Why Don't Language Apps and Tutors Make You Fluent?
Almost every adult who tries to learn a second language fails.
In one study of adults using a commercial language software program, 99 percent quit before finishing even 10 hours of a 200-hour course (Nielson, 2011). Most college students fail to go beyond beginning language classes, and almost no one makes it to the end of self-study courses (McQuillan, 2019).
Why do language apps, tutors, and grammar courses all fail? The reason is that they do not give you the one thing you need to be successful, what linguists call "comprehensible input."
Comprehensible input is language that you can mostly understand, but that still has something new for you to learn (Krashen, 1982). When you understand the message you listen to or read in another language, your brain is able to pick up all of the different parts of the language itself. It does this naturally and unconsciously, without you even realizing it.
What does the research say you need to learn English?
Your brain has two separate types of language knowledge. One type is conscious knowledge of the "rules" of grammar, pronunciation, punctuation, and so forth. This is knowledge ABOUT the language or how to describe the way the language works.
"A noun is a person, place, or thing" and "the third person singular verb usually ends in -s" are examples of conscious knowledge of language. We usually get this kind of knowledge through slow, painful studying.
The second kind of knowledge you have is unconscious. Unconscious knowledge is knowledge you can't always describe, but that you use every day. It is knowledge of the language itself, what you use 99 percent of the time when you speak or listen. It's automatic, fast, and effortless. To be fluent in a language means to have large amounts of unconscious knowledge of that language.
These two kinds of knowledge are completely separate. That means that the rules you study do not become part of your unconscious knowledge (Krashen, 1982). You can have both, of course. You can use the third-person singular -s rule correctly when you speak, and you can describe the rule on an examination.
You can also have one without the other. You can know the rules for the test but not be able to use them in a conversation.
You learn about a language (conscious knowledge) by studying.
You learn a language (unconscious knowledge) by getting large amounts of comprehensible input through listening and reading (Krashen, 1985).
The difference between "studied" rules and unconscious ability explains why most language courses fail. Any product that uses drilling, memorizing, or forced speaking is working on the conscious system or on output. Those products will not make you fluent because they are not giving you comprehensible input, which is the only thing that improves language proficiency.
How do the popular methods compare with the research?
These seven products each give you very little comprehensible input. The table shows what each one gives you instead.
| The popular method | What it really gives you | Why it fails |
|---|---|---|
| Duolingo, Babbel, and other apps | Short, conscious drills on grammar and vocabulary, with points and streaks, which the companies call "gamified" | There is very little comprehensible input. You are being drilled and tested on small bits of English, rather than getting large amounts of English you can understand (Krashen, 1982). |
| Rosetta Stone and similar software | Picture-to-word matching with a few short phrases | There is a small amount of comprehensible input, and lots of testing. In one study, 99 percent of users quit before 10 hours (Nielson, 2011). |
| Pimsleur and other repeat-after-me audio courses | Small amounts of comprehensible input followed by a lot of "now you say it," with constant translation | They ask you to speak before you have heard enough comprehensible input. Speaking does not give your brain new language to learn from (Krashen, 1998). |
| Conversation tutors (iTalki, Cambly, language exchanges) | An audience: you speak, and the tutor corrects what comes out | You can't say what you haven't yet learned, and your own talking gives your brain no comprehensible input (Krashen, 1998). |
| Vocabulary flash cards (Anki, Quizlet, paper cards) | Isolated word-and-translation pairs to memorize | Studies show this kind of memorization does not improve your ability to understand what you read or hear (e.g. Tuinman & Brady, 1974, Pany & Jenkins, 1978, Hudson, 1982, Pany, Jenkins, & Schreck, 1982, and Jahangard, Moinzadeh, & Karimi, 2011). |
| Grammar workbooks and rule drills | Rules to memorize and tests on those rules | Conscious rule knowledge does not become unconscious ability, and real conversation is too fast for rule-checking (Krashen, 1982). |
| Movies, news, and YouTube without help | English that is far above your level, with nothing to make it easier | You need to understand most of what you hear before you learn from it (Krashen, 1985), and materials made for native speakers are too difficult for intermediate learners. |
Do conversation tutors help at all?
A tutor helps you if he or she gives you comprehensible input to listen to. If the tutor speaks clearly and at your level about interesting things, you get that input during your lesson. If you do most of the talking while the tutor corrects you, you get very little comprehensible input, because your own speaking does not teach your brain anything new (Krashen, 1998).
The one question to ask about any product or class is: How many minutes of English that you mostly understand does it give you?
What should you use instead?
Listen to and read lots of English you mostly understand on topics you enjoy and already know something about. Studies by Krashen and dozens of other applied linguists over the past 40 years all support this approach. Every hour you spend on comprehensible input instead of exercises and memorization improves your English.
Take whatever app or course you use now and count the minutes of real English stories or conversations you mostly understood. If the answer is close to zero, the product gave your brain almost nothing to learn from.
Dr. Lucy Tse and I designed Unlimited English to supply the comprehensible input these products do not. We give you more than 1,800 audio lessons of clear English you can understand, each designed with new language you're ready to learn, plus a complete Learning Guide for every one of them. The full science is on our research page.
~ Dr. Jeff McQuillan, from beautiful Los Angeles, California
Selected research
- Hudson, T. (1982). The effects of induced schemata on the "short circuit" in L2 reading: Non-decoding factors in L2 reading performance. Language Learning, 32(1), 1-31.
- Jahangard, A., Moinzadeh, A., & Karimi, A. (2011). The effect of grouping and presenting selected TOEFL words on the vocabulary learning of EFL learners. Iranian EFL Journal, 7(2), 220-235.
- Krashen, S. D. (1982). Principles and practice in second language acquisition. Pergamon.
- Krashen, S. D. (1985). The input hypothesis: Issues and implications. Longman.
- Krashen, S. D. (1998). Comprehensible output? System, 26(2), 175-182.
- McQuillan, J. (2019). And then there were none? Measuring the success of commercial language courses. International Journal of Foreign Language Teaching, 8(1), 2-5.
- Nielson, K. B. (2011). Self-study with language learning software in the workplace: What happens? Language Learning & Technology, 15(3), 110-129.
- Pany, D., & Jenkins, J. R. (1978). Learning word meanings: A comparison of instructional procedures. Learning Disability Quarterly, 1(2), 21-32.
- Pany, D., Jenkins, J. R., & Schreck, J. (1982). Vocabulary instruction: Effects on word knowledge and reading comprehension. Learning Disability Quarterly, 5(3), 202-215.
- Tuinman, J. J., & Brady, M. E. (1974). How does vocabulary account for variance on reading comprehension tests? A preliminary instructional analysis. In P. Nacke (Ed.), Interaction: Research and practice for college-adult reading (pp. 176-184). National Reading Conference.