Our data is released under various Creative Commons licenses.More information
If you love this content, please consider a donation.
Menu
Is there a list of sentences by difficulty? Let's say I wanted to order 500 easiest sentences in Korean, how would do I this?
The Tatoeba collection in any language is not fundamentally sorted. You can search through it with a number of criteria (length, creation date), but none of them pertain to difficulty. Assuming a Tatoeba member could come up with a workable measure of difficulty, they could choose to add tags to particular sentences indicating their perceived difficulty, or they could add a list ("500 easiest sentences in Korean") and assign sentences to it. But they couldn't impose a sorting order, such as easiest to hardest, on a set of sentences.
However, Tatoeba makes its sentences freely available, and I know of at least one site, clozemaster.com, that takes these sentences and groups them into categories by frequency of the least common word within a sentence ("100 Most Common", "500 Most Common", and so on). The rareness of the words in a sentence is not the only measure of its difficulty, but it's probably the easiest measure to calculate in that regard.
There's two ways to grade difficulty, either x grade reading level or language level exam level. For example in Korean, if you passed TOPIK test level 3, should be expected to know this sentence or not?
On another note, there's too many lists. I can't possibly see if someone did create a list I'm interested in, there's thousands of them
Clozemaster sorted the sentences in one way using word frequency lists.
You can do the same by creating a code or ask someone to do the work for you.
In this corpora you can't really sort things the way what is best for your criteria.
◼ For English sentences sorted by vocabulary levels, you can try these lists.
CK's OGTE-Level Lists
http://goo.gl/BnPz6h
Created in 2017.
Reading between the lines:
CK probably used the Online Graded Text Editor (OGTE), a tool that assigns English text a level based on the vocabulary it contains, in order to classify English sentences and assign them to lists that he produced on Tatoeba. Naturally, this doesn't suit your needs directly, since you want Korean text, not English. However, you could presumably look for sentences within these lists that contain Korean translations, assuming that the classification of vocabulary difficulty for the Korean sentences will match those for the English sentences. Alternatively, if you can find such a classification tool for Korean, know how to program, and have a lot of time and motivation, you can do the same kind of thing for Korean that CK did for English. That's a big "if".
As for looking through the lists on Tatoeba, it's true that there are a lot, but you can search through the titles of the lists. I did a search for "TOPIK" and for "Korean", but didn't find anything useful.
If you want to use Tatoeba-sourced sentences, it looks like Clozemaster is probably your best alternative.
> CK probably used the Online Graded Text Editor (OGTE), a tool that assigns English text a level based on the vocabulary
That's right, as the 2nd sentence on that page says, ...
These lists were created based only on vocabulary, using er-central.com/ogte, so some of the sentences will use grammar and idioms that are above the level of the lists.
Since I already had a list of Korean sentences sorted by word frequency, I uploaded the first 500: https://tatoeba.org/en/sentence...&direction=asc
Each sentence contains at least one word that doesn't appear in any of the 499 other sentences, so they're definitely not the 500 easiest ones (which likely contain a lot of repetition) but maybe you'll find the list useful nonetheless.
Ankissa on Tatoebasta (ilman mainintaaa) haettuja ja järjestettetyjä lauseita sisältävä pakka: https://ankiweb.net/shared/info/241481292
Sama henkilö on tehnyt useita vastaavia eikä mainitse Tatoebaa, mutta valtava Tom- ja Mary-lauseiden määrä paljastaa lähteen. https://frequencylists.blogspot.com/
I also believe she is using machine translation. I queried a few short Finnish ones with native Finns, and the English translations were bad.
Ehkä. Olen kuitenkin löytänyt muutaman kirjoitusvirheen tanskankielisissä lauseissa tuollaisen pakan kautta.
Epäilen sen johtuvan enemmän siitä, että Tatoeba, varsinkaan vanhempien lauseiden ja pienempien kielten kohdalla, ei aina ole laadukas. Mutta ehkä siellä on konekäännöksiäkin, en siitä tiedä.
> I also believe she is using machine translation. I queried a few short Finnish ones with native Finns, and the English translations were bad.
Are these bad sentences also present in the tatoeba database? If that's the case, then is anyone doing anything to remove or fix them?
https://tatoeba.org/es/tags/sho...th_tag/561/fin - suomenkieliset lauseet, joita pitäisi muokata. Osa on ollut listalla pitkään. Ehkä pitäisi itse pyytää muokkausoikeuksia niihin, koska kukaan ei näytä niitä muuten korjaavan.
En tiedä, ovatko mainitut lauseet tunnisteella merkittyjä, tokikaan.