What is your experience with Glossika?

[Cavesa]

I am right now using mainly digital tools for review of my languages, because I have tons of other stuff to do, so I want to keep it playful, but still only with valuable tools (so no, I am not trying Duo again ). I've noticed Glossika has changed a lot in the last few years. It used to be just about repeating after audio, now you type too. It looks quite good, even though the short trial doesn't show some stuff.

I've asked them a few questions, it looks like they have 7000 sentences from A1 to B2 for the big languages, 3000 for the small one I asked about (Hebrew. Not available on Speakly at all). Unfortunately, the English based courses are the biggest one, so much for the "learn from whatever language you want" thing.

My expectations from such tools in general are rather realistic. It's about exposure, typing, saying stuff, review, something possible to do on top of other work. I am not a fan of "no grammar explanations", but I don't need those in this app, I have them elsewhere.

But there are a few different things I am curious about, if you'd be so kind to tell me:
-are you content with Glossika, if you've tried it?
-what do you like about it? What is it good for, and what is it simply not meant to do?
-have you noticed some results?
-is there any grading mechanism or feedback to your pronunciation? human checked (like Speechling), or software checked? Is it reliable?

[tacerto1018]

So I tried it a few years ago and have always been eager to use it for real, but it was always too pricy for my poor wallet. Recently, at the end of March, I got a promo email saying that it was 1/2 off per month and promptly took advantage.

So far, so good. I have around 900 reps in Icelandic since I started about two weeks ago and I have dabbled here and there with others.

The good, so far:
- I love the interactive bit. I have been using Colloquial Icelandic and I wanted a supplement for speaking and listening so I could use that book for primarily grammar and vocab and Glossika fit the bill.
- The audio quality is nice. Professional recordings, not too fast, but definitely natural speech.
- It's easy and accessible. No frills website, it's grab and go.
- I've already noticed how much I have been repeating words and phrases in my head outside of my 'reps'. Like today, I couldn't get, "the customs in their country are similar to ours" in Icelandic out of my head haha.
- The amount isn't too overwhelming. It's honestly just the right amount of stuff at 5 new sentences per day.
- There are so many languages. I really love the variety. Makes me want to pick up some more! (I just bought a Colloquial Indonesian, but don't tell anyone. So I might try that)
- Various modes. You have just listening, listening + typing, or listening + typing + recording. (to my knowledge there is no check of audio. I haven't explored much yet, but I would assume it's saved somewhere where you can listen to yourself)

The Meh:
- One speaker for Icelandic, and he sounds kind of old. Not a bad thing necessarily, but I would have liked some variety.
- No app. It's all browser based
- You can't listen to your recorded voice after you've said it, it just moves on to the next sentence
- No C1+. I am C1+ in French and the highest the test in thing goes is B2 low apparently.
- I have more but I can't think of them at the moment. I will update when I do.

Overall, I look forward to doing my reps every day. I have 9+ hours logged already and it's great and portable because I can do it on my phone even when I am trying to quell my screaming newborn.

Oh, I definitely wouldn't use it from the beginning. I tried it with Indonesian and since I have no previous experience with Austronesian languages I had no idea what was going on really. So I would say ideally it's where I started in Icelandic; a few months of getting the basics of vocab and grammar and then hop in. I assume it would be harder with a ideographic language but I haven't tried. I also have studied Dutch, German, and Norwegian before and speak English as a native so Icelandic isn't totally foreign to me.

I like it so far, so I'd say go for it, if you wanna try it.

[IronMike]

I liked it back when you could "buy a language" and be done with it (them). Once they went subscription basis, I quit them. Still have the few I bought (Serbian, Lithuanian, Latvian, Spanish), but won't subscribe.

[Cavesa]

Thanks!

Yes, subscription is not too nice, I'd prefer one time payments too.

tacerto1018, thanks, this is exactly the sort of review I wanted to find. Less wordy than what I tend to write (I have a real problem with that), but based on experience you describe, a nonfigs language, and with pros and cons.

I was offered 20% off, when I asked a few questions. So, I am considering it a bit. But clearly, Glossika is expensive as hell, if you don't get any kind of a sale.

It looks like it could be beneficial, especially for fighting my resistant language enemy-German. And for a new language, even though I really appreciate your very realistic evaluation, that while the content and exercise is very useful, it is simply not accessible for a beginner and as the main resource (I really wish companies would stop lying about that).

Natural speed (even though not too fast), that's a huge win. That's one of my pet peeves with Speakly, which I otherwise really like. The audio (and the website loading) is sooo slow! If Glossika picks a good speed for the learner, grat.

The amount is probably one of the key factors in my decision to subscribe. 3000 for the "smaller" languages, that is not that much, but still 3000 more than majority of the competitors So, I guess the timing will depend mainly on my decision on taking up a new language. Or is the Icelandic course even smaller than that?

The modes are interesting. I'd have to learn how to type in Hebrew, to really use it, because I am not interested in the modes without typing. Yes, you can find your recordings, but on a separate page with your progress or your sentences. It is not too practical. When I asked about that, I was told that Glossika is indeed aware of the "fluency" vs "accuracy" issue, and is more after fluency, thinking listening to oneself right away would make people repeat slower and not develop the fluency part enough. Well, my goal is not to be a fluent neanderthal. This is one discipline, at which Speechling is clearly the winner. But I was told by Glossika, that the audio is somewhat evaluated (it is even mentioned somewhere on the website, where Glossika describes the methods). But I haven't gotten an answer (yet), as to how (person or computer?), and whether it affects for example the SRS, or whether the feedback is in any way communicated to the learner.

The easiness is a huge +, because it is an easy activity (to do even after a long day), but not dumbed down. Typing stuff and saying it, isn't that the base of the scriptorium method, just digital? No picture matching and similar waste of time.

These mehs are actually very reassuring

I am not an Icelandic learner, but I agree that varienty is nice (who knows, what are all the languages I am interested in like).

No app, no problem. I don't like apps, I prefer to type on my computer. The only exception, where an app sort of found a place in my heart and routine, is Clozemaster. But other than that, I'd say most digital tools would be the best, if they stopped wasting efforts and resources on a phone version.

I agree that having to go to a separate page to listen to yourself, instead of having it just played back, that is a clear con. Or meh.

No C1+, that is surely a disadvantage. I agree I'd like this kind of practice at the high levels too. An easy solution: I will not use Glossika for French practice, I don't really need it (but honestly, I might use Speechling for further pronunciation perfectioning, as I think some of the tutors are strict enough to be useful at the high levels!). Content only up to B2 is in general not a big problem in my eyes, even though I don't know, whether my B2ish Spanish would profit from Glossika in that case.

That's the issue with subscriptions for all the languages. I understand the companies, they want to get as much money from each learner as possible, and saying "yeah, but you get 50 courses for the price of one" sweetens the deal. But I am not sure Glossika is a good deal for learners of "just" one language with other options. And for the rest of us, it makes the choices more complicated

Please, update with further points, when you remember them!

[Cavesa]

I got my trial prolonged by another week, and was hoping to get a few questions answered by their support team, to make my decision. Such as: what are the planned changes, are the courses I am interested in planned to be expanded? Is Glossika using or planning any kind of speaking feedback (as it is based on users recording themselves)? And what is the SRS based on, how does Glossika evaluate what should I review, as there is basically no way to make a mistake in a rep (except for typoes due to lack of attention. But you are just copying, you cannot mess that up otherwise). I asked in different words, because my questions were not too clear, and even gave examples to make it clear (paraphrasing: Speechling uses human speech check, some other platforms software, does Glossika use either of them? Or only self-evaluation on a separate page?)

Perhaps it has something to do with becoming older, or living in this too detached world of superficial contact. But I judge companies more and more by the way they treat the customer, the way they communicate. Whether they answer questions (and what the answers are of course), whether they are polite, whether they are trying to provide a good service addressing the customer's needs.

The first support person S. was very professional and useful, answered my first round of questions (ok, some people might find me too curious, but Glossika is rather expensive so questions are only to be expected. Answers to basics like "how many sentences are there in each course" should even be on the frontpage!), and this person actually made me really swing towards paying for Glossika.

But today, I got answer from M. I am not sure, whether it is another support person, or it is the founder, who is on youtube. The name and the foto would probably fit, but the person on youtube sounds educated, polite, and speaks mostly straight to the point. And why would the boss of the company answer questions from a random user, who hasn't even paid yet. The person in the Glossika chat is arrogant, wastes a lot of time typing irrelevant opinions that tell me nothing about Glossika (nor anything useful for my learning), and probably didn't even understand the questions. For some weird reason, he even apologized for not answering in French. I had been communicating in English the whole time, never demanded French (or other language). It really looked as if the person hadn't even read the previous post in the chat properly.

I'll quote the answer. The later parts are useful and contain very good info about Glossika's SRS algorythm. But a large part of the answer, especially at the beginning, is just a pile of arrogance, of answering something totally different, and giving me unsollicited advice. I am deciding whether Glossika can help my vocabulary and speaking, I don't need superficial lectures about listening (something I already excel at in my languages), or a condescending "explanation" how detailed pronunciation feedback would be either too complicated or just counterproductive to me anyways

Awareness
Improving your listening
Glossika Spaced Repetition
What to look forward to in 2021

I'm sorry I cannot respond in French.

1
Everyone starts their language learning journey with awareness.
Without awareness to details, you're unable to make improvements, no matter how many times somebody tells you're wrong.
For example, you can't say "it bases the choice of what should I repeat", because your verb "should" must come after the subject "I". But that doesn't mean you are now aware of the mistake and won't repeat the same mistake again tomorrow. And "speakly" is not a word; neither is "speechling".

2
By recording your voice and using the interface that we provide for comparing all your recordings (open up your recordings by clicking on the sentence in the sentence list) and actively listening for how you sound against native speakers, and becoming aware of what's happening in your voice, is how you improve. No system will give you accurate feedback, not even humans, unless they are trained experts in IPA and can accurately hear all of your tongue positions, and then you probably won't even understand the feedback you receive. But this is a very technical skill that few people can do, and it's very expensive. I can do it for you, but it would cost you dearly. I use technical vocabulary like fricatives, palatals, and vowel height, because it's very accurate and tells you exactly what you need to improve. That information, unless you're aware of it yourself, is counterproductive and will not actually help you improve. You need to listen to your own vowel height, the quality of your points of articulation, etc, in order to make the necessary corrections. If you find a service that is offering human feedback at a very cheap price, then cheap results are what you're going to get.
Create a more productive approach for yourself: listening comes first, and the accent and the grammar and the intonation you are imitating becomes so ingrained that any slight alteration from it leads to a sharp feeling of discomfort in your own ear. That's my experience. If that discomfort doesn't happen to you yet, you don't have enough experience with the language yet. And it requires great amounts of dedication to reach that level.
In late 2020, we developed an intonation tool that shows the intonation graph of the native speaker against the intonation graph of your own voice, and this is currently queued for launch for all of our tonal languages. Again, this tool provides a way for you to discover and become aware of what you're doing different from the native speaker, rather than telling you what you did wrong, which is counterproductive.

3
Note: when I say "review", I also mean "revise" as equivalent in meaning.
At Glossika, we use Glossika Spaced Repetition, a set of algorithms that do several things. For every item, you start with a rate of decay calculated with the mathematical constant "e", which decays at a rate over time consistent with your level of stability for this said item. Since when you first begin, your level of stability is 0, your rate of decay from 100% falls quickly. The calculated stability growth is dependent on the sum calculation of success from every user in the system, so that items that have a higher rate of success across all users, indicate the item's relative difficulty, which influences the stability growth.
All items are given a priority ranking for review, which means that no matter how quickly an item decays, the items that were most recently learned will show up with highest priority. The reason for this is because if you do not have enough time today to get to all items, it's better to review weaker stability items to increase their stability, than to review items from a week ago that already have stronger stability. This prevents losing brand new items to having absolutely forgotten them. Once the highest priority items are cleared, you move on to older items that have stronger stability levels. Once the stability rises, the rate of decay slows after each successful repetition, meaning that the item will appear less frequently for review, to the point that several weeks or a month may pass for the system to recognise that the item should be reviewed again.
This in effect is a predicting system, which predicts a point a time just before you are about to lose an item to absolutely forgetting it, and by boosting your memory with that repetition, thereby building stability once again and slowing the rate of decay once again. Once the stability surpasses the rate of decay threshold, it is determined that the item no longer needs to be practiced, and can safely be said that it has successfully passed into long-term memory.
Successful reps compound stability growth, unsuccessful reps do the opposite. Listening and speaking reps compound stability growth, but many users do it passively rather than actively, and this compounds at a much slower rate than full practice mode, which means that more reps are required to achieve the same levels of stability. Stability influences the rate of decay, and once decay has passed the threshold for review, the item is assigned a priority rank and sorted with all your other review items and presented in your review. This algorithm provides a unique experience for every item for every user in the system, but we do collect the sum of all users experience to apply difficulty weights for each item, which then affect the compounding of stability.
In comparison, Anki does nothing at all. It doesn't calculate stability, it doesn't know the comparable difficulty of each item across millions of people, and it doesn't calculate the rate of memory decay. It simply asks you how you feel and when you want to review it again. Glossika on the other hand removes all the guessing for you and manages thousands of items and sorts their priority review rankings for you automatically, which in turn saves you a lot of time dealing with which items you should review, and when you should review them.
A lot of automation has been built into Glossika so that you spend less time focusing on organizing or finding your materials, and spend more time on training. Glossika does a lot more than this in terms of sorting and filtering content for you. If that time saving is not worth the small price you pay to have all of this automated for you, then the subscription is not for you.

4
What we're releasing in 2021, besides intonation mentioned above, is simply lots and lots of more content, especially in the higher levels B1 to C1. We're working on adding lots of more voices in both genders, and adding discourse register information for each sentence. All of this is currently getting automated to interact with our thousands of associates worldwide in order to streamline the content delivery to every user. We will also provide a newly designed update to the training session with new skills and options to focus on specific skills. A lot more syntactic and semantic data and filtering of sentences based on this data will be available, so that you can target specific sentence structures.

I am not sure I feel like paying this company anymore, at least not now, because the most interesting stuff (tons of new content at the B1-C1 levels, and a pronunciation checking tool) are yet to be released.

My hesitation is not just about an annoying weird person, who doesn't understand questions and assumes every user must be an idiot, or a toddler. But Glossika method's focus seems to be much more about listening, than speaking. I don't really need that. Glossika is preparing to lauch a pronunciation (or rather intonation) checker, but is at the same time arguing how useless it is anyways. The SRS based primarily on the data from all the users does have some advantages, but isn't that appealing to me. I am more after individualised tools, not ones assuming everybody is the same. I still haven't gotten any answer as to "How does Glossika tell the difference between successful and unsuccessful repetitions". Really, the only way to make a mistake is to make a typo, you cannot make are real mistake based on failing to recall something, so how can Glossika adapt the general pattern to each learner.

[PeterMollenburg]

-are you content with Glossika, if you've tried it?

I used the old version of Glossika (so I'm not sure how useful my comments will be if you're considering the newer format), completing all 3 levels of French and the first level of Dutch.

I used it more often than not as an audio only based listen and repeat programme to enhance my output and did this while commuting (driving). At other times I'd read the sentences either before or after listening to the audio when I had time to review, or in order to go back and refer to specific written sentences where I wasn't 100% sure what I was hearing was correct or to clarify unknown words or phrases.

I'd rate it 6/10. Too many errors for my liking. It's useful, but I wouldn't recommend it to a beginner as you could seriously solidify errors if this was your go to programme. I don't know if the newer version with the subscription format has improved much, or in other words, I don't know whether those errors have been ironed out. It's great that it has such a wide array of content or varied phrases and sentences, but I became a little nervous about it, given I'd spotted semi-frequent errors in the French content (I used it when I was well beyond the beginner learner stages) and therefore wasn't sure with the Dutch content whether I could trust the material as being accurate/correct.

Errors included the complete wrong phrase as if a few English/Dutch, English/French sentences had gotten mixed up before the final version was published, major issues with the IPA transcripts (in the French version I used, you could not rely on the IPA, as there from memory were errors in at least 25% of sentences), and clunky translations that seemed foreign or unnatural at times.

-what do you like about it? What is it good for, and what is it simply not meant to do?

Given the above hesitation, if errors are common with all the languages and have not been ironed out, I'd recommend Glossika for intermediate learners who want to practise output through repetition but know when something doesn't sound right, therefore they won't just accept what they hear as correct speech.

If the errors are largely gone, I'd recommend it to serious beginners and intermediate learners who need to push their output but don't have access to or don't want to converse with other people at this stage. Perhaps it could be a stepping stone before using italki or convesation partners to help the learner prime their output by learning a good deal of pre-recorded phrases.

-have you noticed some results?

Through repetition of the audio, confidence with speaking improved a little, especially with longer more complex sentences in the Dutch version (I was already more advanced in French) than what I was used to up to that point.

-is there any grading mechanism or feedback to your pronunciation? human checked (like Speechling), or software checked? Is it reliable?

This likely applies to the newer version. I'd be keen to read other peoples' thoughts and experience....

[Cavesa]

Thanks for a great reply as usual, PM!

It really looks like Glossika has evolved a lot. Which explains the lack of relevant reviews in the usual places, they are mostly based on something no longer offered.

The use to get more input and output (!) looks like the primarily intended one. I'd say it's making me consider Glossika as probably great for the less popular languages, but not much for the big ones. I miss mainly the Active Recall exercises, just repeating doesn't justify the price tag imho.

The errors are a very good point. I wonder, whether it is a big problem. It may be partially solved, Glossika seems to be hiring a lot of natives to participate in the "Viva" project. I haven't digged much into it, it looks like a lot of new content, checked by natives, which is a good thing. But it is not too clear. Yes, you're right that it is hard to trust something, when we can clearly see the problems in an already known language.

An intereting point raised by the S. person from the support, I'll paraphrase "you should use the English base, to get the best quality and amount of content for the target language". I am too lazy to look up the quote right now, but it was clear. Were you using it with the English base? And in any case, I find it slightly disgusting, when you build a part of your marketing on the "learn any language from any language, we are the biggest" but know that every base language but English hasn't been given the proper care.

Given the above hesitation, if errors are common with all the languages and have not been ironed out, I'd recommend Glossika for intermediate learners who want to practise output through repetition but know when something doesn't sound right, therefore they won't just accept what they hear as correct speech.

If the errors are largely gone, I'd recommend it to serious beginners and intermediate learners who need to push their output but don't have access to or don't want to converse with other people at this stage. Perhaps it could be a stepping stone before using italki or convesation partners to help the learner prime their output by learning a good deal of pre-recorded phrases.

Thanks, this is an excellent recommendation.

I'd say we are seeing another case of "good content, bad method". That's something I've heard even about Assimil actually . A ton of sentences to repeat after audio and copy by typing is nice. At least if most mistakes have been corrected. But it is very far from a complete course, very far from the "you don't need other stuff, such as grammar explanations, or pronunciation feedback".

The most positive thing I see about it: the message "you should say everything out loud". But that's doable with other tools too.

Through repetition of the audio, confidence with speaking improved a little, especially with longer more complex sentences in the Dutch version (I was already more advanced in French) than what I was used to up to that point.

Sounds good. But I am not sure it is good enough, for the price tag. What I find a bit sad is how much better could the tool be, if it actually went just a bit further. It requires typing (the new version), but not active recall. It requires speech recording, but gives no feedback (and makes even self-evaluation slow and not too obvious). It offers many language combinations, but is sloppy with them compared to the English base.

[PeterMollenburg]

Thanks for a great reply as usual, PM!

You're welcome, Cavesa. I tend to agree with all of your conclusions.

I agree that although they offer multiple languages which from a sales perspective might sound inviting for someone considering Glossika, but for someone learning one language and even more so one of the popular languages, it's a lot of extras of no use, or worse, simply distracting extras.

Agreed also that it would be better served for rarer languages. I have it on my list of Norwegian resources to perhaps someday try it out.

It is definitely disapointing that English-based content is more widespread on their platform than other base languages, but hopefully that will change in due course, but unlikely given English has the most reach in terms of attracting new customers. Still, personally I find it a little disappointing, although the content is still accessible at least albeit in English only perhaps.

I used English based content for French and Dutch learning. Should I use the subscription based platform in future, I'll definitely be taking a good look at French or even Dutch as potential base languages for Norwegian and should one or both base languages be available that would certainly encourage me to fork out the cash-ola and buy a subscription. Even if learning one language, redoing all available Norwegian activities in French and Dutch would reinforce the new Norwegian phrases all the while practising some French and Dutch. It'd be like SRS'ing everything three times but from different base languages.

-----------------------------------------------------------------
Edit: I just looked and discovered that the base languages are more limited than I realised. Anyway, I compared EN and FR base languages for learning both Norwegian and Dutch. The number of sentences and words available for learning is higher for Dutch than it is for Norwegian but it doesn't change (if their numbers are adjusted depending on the base language) regardless of whether learning from an EN or FR base. Additionally both the NL and NO courses supposedly take the learner to C1 despite a significant difference in the overall words and sentences one can learn in each of these languages on the platform (NO = 3,800+ sentences, 2,500+ words, NL = 6,700+ sentences, 3,000+ words). Seems questionable.
-----------------------------------------------------------------

"You should say everything aloud" is something I strive to do anyway with any language learning content, as you have also said is possible Cavesa with other resources. Still, as far as I used it for audio drills akin somewhat to SRS recall 'quizes' (ie I'd listen to the English sentence and attempt to say out loud the French or Dutch equivalent before hearing it on the recording), it's quite expansive. It really did go where other audio programs I'd used had not. My prior experience up to that point with audio based content in French included Pimsleur 1-5, Michel Thomas (all levels), Rocket French (all levels), Paul Noble, French All Talk by Linguaphone and other programmes that aren't necessarily strictly audio such as Fluenz and FSI. Despite the errors it was unique enough to be worthwhile with a wide variety of phrases, but still best used for rarer languages. Still it's useful for any language, despite the errors if you're looking for help with output.

[Cavesa]

It is definitely disapointing that English-based content is more widespread on their platform than other base languages, but hopefully that will change in due course, but unlikely given English has the most reach in terms of attracting new customers. Still, personally I find it a little disappointing, although the content is still accessible at least albeit in English only perhaps.

That's not the core of what I find disappointing. Openly saying "English base brings the most users, so we don't offer X and Y base for Z or W language", that would be totally fair.

But Glossika is making the "learn any language from any language" part pretty important in their marketing. And it is technically true, you will get content, if you click on the other options. But you need to dig a bit deeper, try them out, or even ask about them, to get the "yeah, but they are worse than the English base" truth.

The dishonest marketing is what I dislike. It is not an open lie, but it is a clear suggestion that everybody will finally get equal treatment as the privileged learners, and you discover only too late, after the payment and some use, that it is not so. I highly doubt that you'll get your money back, if you find out that a course X from Y is significatnly smaller, than the primarily advertised X from English.

I used English based content for French and Dutch learning. Should I use the subscription based platform in future, I'll definitely be taking a good look at French or even Dutch as potential base languages for Norwegian and should one or both base languages be available that would certainly encourage me to fork out the cash-ola and buy a subscription. Even if learning one language, redoing all available Norwegian activities in French and Dutch would reinforce the new Norwegian phrases all the while practising some French and Dutch. It'd be like SRS'ing everything three times but from different base languages.

But they are available. On the surface, it is awesome. But what Glossika doesn't tell you right away:
1.will you get the same amount of content in each of the base languages?
2.And are you likely to encounter more errors in for example the French to Norwegian course, compared to the English base?

As I said, the problem is not a visible lack of some of the base languages. It is the varying quality, and the lack of information before you actually try.

-----------------------------------------------------------------
Edit: I just looked and discovered that the base languages are more limited than I realised. Anyway, I compared EN and FR base languages for learning both Norwegian and Dutch. The number of sentences and words available for learning is higher for Dutch than it is for Norwegian but it doesn't change (if their numbers are adjusted depending on the base language) regardless of whether learning from an EN or FR base. Additionally both the NL and NO courses supposedly take the learner to C1 despite a significant difference in the overall words and sentences one can learn in each of these languages on the platform (NO = 3,800+ sentences, 2,500+ words, NL = 6,700+ sentences, 3,000+ words). Seems questionable.
-----------------------------------------------------------------

THANK YOU! You've found a lot of info not findable elsewhere. Btw, where do you find the number of words or sentences? I was searching and failed to find it.

Well, the varying numbers are a part of the issue. The fact there are levels up to C1 in two languages doesn't mean both courses cover the levels equally well. I'd say nobody here is foolish enough to believe that Norwegian is twice easier than Dutch for an English speaker, and you'll get equal value and equally solid progress from just half the sentences.

For the base comparison: I assume (based on experience with such tools), that the difference between the En and Fr base is likely to be much smaller, than between En and Cz or En and Turkish. The people more in need of such a service get constantly worse service, due to being seen as not an important market.

"You should say everything aloud" is something I strive to do anyway with any language learning content, as you have also said is possible Cavesa with other resources. Still, as far as I used it for audio drills akin somewhat to SRS recall 'quizes' (ie I'd listen to the English sentence and attempt to say out loud the French or Dutch equivalent before hearing it on the recording), it's quite expansive. It really did go where other audio programs I'd used had not. My prior experience up to that point with audio based content in French included Pimsleur 1-5, Michel Thomas (all levels), Rocket French (all levels), Paul Noble, French All Talk by Linguaphone and other programmes that aren't necessarily strictly audio such as Fluenz and FSI. Despite the errors it was unique enough to be worthwhile with a wide variety of phrases, but still best used for rarer languages. Still it's useful for any language, despite the errors if you're looking for help with output.

Yes, I'd agree with that, even though I somewhat fail to see the huge difference. Not until there is any kind of feedback. Even the auto-feedback, by playing your recording immediately to you (instead of hiding it on a separate page) would already do what you say. As it is now, you are demanded to record the output but that's it (if you cough instead, you'll move on in the repetitions too), it's not that much of a success.

[PeterMollenburg]

It is definitely disapointing that English-based content is more widespread on their platform than other base languages, but hopefully that will change in due course, but unlikely given English has the most reach in terms of attracting new customers. Still, personally I find it a little disappointing, although the content is still accessible at least albeit in English only perhaps.

That's not the core of what I find disappointing. Openly saying "English base brings the most users, so we don't offer X and Y base for Z or W language", that would be totally fair.

But Glossika is making the "learn any language from any language" part pretty important in their marketing. And it is technically true, you will get content, if you click on the other options. But you need to dig a bit deeper, try them out, or even ask about them, to get the "yeah, but they are worse than the English base" truth.

The dishonest marketing is what I dislike. It is not an open lie, but it is a clear suggestion that everybody will finally get equal treatment as the privileged learners, and you discover only too late, after the payment and some use, that it is not so. I highly doubt that you'll get your money back, if you find out that a course X from Y is significatnly smaller, than the primarily advertised X from English.

Ah, yes, I see what you mean. Agreed, this is dishonest marketing.

I used English based content for French and Dutch learning. Should I use the subscription based platform in future, I'll definitely be taking a good look at French or even Dutch as potential base languages for Norwegian and should one or both base languages be available that would certainly encourage me to fork out the cash-ola and buy a subscription. Even if learning one language, redoing all available Norwegian activities in French and Dutch would reinforce the new Norwegian phrases all the while practising some French and Dutch. It'd be like SRS'ing everything three times but from different base languages.

But they are available. On the surface, it is awesome. But what Glossika doesn't tell you right away:
1.will you get the same amount of content in each of the base languages?
2.And are you likely to encounter more errors in for example the French to Norwegian course, compared to the English base?

As I said, the problem is not a visible lack of some of the base languages. It is the varying quality, and the lack of information before you actually try.

Yep, I see what you mean here too. While they certainly do what many other language learning resources don't do (in particular their coverage of rarer languages), their platform could be so much better in so many ways. I almost want to take their design and improve on it 1000%, but... I don't have time for that. Still, at least they exist than nothing at all, but yes it could be better in many ways, and one would could assume (but might be wrong in their assumption potentially) that with a little work a much more well rounded product could be developed.

-----------------------------------------------------------------
Edit: I just looked and discovered that the base languages are more limited than I realised. Anyway, I compared EN and FR base languages for learning both Norwegian and Dutch. The number of sentences and words available for learning is higher for Dutch than it is for Norwegian but it doesn't change (if their numbers are adjusted depending on the base language) regardless of whether learning from an EN or FR base. Additionally both the NL and NO courses supposedly take the learner to C1 despite a significant difference in the overall words and sentences one can learn in each of these languages on the platform (NO = 3,800+ sentences, 2,500+ words, NL = 6,700+ sentences, 3,000+ words). Seems questionable.
-----------------------------------------------------------------

THANK YOU! You've found a lot of info not findable elsewhere. Btw, where do you find the number of words or sentences? I was searching and failed to find it.

If you go to their website (I did this both with my phone and with my laptop), select a language and scroll down past a lot of the wordy sales pitch stuff you'll find a section stating A1 to C1 for example and next to it the number of sentences and words Glossika offers.

Well, the varying numbers are a part of the issue. The fact there are levels up to C1 in two languages doesn't mean both courses cover the levels equally well. I'd say nobody here is foolish enough to believe that Norwegian is twice easier than Dutch for an English speaker, and you'll get equal value and equally solid progress from just half the sentences.

Agreed.

For the base comparison: I assume (based on experience with such tools), that the difference between the En and Fr base is likely to be much smaller, than between En and Cz or En and Turkish. The people more in need of such a service get constantly worse service, due to being seen as not an important market.

Oh, and the way I understand it, the content is available only in these base languages: English, French, Mandarin, Russian, as these are the languages listed at the top right of the language drop-down menu to run the website. Seems odd why you wouldn't have some more such as German, Spanish or Japanese, but I guess four base languages is better than one. As for Turkish and Czech or other such smaller languages, I don't believe such languages are offered as base languages, again due to their drop down menu being only in four languages, so by all means point me in the direction which indicates otherwise if you've discovered more base languages.

"You should say everything aloud" is something I strive to do anyway with any language learning content, as you have also said is possible Cavesa with other resources. Still, as far as I used it for audio drills akin somewhat to SRS recall 'quizes' (ie I'd listen to the English sentence and attempt to say out loud the French or Dutch equivalent before hearing it on the recording), it's quite expansive. It really did go where other audio programs I'd used had not. My prior experience up to that point with audio based content in French included Pimsleur 1-5, Michel Thomas (all levels), Rocket French (all levels), Paul Noble, French All Talk by Linguaphone and other programmes that aren't necessarily strictly audio such as Fluenz and FSI. Despite the errors it was unique enough to be worthwhile with a wide variety of phrases, but still best used for rarer languages. Still it's useful for any language, despite the errors if you're looking for help with output.

Yes, I'd agree with that, even though I somewhat fail to see the huge difference. Not until there is any kind of feedback. Even the auto-feedback, by playing your recording immediately to you (instead of hiding it on a separate page) would already do what you say. As it is now, you are demanded to record the output but that's it (if you cough instead, you'll move on in the repetitions too), it's not that much of a success.

Yes, agreed. While I found the old Glossika to be somewhat unique and not a carbon copy of other audio programmes out there (for example, Michel Thomas French and Paul Noble French are almost identical to each other in format and content), it's not that different to be something to rave about. I actually found a lot of the audio phrases to be similar in nature to audio found in Fluenz French in terms of sentence length and complexity (not the actual phrases). However, Fluenz has the audio recording with some kind of voice recognition technology that will mark you incorrect if your recording of your own voice repeating certain phrases doesn't match that of the pre-recorded Fluenz one sufficiently close (including intonation). Fluenz is well put together but doesn't advance in difficulty as much as it could from the first level to the last.