Lingvist 2.0: thematisation of vocabulary

Hi Folks, I found Lingvist just a month ago and I am hooked! It's the most fun I've had practicing french in my life (aside from real-life conversations)! And my vocabulary is increasing rapidly - very effective learning tool.

Here's my wishlist for Lingvist 2.0: Have a series of Topic Sections, for example: all the words one would encounter in the kitchen (or any other room in the house), all the words one would encounter in any subject you find in a newspaper (politics, art, science, news, drive, crime, etc...). I would happily pay $10ea for permanent access to all those sections.

What do you think? Is this possible?

last edited by Marina

I would like to second Kira' opinion. There are lots of studies done that language learned in category can be retained better. It would be really nice to have the category/theme feature added. However I don't think having to pay for this service is a good idea yet. Duolingo is a direct competitor for Lingvist and personally I like Lingvist a lot better than Duolingo simply because it is more effective.

One needs to keep in mind that, even till now Duolingo has been developed for a much longer time, it still doesn't have a pay wall behind its content but more of the gaming features that you need to pay.

An application like Lingvist needs to be careful about its vision before charging, otherwise it is risking of becoming another Online Course site that not many people will use :D

But looking forward to the incoming features for sure.

Nice ideas. I have another idea. Lingvist tracks which words we know. That is a good thing. I wish they used it to search for presenting suitable texts to read for us on the internet. That would be an amazing feature. :)

@Calvin-Yuan I've come across the opposite: that learning by category reduces acquisition because the words, being closely related, interfere with one another in your memory. Better to learn by random acquisition as Lingvist teaches.

At the start of my Lingvist journey, I had a number of similar words pop up, and I still have problems remembering which is which more than six months later.

last edited by Kyle Goetz

They were structurally rather than semantically similar, but the same idea holds. For me, two- or three-letter prefixes followed by a four-letter root that typically ends in an e and is feminine. Aufgabe, Angabe, Anlage, and a few others still give me fits even though I can break them apart and understand why they mean what they mean, I nevertheless struggle with them. And the stats bear this out:

I've had to review Angaben a whopping 56 times. Aufgabe 47. Vorlage 38. Antrag 36. Vortrag 38. In my vocabulary list, these words all fall in the top 1% "most often got wrong." I came across all of these in a narrow window and have been getting them wrong ever since despite successfully learning thousands of other words.

last edited by Kyle Goetz

Hi Kyle, interesting points. I was more thinking along the lines of all the words that you'd use to talk about a certain subject, such as, stove, fridge, utensils, cooking, pans, pots, mixing, baking, etc... OR reading the crime page: criminal, theft, investigation, arrests, OR weather terminology, or way-finding, etc...
Ideally, Lingvist keeps the random 5,000 words, and then on top of that, we could click to choose different subject areas to go deeper on... giving us access to another 5,000 words accumulatively. Best of both worlds!

No, I know. It doesn't make a difference: I've seen references to studies that this kind of vocabulary grouping slows down retention, not increases it. Like if you learn a bunch of colors at once, and then a bunch of different birds, and so forth, it's less efficient than learning them somewhat randomized (although still controlled by frequency/importance so you aren't learning "African swallow" before "purple").

Guys, I am loving this conversation and the ideas you guys are having. Will have a thorough and hopefully interesting response to this on Thursday!
Keep thoughts coming until then regarding thematization, pros and cons to learning vocab by topic, what topics would be of interest, how you would want to see it in the app, etc. @Kira-Callahan @Calvin-Yuan @Alexander-Hansson @Kyle-Goetz and anyone else :-) !

last edited by Lisa-Lingvist

I would like to throw another aspect into this topic. I don't really know which method has the better retention rate, but I'd like to think about the usefulness of the time spent.

Word collections sorted by subject areas have, in my opinion, two big advantages over pure frequency-based learning lists:
Firstly, you can choose which words you may need in your specific use context.
Secondly, I feel like a more ore less comprehensive set of words for a specific topic is more useful than sprinkled, single pieces of vocab distributed randomly over topics.
Thirdly, it should be easier to implement for a variety of languages, as these word collections and example sentences could be more or less the same for all languages.

I really like the idea of theme vocabulary. I think it's very motivating for people to study things they are interested in. I think Medical, University/Education, Computer/Tech, Food/Cooking and Travel would be some interesting topics.

Hi Lingvist team, if you're checking in, my hope is that category-specific vocabulary would be in addition to what you've already developed. I LOVE the 'first 5,000' and I'm glad to have that as my base. (I'm at about 4300 words right now.)
Perhaps the first 5,000 can be a gateway. You complete it, and that unlocks a menu of them vocab.
As far as retention goes, I believe that if we are choosing a category that is relevant and useful to our daily life, it will quickly be reinforced. For instance - travel vocabulary. I travel on Air Canada flights fairly often and would quickly come to understand the announcements in French, read the safety card and in-flight magazines in french, Or if we have a kitchen/cooking vocab, we can watch YouTube cooking videos and be able to understand it all!

@Kira-Callahan I think it is a great idea that the specific vocabulary comes after the 5000 more commun words. For me, it has to be something to really enhance your vocabulary. For example, the french vocabulary about cooking won't be "cuire, couper"( because you have already learned it in the 5000 words) but "rissoler, mijoter" (Broil, simmer).

Hey hey!
Coming back to the discussion around learning vocabulary by topics, thematised courses and so forth.
This discussion really shows us that for many of our users, learning language around things that they are interested in, is a big desire. And we agree! So here goes the low-down ;-):
Thematisation has been on our table for three quarters now and we have twisted and turned it to understand what would be best for users, what is innovative and what is doable from a technical/product perspective. As @dev_temp warned me to not over-promise things, let me tell you the things we are working on currently as well as the mid-term projects regarding this (subject to change - as is often the case in a start-up).

We are building a medical module targeting travellers going abroad, needing help in a hospital or at the pharmacy, as well as for medical professionals studying in English speaking countries, interacting with English-speakers and so forth. This will be 500 cards of fairly dense medical vocabulary (we call it the ‘gross’ module… you can imagine why).

We are also building a sort of travel/moving abroad/hospitality module. This is for the travelling adventurer as well as hotel staff, flight attendants, etc. interacting with English speakers.

Now, as far as a mid-term project, we are working on a much more personalised version of this. You may not be interested in syringes & co, and you may not care to learn travel related words. Maybe you are interested specifically in the music industry, gardening and astrophysics. Our goal is to let you choose what you want to learn and provide a custom product that suits your interests. After all, we are the most enthusiastic about learning something that interests us, right?!

And @Kira-Callahan Ideally, Lingvist keeps the random 5,000 words
No worries, we plan to keep a "general" course as an available option for the foreseeable future. Our EN course now has actually nearly 9000 most frequent words in English and like @Kyle-Goetz and @Jonathan have mentioned, there is research done on both sides - retention of vocab learned randomly vs. by categories. But regardless of how you feel about it, giving you the choice can’t hurt, right?

Thoughts, ideas, feedback welcome. I have the pleasure of heading this very exciting project so this is your chance to help shape it.

Awesome @Lisa-Lingvist, thanks!! I'm so excited about personalized vocab;)

@Lisa-Lingvist Thank you for asking. Regarding my idea, I would like to see texts highlighted based on the level of comprehention. I would suggest you checking out the program "Learning With Texts" aka. LWT (website: or LingQ (website: to see what I mean. In that case, unknown words in a news article or text should be highlighted with a color and "well known" / "already known" words should not be highlighted. That is a great way to develop the reading skills. That is a nice visual representation of what you can understand. I hope you like my idea. I love Lingvist! :)

@Lisa-Lingvist Hi Lisa! Woo Hoo! So glad to hear that Lingvist is working on thematic vocab. Questions in response to your post:

  1. When can we have 9.000 words in French? Would love another 4,000 on top of the initial 5,.000 you have now.
  2. What's the timeline on launching the travel/hospitality and medical themes?
  3. How much longer after that do you think it might take for any other topics, and how can we influence which topics come next? (months? Years?)
  4. What can we do as a community to help you speed things up?

My votes for next topics:
Home (there's got to be 1,000 words to cover every object in every room in the home...)
Business language.
Daily news (use the sections in the paper as subject guides)
***Directions! Way-finding language - perhaps this would fit right in with your travel.
Thanks so much for all your great work!

I'm not sure if Lingvist is ready to "talk business" or it's still pleasure for the "family crew". If I were in startup aiming to build a user base and maybe, just maybe, gain some profit, the questions I would constantly ask myself would be:

  1. how unique is that (was it done before);
  2. how popular it could possibly be;
  3. how difficult is that to implement

I see @Lisa-Lingvist have mentioned some of those questions in her text, but let’s take a look at particular feature “thematized word sets” (I guess it’s too early to call it “personalized “ as in my mind it has very different meaning). So, how unique is that? There are well known players on that market like Memrise, Quizlet, Cram just to name a few. There is certainly a demand for that feature, but how do we stand out among those competitors? All of them share a common problem – it’s user generated content and quality may vary. So, obviously, Lingivst guys were thinking “how do we solve it”?

First thought: let’s create more complete word sets, which cover everything, like “500 words to ask for help” complete. Well, that’s insanity! I can imagine someone taking a course consisting of “Ruf den Artzt” and “Es tut mir überall weh“, but 500 cards of diseases, medical instruments and kinds of medical professionals that is for fanatics! I doubt most average people will ever end that course.

Second thought: let’s make a bunch of highly professional sets like Pro Medicine with thousands medical terms, or Pro Botanic, or Pro Mechanics etc. Well, it’s utopia. You can’t really make general medicine course without making it uselessly big, or splitting it into more specialized subsets, like Psychology, Surgery, Anatomy, Patient care etc. Just to give you a rough idea of the possible variations you may take a look at Cram’s medical section. You gonna need a professional in every specialty to create a proper word-set, it’s not enough to have a computational linguist that know how to crunch numbers from a book, you need someone who knows the field. And it implies money and time. But let’s estimate a demand. There are definitely some professional that need that kind of training, but it’s just a modicum of general language learning community. If you were a huge profitable company like Google, Amazon, Facebook, you may allow yourself to spend time and resources to satisfy that tiny groups, but it’s a luxury for a startup.

I have nothing against general topics like home, tourism, pets, children, idioms, but you ether should keep the courses small (like 50 – 100 cards) or divide them into sections (Pets 101 (for beginners) , Pets 202 (for advanced learners). That wouldn’t attract lots of new users and you have to keep that in mind, but it at least some kind of answer to a question I've finished the course. Now what? Also, I like the idea of small general introductory courses like you did with Essential Estonian. I would be terrified to sign up for a 9000 words course, but to give it a try and learn 100 words, why not? I would love to learn 100 Spanish, or Czech, or Portuguese words. It’s fun, it’s engaging, especially if there is a possibility to upgrade to a full course.

What could really make you shine is specialized courses for test, like 2000 most popular TestDaF words, or 5000 TOFEL words, you name it. There is a huge user base preparing to those tests, much more then surgeons or plumbers moving to Estonia and going to learn some professional Estonian. My guess is that this effort would pay off much more quickly.

All that is feasible with the current technical platform, but if we are talking about Lingvist 2.0 in general, there is so much to do to improve the system as a whole. First of all, beginner needs some guidance, if you can tell (you can ask) that a user is a complete beginner, he probably needs some guidance. When I first time logged into the webapp I chose a French course (there was no German at the moment) just to see how system works in general. Having no previous knowledge of French I had no idea even how to read, but I was supposed to memorize and write words. I believe there should be some kind of a small introduction to the language that guides you and explains the basics and isn’t hidden inside the grammar tips.

I like the idea of being able to read in less than 200 hours, but if you want to evolve, you may also think how to cover those learners who want to learn how to speak and listen, but don’t bother about writing. There could be some kind of selection instead of typing in modes or even some kind of speech recognition.

There is also a huge demand in some kind of replacement for closed down Reading and Listening section. Because that is irreplaceable for learning and at the same time, something that can make user stay on the service after finishing the course. It may be user-generated content or authentic materials in partnership with educational organizations that are more interested in language popularization than in gaining huge profits, like DW or BBS, for instance.

P.S. @Lisa-Lingvist first of all, you’ve taken my message about overpromising too personal, second of all, there are different kinds of overpromising:

In most industries, unachievable promises are a sign of bad leadership. But in tech, where companies are built on impossible ideas, unreasonable pledges are just a part of doing business.

If you promise to create a thrill and excitement it’s good for business. Learn language in less than 200 hours is an exaggeration, but it creates hope and excitement. Only very naïve people would think that it’s enough to go through a bunch of cards to be indistinguishably close to a native speaker. For most of people that “overpromise” doesn’t create a frustration at the end of the course, ‘cause you’ve still learned a lot and can do whole lot more than at the beginning. When you have a button in the interface that says it does something, but pressing it gives you squat it creates frustration and it’s a bad thing. I hope my point is more clear now.

last edited by dev_temp

Hi everyone,

I would also like to leave my opinion here. I have been using this website for a year, with some up and downs of course, but I have enjoyed it since I learnt of its existence.

First of all, let me tell you my story. I had been trying to learn a language for a while when I found lingvist. I tried using some offline sources, such as a book with a CD for audio exercises as well as some online sources such as duolingo, memrise, lingq, glossika and couple others I cannot recall right now. None of these methods was really working, so I kept digging for other sources of learning languages and I found lingvist. At that moment, they didn't have the language I was studying, hence I tried learning French just so that I could test this website for language learning and, in a few days, I was hooked to the website, in spite of learning another language on it.

I think you have one of the best websites for language learning currently online. You are somehow unique and the approach you take to teach is simple and effective. Though, there is still plenty you could do to improve. Thematisation of vocabulary is NOT among the approaches you should take in the near future, or ever. One thing I hated about duolingo is that you study by themes. But we don't live ours lifes in parts, like we are going to talk about fruits now, then about means of transportation and lastly about family. We live our lives talking about stuff in general, so when studying a language, there is no topic that should come first or second. There shouldn't even be themes because in real life words come as they come. Of course, knowing water in some language before knowing football in that same language may prove to be useful, but that does not mean we should learn all beverages prior to know how to spell the most common 2/3 sports.

I believe the approach you have been taking to teach is the right one and, in my humble opinion as someone who has tested other softwares/websites, the most effective. However, it should be better complemented. Another website I like but I don't use because it is paid is LingQ. For those who are not concerned about this platform, in this website, the student reads texts and underlines unknown words for further revision. By doing this, the student must interpret the information and try to understand it. Whatever is not understood, he or she can review later the words he or she didn't understand. Besides, there's translations linked to each word. In lingvist, there is the opportunity to do something similar. However, the student cannot study the unknown words later on the website. Moreover, the study of the texts is not yet linked to the study of the words (when we see words in texts, they are not considered as learnt, as in the course).

As I am a believer of learning content from context, I strongly believe that increasing the number of words in the courses is not the best strategy because they are already long enough. Maybe one or two could be extended a little bit, maybe. But I don't think it is really the problem. I don't know how much is enough, but somewhere in the few thousands words (3-7k) should be more than plenty. This amount of words should stand as the base for further incresing the knowledge about the language to learn it from real sources, such as texts. These texts could be from newspapers, music songs, articles, etc... Essentially content that is important to the learner. So, if possible, extending the library of content should be a priority. Maybe, even dividing the contents by topic or type, so that the learners can choose.

To wrap up, I would like to congratulate you on this website. It is amazing! And, I don't know why you didn't create themes for language learning up until now, but whoever made that decision, in my opinion, made the right decision. In the end, it is up to you. Do what you believe is the best approach to language learning and we will all be ok with it :)

Best regards,

I like the pure spaced repetition approach. I suspect it's the ideal way to learn vocab, and themed practice, as intuitively attractive as it may seem, can only throw a spanner in the works.

Well, sounds like we've got two camps. The 'NO, don't ever thematcize, and the PLEASE, provide thematic options (I'm in this group) as soon as possible! Hmmm... how to cater to both camps, I don't know. Can't wait to see!

  • 41
  • 11648