Language Learning Is Broken | Kumar Thangudu

I’ve learned a lot about languages over the years. I’ve independently studied the following languages and achieved varying levels of fluency. By competency:

Spanish French German Swahili Russian Hindi Farsi Arabic Afrikaans Polish

I estimate that there are about 250 players in the space ranging from innovators like Duolingo and Verbling to overpriced behemoths like Pimsleur and Rosetta Stone.

Language learning software is broken in terms of reaching its ‘holy grail.’ I believe the holy grail of language learning is the ability to learn a language outside the country that speaks it, to a level of fluency that puts the user within 30 to 45 days of advanced reading, speaking, and vocal comprehension fluency once they are in the country. 

This is a problem worth solving. I read somewhere recently(if you find a link please message me) that the demand to learn English is so great in China that if they hired every US college student to teach English, they would not fulfill the demand.

There are some fundamental reasons why many of these software platforms fail at the attempt to reach the holy grail.

Flashcard systems don’t work. Whether it’s words or phrases flashcard systems don’t work. Many people who are learning a language need to see two words together. For example, Estudiante de Postgrado denotes a grad student in Spanish. Even if a user sees these words in a flash card system with a picture and a translation, their retention and later usage of the word is unlikely.  Audio systems don’t work. Language learners often want to see, hear, and understand how the syllables of a word are put together. Teaching systems that rely on audio break this mental image. This puts some language learners at a particular disadvantage. For example if I was learning Polish, the language is extremely phonetic but requires more complex tongue movements. Without seeing how the word is spelled, the user is at risk of slaughtering the already difficult pronunciation. A simple example: Zapomnij o tym (Forget it) (Zapom-nee- Oh- Tim). Even with the pronunciation key, there are elements of it that are lost in translation. Language learning schools ‘kind of’ work. I’d probably only take this option if it was learning English or German. If a user has to go live in the country and take classes with a large number of students that are also learning the language it’s often too time and capital intensive for most people looking to learn a language. Costly systems. Costly systems aren’t priced for the users who need them most. Worst of all they are usually excessively boring. Rosetta Stone, Pimsleur, and Michel Thomas are prime examples.

My friend and I built a simplistic prototype(Version 0.1- try this link if it doesn’t work) that uses a popular vice documentary about  Mormons in Mexico and subtitles to give users viable examples of some of the 1000 most common words in English. If you are building something awesome in the language learning space, please feel free to message me about it.