Original Canny votes: 11 | Status: Under Review
If I already know “pen”, it would be nice if it didn’t show “pens” as a new word. Especially frustrating in korean since the same word shows up so many times because of different conjugations
Original Canny post: Conjugated forms shown as new words | Feature Requests | LingQ
Agreed. Plurals and verb conjugations would be nice to eliminate or at least have the option to eliminate
I’m guessing the downvotes on Canny are because the titles sounds like the suggestions WANTS conjugated forms to show as new words
日语也有同样的情况，不同变形不同后缀的词汇都会进入词汇库，同样的动词不同的变形可能会有5到6个词汇学习后进入词汇库，这在语言学习初期，变形掌握不熟练的时候，用户可以为不同变形的词汇添加不同的含义，来不断的学习，掌握词汇变形。
但是掌握词汇变形以后，这个特性就没什么用了，因为用户已经掌握了动词变形，所看到的词汇不需要添加释义便可以理解变形，后续的学习只需要不断增加词汇量即可，但是这样一个特性又让lingq的词汇系统变得非常繁琐。
我认为你们应该添加这样一个选项，在词汇释义的选框里增加一个只需要点击展开，就可以看到该词汇的全部变形的选项，可以完美兼具查看词汇变形的有点，又不用添加多余的词汇.
这个应用叫做[moji辞書]
A possible modification would be if the app would check whether there definitions already been made for similar words, and propose these as definition for the “new” word, too. To reduce workload this could be restricted by considering only words that start identical and/or definitions made within the current lesson or lessons that belong to the same course (possibly restricted to the last X lessons within that course).
I actually really like how inflected forms map as “new” words. I study ukrainian from russian, and defining each new case of each word with a word in the appropriate case has really helped me understand the case endings. I wouldn’t mind there being an option to reduce this but I would hate to see it mandatory.
I feel the same way about German. To be honest, I’ve stopped caring about the ‘known words’ count; it’s not as if I’ll introduce myself in real life by saying, 'I actually know 32,000 words on LingQ 
'. My goal is to live in the country where my target language is spoken, rather than just seeing numbers on a page to feel better. We really need to detach ourselves from that mindset.
Regarding the topic, I don’t think it’s that important since we already have the option to ignore words. If, instead of marking words as ‘known’ when flipping the page, they were moved to ‘ignored,’ I think the process would be much smoother. Also, the suffix ‘-s,’ as mentioned in the post, isn’t always for plurals. As you noted, it might be a declension form that we are still trying to master.
It’s actually complicated to determine whether a word belongs to one family or another or stands alone.
In various videos Steve Kaufmann has explained that it was an early design decision to keep LingQ simple and fast by treating all words as unique. The intention is to provide some feedback on a user’s vocabulary progress, not a perfect vocabulary count.
I think it was a good decision.
I prefer the way it is now, because not all languages are the same, and our brain doesn’t operate in the same way all the time.
For example, sometimes I can recall without any problems a series of verb forms but for some reason, others are not so easy to remember, or they generate confusion. Maybe because there is a conflict with other words in my mind, I don’t know. I just observe the difference or struggle.
For me, each word is a different word.
So, pen is NOT pens, word is NOT words.
Because the mind might understand the general concept, but pen, pens, and word could be easy, and for some reason, words is not for a much longer time.
And there are many other reasons why one simple letter or declension can actually be other words or meaning, depending on the situation and language.
LingQ choice is a more relevant choice, for me. Imho.
I use Anki with Ankimorph addon it is able to create a database of all the words from my sentence card and as I use anki it is able to show number of Lemma and Inflection that I know.
It was fairly complex to setup this addon. I’m not sure how taxing it would be on the servers because every time I want that stat to update I need to click on that recalc button which takes a short while to calculate.
I am fine with marking and treating every word as unique, it makes things simpler.
I think it would be cool if we can see a detail break down under known words stat into base words, lemma, inflection or whatever other morphology but it is not a critical feature, just a nice to have feature. if it is really important we can just export all our vocabulary list into other external tools to get those stats.
I don’t think a computer can really identify new words anyway.
(A)
Homographs are different words in a stronger sense than inflected words, but the computer can’t easily handle that.
I think it is easier for beginners in Russian to be told that это the invariant particle is an entirely different word than the inflected “этот / эта / это” the determiner or the pronoun.
But the computer is not going to be able to handle that.
(B)
Also it is not always clear which inflected forms belong to which root word.
Some words have overlapping forms: плАтит & платИт are forms of two different verbs in Russian that look identical but sound different and are unrelated, and lead (noun) and lead (verb) are two different words in English that look identical, but sound different and are unrelated.
And then you can get really tricky cases like… is “человек” really the same word as “люди”, because they are handled in declension tables as forms of the same word?
I don’t think the main issue is the known word count. I personally ignore that, too. The issue I see is that if you want to have definitions for different conjugations of one word and the suggested definitions (ai or not) aren’t that useful, you can get into a lot of typing work. This can get tiresome and slow down the reading unneccessarely.
What I do is to just create definitions for the first forms I encounter, and if I see a word coming in a lot of different variations within the subsequent text (I use page view only) I use copy and paste. But this may cause the word occouring undefined in an upcoming text, requiring me to find out the proper meaning of the word that I had searched for already, somewhat rendering the whole LingQ approach a bit pointless.
In regards to how words are unique: The effect described in this thread, that one word can take several form without a real change in meaning also exists the other way around. A form can equal several different words, either because they are homophons (different words that sound alike) or because different conjucations, declinations etc. of different words result in the same form.
So I would like to repeat my aforementioned suggestion in one way it might work:
Add an earlier in the same lesson created definition to the suggested definitions tab of words that start with the same letters, where start means the words share the first half or more of the letters or symbols (like Kanji).
This approach wouldn’t change anything substantial for those who like it the way it is now, as there would just be one extra definition suggested (hilighted with some symbol as it’s done with the ai suggestions so people are aware of it) while at the same time improving the workflow for those who struggle with this.
I like your idea of it being optional, user controlled instead of auto.
So when we click on a word in the popup that appear there could be another section there in the popup where we can see all the other conjugations of selected word (there could be alot of them though). This list of other word forms could be loaded from external sites or some database if they exist. (To add, if we had a LingQ app that can read and work on webpages, we can easily mark all the various forms on that conjugation site.)
For those learner who want to mark all the other form of the selected word as known they can do so there automatically by clicking on a Mark-all-forms-known button.
They can always mark any form as a new LingQ if they were to encounter later and realize they don’t know it.
That wasn’t what I’ve meant. Maybe I should give an example.
Let’s assume you are learning German and read a lesson in which you encounter the word fahren. You create the definition to drive either directly or choose it if it was suggested. Later on in the same text you come across the word fahre. Among the definitions you should - in addition to the ai and the suggested ones - find your already created definition for fahren, as fahren and fahre are similar and thus could be the same word (the app cannot know this, but the user can decide). So in addition to what we have now, you could reuse an already created definition.
Obviously fahren is a rather simple word where this feature is unneccessary. However, for more specific words there might not be good user created definitions, if any, and the ai may also suggest nonsense. On the other hand, if you are at a level where you start to stumble into such issues you are most likely advanced enough to be able to judge on whether a word you encounter is just a different form of another one you already came across, and all the app now has to provide is to list already created definitions of words that are similar enough. As there could be several such words, it might be good to not only suggest the definition but also show the word that definition was created for.
In regards to similarity there some language specific aspects that might need consideration, though. In German, if we stick with that example, it would probably be good to deal a, o, u and their respective umlauts ä, ö, ü as identical (so a = ä etc…), as german words tend to switch to umlauts on conjugation or when forming the plural. Another aspect are irregular verbs. Those are relatively basic words, like the aforementioned fahren, and thus this might be ignoreable.
However, such specific aspects would need to be taken into consideration for this feature, if it would come, to be of use, imho.
Got it, So LingQ goes through your library and propose definition from words already marked of belonging to same family. That can be very useful for data entry to keep thing consistent.
I didn’t read all the above, sorry, but I remember I proposed something similar in the past. To add our own vocabulary database where we choose our dictionaries. So we can add our database too and quickly pick from there already existing definitions if they match.
