Hi everybody! May I please ask your opinion on the following idea.
New users come to a system like LingQ. They already know s-thing in the target language (the typical case for a learner of English). They are asked, in a polite English manner, of course, to read through a short enough text and click all the unknown words while they are reading it. (They may and may not be lingQuing or translating the unknown words while doing this – it does not matter). When the user is done, the system says s-thing like:
“ I estimate that you know 3300 ±100 different words in English”.
Could it be useful, and if yes, how, where?
In the case of LingQ, I could have imagined the usefulness in the following. Today, when a new user having a knowledge in a language comes to lingQ, the system starts her or his known words’ counts from zero. With the above estimate, the counts could have been started from the estimate, like from 3300 ± 100. And henceforth the count would be only increasing farther (with, optionally, the value of the uncertainty ± 100 being constantly decreased). This could be felt more stimulating and true by the learner.
Or, with no any relation to LingQ, this could make a basis for the quick but objective estimation of a learner’s (passive) vocabulary.
What do you think? Could you please suggest other uses? Thanks to all.
Ilya, great suggestion! I’ve been studying French for 2.5 years in academia as well as independently, and have a high recognition vocabulary, particularly since I read pieces of texts beyond my “comprehension level” i.e. Kristeva, Foucault, Fanon, etc., but there was no way for me to tell the system so, if you will.
What I did do, however, and what I found tremendously helpful, was to start at beginner level to get those “known” words into my “vocabulary bank”. The system not knowing my level did work to my advantage because I backtracked and got in some good review time, where I notated fundamental structures that previously eluded me. Oddly enough, and digging into my comfort zone, I picked up some great words along the way in Beginner 2. ^^
Yet for those newcomers with a solid foundation and with little to no doubts about their level of knowledge in a particular language, it would be great to have a meter to register one’s estimated value of known words.
It was never a problem for me to start with 0 known words here at LingQ. It is logical in my opinion that LingQ couldn’t know which words are known and unknown for me, and at the beginning the number of known words increased very fast. I believe that this isn’t a huge problem for learners. The number of known words will increase very fast if they choose lessons from their level.
@Yvette: I started English with the beginner lessons two. After 30 years of about no contact with English they gave me a great start, and self confidence.
My algorithm, which allows to quickly estimate the number of words a learner knows in a foreign language, says what it says already a couple of days, but what it says still surprises me.
It says in particular the following. Let you and me were given to read a page in English. You have encountered 10 while I have encountered 20 unknown words in it. It is clear that you, obviously, know more English words than me, but how much more? Please make a guess before reading farther.
The algorithm says you know 16 times ! more English words.
This surprised me. What would you say?
Though all the dependencies obtained seem to me qualitatively correct, I had not expected it in the quantitative terms. And I simply am afraid to report you what the algorithm predicts for the case of the Russian language.
I don’t know what to compare these results with. If someone has heard or read about somethingrelevant, a way to somehow estimate the number of words a learner already knows in a foreign language, please let me know.
And I still lack motivation to implement it, so if you’ll see good uses for it, consider it a second call
Alex, what does you comment mean?
