Names and identifying details of every student here are changed, as they are everywhere on the blog.
Put a chatbot and a teacher side by side on explaining the instrumental case and the contest is closer than most teachers would like to admit. The machine is patient, available at two in the morning, and will produce forty examples while I am still deciding whether eight is enough.
The comparison everybody runs is the wrong one
Then look at who is still studying Russian in March. Different contest, different winner, and the only one that decides whether any of the explaining mattered.
The tools are not competing on teaching. They are competing on whether you keep going.
What the learners themselves said
There’s a study I keep coming back to because the finding is so undramatic that nobody quotes it, and because the students in it explained the result themselves rather than leaving it to the researchers. University students were given Duolingo to use on their own alongside their language course. They rated it highly. They barely opened it. Asked why, they named other obligations and a need for outside motivation, which the researchers connect directly to there being no teacher attached. Not it was badly designed. Not I stopped learning anything. Just that nothing outside them required it.Neither study is about Russian, and I’m not going to stretch them into being. The mechanism is not language-specific, though, and every teacher of anything recognizes it on sight.
I have never finished a course alone
Sixteen years of teaching, and I’ve started three professional courses on my own. I finished none of them, which is a thing I would rather not put in writing. My last attempt was a course on assessment design, genuinely useful, twelve modules, no cohort and no deadlines. I got to module four in eight days and then to module four for the next eleven months. I’ll do it this weekend is a sentence I said to nobody, roughly forty times. Nothing about it was too hard. Nobody was ever going to ask, and that turned out to be the whole difference. What I did finish had four other people in it and a call every Wednesday at seven. I was not more motivated that year. I was more observed.Nobody drops out of a class they have to say they are not coming to.
Where each one is genuinely better
Self-study with tools wins on volume, patience, and cost. It will drill you on aspect until you are sick of the whole category, never once sigh, and cost less in a year than two weeks of tutoring.
Most adults have plenty of access to explanation and almost none to being expected.
What each side is actually selling
An honest version of the comparison is not one against the other.
Every app on the market sells explanation, because explanation scales and photographs well. Nobody advertises expectation, partly because it sounds like an imposition and partly because it is difficult to deliver at a distance without a human on the other end of it.
Which is roughly why the honest recommendation ends up sounding like an advertisement for classes even when the person making it would rather it did not.
The year I moved a class by ninety minutes
Some years ago I shifted an adult group from Tuesday at seven to Tuesday at eight thirty, for reasons that were entirely sensible and had to do with my own week. Four of the seven stopped coming within a month. Not one of them said the new time was impossible. Two said they would catch up and did not. One wrote that she was just going to do it on her own for a while, which I’ve since learned to hear as a goodbye. Nothing about the teaching changed. I didn’t become worse at explaining aspect between one term and the next. What changed was that seven o’clock had been welded to something else in their evening, and eight thirty was floating loose, and a floating appointment is one you can decide about. Our conversation school keeps its slots fixed for exactly this reason, and I used to think that was inflexibility. I had spent years thinking my value was in the forty minutes. A good part of it was in the fact that the forty minutes were always in the same place.A fixed time is not an administrative detail. It is most of what a class does.
Making the tools carry some of the structure
If classes aren’t an option, the workaround is to build the expectation artificially, which is harder than it sounds, less effective than the real thing, and considerably better than relying on wanting to. A streak is not accountability. It is a number that goes up when you do the easy thing, designed to make quitting feel expensive rather than to make showing up feel obligatory, and those are different mechanisms with different failure points.What it looks like when it works
Tuesday group, third month
Student I did nothing this week. Genuinely nothing.
Me Fine. Read the first line of the homework out loud and we’ll go from there.
Student …I did do that bit, actually. On Sunday. Because of tonight.
Nobody would have checked. She knew somebody would ask.
She had not planned to study. She had planned not to arrive empty-handed.
None of that is a triumph of pedagogy. It is Sunday evening and a small social cost. Because of tonight did more for her Russian that month than anything I explained. The chatbot is still the better explainer, which I have argued at length elsewhere and will keep arguing. It’s simply not in the business of expecting you, and nothing in its design suggests it ever will be. At Palme School we teach online, in groups of up to eight, two forty-minute classes a week, and the first two lessons are free. You can book those here, and there are printable materials for learners in our resource library.Frequently asked questions
For explanation and drilling, close enough that the difference rarely decides anything. For finishing, no, and the research on attrition points at accountability rather than teaching quality.
Plenty of people reach a working level alone, and the ones who do usually have some external structure they built themselves: a partner who asks did you do it, a group, an exam date on a calendar.
Tools daily for volume, a class weekly for the appointment. Bring to the class whatever the tools could not settle, and let the class supply the reason you opened them at all.
Because month three is when novelty ends and nothing external has replaced it. That is the ordinary pattern, not a fact about you, and the fix is structural rather than motivational.
Sources
- García Botero, G., & Questier, F. (2016). Self-directed language learning: Students’ experiences with Duolingo. ERIC ED572176. files.eric.ed.gov
- Hwang, H.-B., Coss, M. D., Loewen, S., & Tagarelli, K. M. (2024). Acceptance and engagement patterns of mobile-assisted language learning among non-conventional adult L2 learners: A survival analysis. Studies in Second Language Acquisition, 46(4), 969–995. doi.org
Palme School: Russian for bilingual children · Conversation school · Free materials
Lily Rozberg has been teaching for sixteen years, first English in Russia and now Russian to bilingual children across the US and Canada. She writes for Palme School about heritage language, reading, and the things report cards do not measure.
