Topic on Extension talk:WikibaseLexeme/Data Model

Language learning features

5
Ajoposor (talkcontribs)

Is it possible to include fields at the Lemma or Senses level that may later support language learning?

I'm a daily user of wiktionary mainly because I've been learning languages already for many years. I would say that the use of wiktionary for my mother language could be a tenth or a hundredth compared to its use for second language learning.

In learning a language it could be helpful to be able query the wiktionary for sense by its frequency. There are frequency list already, but I've not seen yet one that does so by sense.

In addition to examples by sense, there could be a section, also by sense, supported by the learners community to include . For instance, mnemonic sentences and or images that may help while learning a word. This mnemonics could be rated by users, so that the most voted are displayed. Here there may be an option to set the mother language, since a mnemonic could be very good in your mother language is French but have not sense for someone that speaks Japanese.

With this data, some sister projects may be developed as learning tools, like learning cards, improved from what is already available outside wiki, using not only repetition learning, but good mnemonics.

Denny (talkcontribs)

I like the ideas a lot. Regarding the sense frequency, we would need some source to get this data from - is there something like this?

Tropylium (talkcontribs)

I'm not sure I follow entirel, but this is not data that is possible to record "about a word" in isolation. Frequencies always apply to a specific corpus, e.g. prose, newspapers, technical writing, online forums…, not to a language as a whole. It therefore seems inappropriate for Wikidata. I agree though that the Wiktionary coverage could be improved a lot, but that's something to take up at the individual Wiktionaries you're interested in.

Ajoposor (talkcontribs)

Hi, yes, a frequency is specific to a corpus, but it doesn't mean we must discard its use. We may need to agree on a Corpus, for instance, the whole wikipedia could be used as a proxy. There are some dictionaries that already display a frequency. So frequency counting could be done with algorithms so that frequencies may be adjusted over time.

But going beyond the usual frequency, I would suggest having a methodology so that frequencies be assigned to each meaning (there are many words with multiple meanings, some of them rarely used, an that poses a problem to language learners). In order to do so, there could be an algorithm that takes sample texts containing a word, this sample would be left to users to be analyzed and assigned a meaning numeral to each appearance. In this way, there could be a calculation of frequencies by meaning. It could be made through a gamification of this task.

I tried to find out who are the users of dictionaries but couldn't find an answer. We may be so used to dictionaries that consider them as a given. But it is important to know who is using the dictionary and what are their needs.

I'm interested in addressing the needs of language learners. Learning a language is a task that takes A LOT of time. So it is a perfect target for improvement and optimization.

BoldLuis (talkcontribs)

I agree to use this frequency data.

Reply to "Language learning features"