Is there a relatively objective way of telling how much you know?

General discussion about learning languages
Tillumadoguenirurm
Orange Belt
Posts: 193
Joined: Fri May 06, 2016 3:07 pm
Languages: English
x 235

Re: Is there a relatively objective way of telling how much you know?

Postby Tillumadoguenirurm » Wed Sep 13, 2017 1:46 pm

mcthulhu wrote:You could try to find out where you are with some spot checks to see where words you know and don't know in a text are ranked in a word frequency list. There are others, I think but the Leipzig corpora collection can also serve as one. It has a large French corpus at http://corpora.uni-leipzig.de/en?corpus ... mixed_2012, with word frequency rank and frequency class associated with each word. For instance, arbre is ranked 6,102 and ficelle is ranked 31,847, so if you knew arbre but not ficelle that would suggest that your vocabulary size is somewhere in that range. (This is a wide range, I know - I just picked a couple of words at random. You would need to do more than just two searches, anyway.)

I guess it's theoretically possible that a person might know ficelle and not arbre, but that seems less likely...


I'm only on my phone/tab, and I couldn't find anything that explained the terms on that site. What does frequency class or rank mean?

Word: .....
Number of occurrences: 42 
Rank: 367,046 
Frequency class: 21
0 x

mcthulhu
Orange Belt
Posts: 228
Joined: Sun Feb 26, 2017 4:01 pm
Languages: English (native); strong reading skills - Russian, Spanish, French, Italian, German, Serbo-Croatian, Macedonian, Bulgarian, Slovene, Farsi; fair reading skills - Polish, Czech, Dutch, Esperanto, Portuguese; beginner/rusty - Swedish, Norwegian, Danish
x 590

Re: Is there a relatively objective way of telling how much you know?

Postby mcthulhu » Wed Sep 13, 2017 9:33 pm

The rank should be the word's absolute position in a corpus's frequency list; "et" has rank 3 in the French corpus, and in the German corpus, "der" has rank 1, and "und" has rank 3. You'd expect those to be around the top.

The frequency class, in turn, indicates the relative frequency of the word, compared to the most frequent word in the corpus; the higher the class, the less frequent the word. Here's the explanation from a paper on the structure of the database:

"The basic structure of entries in the corpus database
includes information on the absolute word frequency for
each entry (i. e. each inflected word form or each identified
phrase like the proper name Helmut Kohl). Additional
frequency class is calculated based on a logarithmic
scale relative to the most frequent word in the corpus.
For the English corpus, the most frequent word,
the, has frequency class 0, while an entry like Acropolis
with an absolute frequency of 20 belongs to frequency
class 18, as the occurs approx. 2^18 times more often."

This paper is at https://www.researchgate.net/profile/Ch ... 000000.pdf.
1 x

Finny
Green Belt
Posts: 260
Joined: Sun May 22, 2016 3:01 am
Languages: English...
x 582

Re: Is there a relatively objective way of telling how much you know?

Postby Finny » Thu Sep 14, 2017 3:09 am

I like this question, although I can't answer it. :D I find one decent way if you've learned (or are learning) multiple languages is to compare them to each other.

For Spanish, I've never CEFR-tested it, but I'd estimate I'm at the C1/C2 level for speaking/writing, C2 listening, reading (at least for a Mexican-American Spanish blend...they'd drop somewhat for variations I'm less familiar with due to accent, etc). For French, I'd estimate B2/C1 speaking, C1/C2 listening, C2 reading, and B1/B2 writing with spell check). However, I don't really like putting CEFR letters to my French because I'm not nearly as sure of where things fall compared to with Spanish, where I have much more self-efficacy. It's much faster to think of what I can do in Spanish and refer back in French.

In French, I can follow the radio almost as well as I can in Spanish, I can read almost as quickly, but do come across more words per page that I don't know than in Spanish (where it's rare unless I'm reading a specialized topic). When speaking--which is pretty much just with my kids--I don't run into grammar roadblocks very often, but it does still happen much more than in Spanish, and it's more dramatic. For example, the other night, I wanted to say "tu pourras choisir un livre quand tu te seras brossée les dents"...I knew there was a way to construct the "when you have brushed" section, but couldn't figure it out on the fly and had to look it up later...I said it incorrectly at the time. Later I of course realized I could have said "apres s'etre brossée", which would have gotten rid of the need to know that tense. Similarly, I run into a lot more vocab roadblocks in French than in Spanish, because I simply haven't read nearly as much in French yet. I can write pretty much anything I can say in both languages, but my French writing is limited by my smaller vocab, shakier grammar, and much poorer spelling. So I can use the language with my kids and they're learning it very well, but I still become aware many times a day on French days of what I don't know yet. In Spanish, that's a lot rarer, besides some niggling grammatical doubts (e.g., "molestarlo" vs "molestarle") and the occasional word I don't know but reach for in a sentence (e.g., "horquilla", "vincha", and other things that have to do with my daughter's hair).

I'm not sure if that helped anyone, but to sum it up, instead of trying to objectively rate yourself, perhaps consider comparing your abilities to another L you have. I suppose this would work with your L1 too. I remember when learning Spanish, I noticed a big shift once I could speak fluidly due to building up enough grammar. It didn't matter that I didn't have all the words I wanted; the fact that when I looked them up, I'd know how to phrase them naturally made much, much, much more of a difference. I look forward to reaching that point in French; I still need more internalization of which verbs trigger the subjunctive, though, as well as when it's audible (e.g., "Je suis content que tu viennes aujourd'hui") vs inaudible (e.g., "Je suis content que tu m'ecoute"). Oh, and things like the de/du/d' partitive, and how the future tense is used in situations where you'd just use the present tense in Spanish, and...it goes on.
4 x

Cavesa
Black Belt - 4th Dan
Posts: 4960
Joined: Mon Jul 20, 2015 9:46 am
Languages: Czech (N), French (C2) English (C1), Italian (C1), Spanish, German (C1)
x 17566

Re: Is there a relatively objective way of telling how much you know?

Postby Cavesa » Thu Sep 14, 2017 6:38 am

Finny wrote:I'm not sure if that helped anyone, but to sum it up, instead of trying to objectively rate yourself, perhaps consider comparing your abilities to another L you have. I suppose this would work with your L1 too. I remember when learning Spanish, I noticed a big shift once I could speak fluidly due to building up enough grammar. It didn't matter that I didn't have all the words I wanted; the fact that when I looked them up, I'd know how to phrase them naturally made much, much, much more of a difference. I look forward to reaching that point in French; I still need more internalization of which verbs trigger the subjunctive, though, as well as when it's audible (e.g., "Je suis content que tu viennes aujourd'hui") vs inaudible (e.g., "Je suis content que tu m'ecoute"). Oh, and things like the de/du/d' partitive, and how the future tense is used in situations where you'd just use the present tense in Spanish, and...it goes on.


I think this is a very good point. This kind of information serves well for some purposes. The original question does not necessarily refer to CEFR levels. I think "learning French to the same level as Spanish" can definitely be a good goal. I compare my languages like that too, and it helps identify my gaps and mistakes.

The other point that things like grammar cannot be overlooked, that is true as well, and there were various posts trying to make the distinction between the overall level and vocabulary level. Both have their place in self-assessment, whenever the learner needs them. I think it is perfectly ok, when someone prioritises other aspects of learning, but I have simply encountered the roadblock of knowing too few words repeatedly. It tends to be one of the main challenges of the intermediate plateau.

For the vocabulary level, there were various methods mentioned. I've just remembered another resource could be wordlists in coursebooks. The pro is a clear distinction of the individual levels, for people interested in cefr, and the words being sorted by an order that may be much more relevant than frequency. A native is unlikely to talk or write about toothbrushes very often, yet it is a much more important word for a learner than many of those higher on the frequency lists. The clear con is the fact such lists are not exhaustive. Both in terms of included all the vocab in the course (the authors seem not to include some words either by mistake or due to thinking "oh, this suffices for passive knowledge"), and "all" the vocab for the given level. Even several coursebooks and vocabulary builder books will vary a lot in content and complement each other to some degree, and that happens even at the A1 level. The higher the level, the more complicated this gets, until the C1 or C2, where you cannot consider this kind of resources exhaustive at all.

For the overall level, I really like the idea of exams like SIELE, TCF, and others, where you have one exam with tasks of various difficulty, and you get assigned a level based on your score. Their validity for just a few years could be an advantage, even though I think 2 years is simply too little. Doing them repeatedly could be valuable. The main problem is their prohibitive price. Unless I am clearly required one particular exam that happens to have an expiry date, I definitely cannot put that much money into something that ephemeral. But I could definitely imagine myself, not now but when I am older and have a stable job, to pay 50 euros several times, always four or five years apart. Two years is still not that much even for the presumed forgetting of stuff, and the repeated test taking could affect the results too much.
2 x

Tillumadoguenirurm
Orange Belt
Posts: 193
Joined: Fri May 06, 2016 3:07 pm
Languages: English
x 235

Re: Is there a relatively objective way of telling how much you know?

Postby Tillumadoguenirurm » Thu Sep 14, 2017 8:33 am

mcthulhu wrote:...


Thank you!
0 x


Return to “General Language Discussion”

Who is online

Users browsing this forum: jackb, nathancrow77 and 2 guests