the thing to death and keep the customers coming back and working through a bunch of multiple choice questions and fridge magnet exercises.
And I don't think Duolingo have opened up the market for language apps either, because by running for years on a $0.00 price tag funded by immense amounts of venture capital, they've killed the value of language apps in the public imagination. There have been people trying to do interesting things with language technology, but who's really got anywhere? Babbel is actually pretty old-school when you look at it -- it's just a slick version of the sort of language course that's been around since the 90s. Anyone who has tried to make an incremental improvement on Duolingo is killed by being too small and invisible, and not being able to charge enough money to stay alive.
As I say, I reckon Duolingo is a squandered opportunity that has actually hampered innovation.
[Edit: fixed unclosed [I] tag.]
Last edited by
Cainntear on Tue Jan 22, 2019 5:35 pm, edited 1 time in total.
14 x