Language Transfer

[Expugnator]

Let's see it on the positive side. Once they are done with those FIGS there will be mostly cooler languages to be added. The sooner they release those FIGS courses, the sooner we'll have unique material. We should bear in mind that Language Transfer fills a marketing segment that so far has consisted solely of expensive commercial courses such as MT and Pimsleur, so it's natural that there is a high demand for FIGS courses as well.

[Random Review]

Anyway, it's not all bad news :

[Cavesa]

I really like the schedule on dashboard. Sure, it cannot be precise but it is a nice overview. I trully hope German gets finished this year. Next year, I hope to be too advanced to use it

And there is the Swahili course being prepared! I am happy for the Swahili learners!

[Expugnator]

Both Georgian and Papiamentu got voted, and it wasn't me! (I would have voted for Estonian).

One can always dream...

[Cavesa]

So, I am stalking LT a bit for new German lessons (I don't want to start the course until it looks like it might trully be complete in future) and I am slightly disappointed. The dashboard overview was updated and it is clear German is the least important project. It won't be finished this year unlike the Introduction Italian (which will be finished in September, it seems, so at least some good news, even though for someone else than me). The Complete Greek is obviously more of a priority, which I understand (and the course tempts me to start Greek ). And perhaps Swahili will be too. But I suppose we might perhaps get new German lessons, once all Complete Romance Languages get finished.

Btw it looks like the next course might be introduction to Mandarin. That sounds great. Even though Complete German looks even further away now

[Brun Ugle]

I know. I've been waiting for the German course a long time. At this point, I think I'll already know German by the time he's finished.

[Cavesa]

I know. I've been waiting for the German course a long time. At this point, I think I'll already know German by the time he's finished.

Considering my German pace, I might not be that advanced at that moment. But perhaps I'll have learnt a few more languages by then!

[Cainntear]

...

Seneca asked me my opinions of LT recently, so I had a quick look at a couple of lessons to try and get an initial impression. First up, I tried the German course, as that's a language on my started-but-never-mastered list. I bailed almost immediately when he pronounced "lernen" almost as though it was an English word, that is with ER rendered as a single vowel phoneme, rather than a vowel plus a consonant. (I'm aware that ER occurs in German as a single phoneme as well, but that's in word-final position, and it's a radically different phoneme from the English one, so this is doubly wrong.)

Michel Thomas may have had a horrible accent in all of his courses, but accent is only one component of pronunciation, and a superficial one at that.

I have long been a fan of this guy, I did his early Greek course, which is basically Michel Thomas with a few minor improvements. That in itself is no small thing! Everyone else who tried to continue what MT was doing (whether Hodder & Stoughton, Harold Goodman, Paul Noble, etc) missed the point completely!

He still makes some of the classic errors though -- he starts the Spanish course, for example, with an overly long section on cognates, leading to a pretty scattergun selection of unrelated vocabulary before starting to form phrases and sentences. I call this a "classic" because pretty much every post-MT MT course seemed to start with this. (With the exception of the Chinese one which started in an even less MT way -- coloured thumbs and saying "dog" with different tones...)

Among the small improvements were Mihailis' own command of the languages he teaches (better than MT and with a much better accent), only having one student (I don't know why that proved to be an improvement, but it did) and smarter editing (unless I am mistaken, he sometimes left a correction in but removed the mistake that prompted it, thus keeping the student feedback element without the listener having to hear every single mistake made by the student on the audio).

Well, in the Spanish course, his "student" appears to have a better command of the language than he does -- her accent is pretty damn good. And then after the better speaker says the answer, he repeats it in a worse accent, and the listener will incorrectly assume that his model is the one to follow, rather than hers.

The other thing about this is that it calls into question the validity of any errors and correction. When someone pronounces "fatal" like a native, right down to the point of articulation of the /l/ phoneme, then erroneously includes a y-glide in "natural", you can be pretty sure it was a purposefully included "error". The effectiveness of mistakes in MT's course derived from the fact that learners' mistakes are actually pretty systematic, and a learner following an MT course is likely to make similar mistakes to the students on the recording (in fact, I remember being stunned at how frequently I made the same mistake as the guy on the Spanish recordings to start off with) and so you need to be dealing with genuine mistakes, not scripted ones.

Think of it this way: if students tend to make the same errors, there's something missing in your teaching. If you can identify it, you should eliminate it, meaning no need for correction. If you can't identify it, how do you know what the optimal place is for a fake error and correction?

To do all that and provide these courses free is already really good,

There's a problem with free: it takes time and money to make a good course, and you just can't put that in when you're working for free, and even Patreon donations can't change that. I'm very dubious of the ability to create a coherent single course if you're doing it in stages, and he's looking for Patreon funds even now to finish his Greek course, which should be the easiest one for him.

The reason that this is an issue for me is that if he hasn't finished it, it means he can't have tested it, and therefore he's committing to tape* something he can't know works. (*Yeah, I know, not really tape, but "committing to digital file" doesn't really sound right.) Once it's recorded, it's fixed, and the opportunity for improvement goes out the window.

Another example of this is SaySomethingInWelsh who've been "working on" an improved version of their original course for almost a decade, but still haven't finished the new version, so the original course is still the main product. They're struggling for cash, and they're not even entirely free. What's happened with them is that they've moved on to a subscription system and people are paying more over time than they would for a similar one-off-purchase product. They're happy to do this, because they see themselves as contributing to the future of the language learner community, but in the end, even with all of them paying over the odds, and some of them offering free labour to the company, it's still not enough to get the additional material funded, completed and out the door.

LT's Patreon campaign and community model has unintended consequences in terms of buy-in:

Let's see it on the positive side. Once they are done with those FIGS there will be mostly cooler languages to be added. The sooner they release those FIGS courses, the sooner we'll have unique material. We should bear in mind that Language Transfer fills a marketing segment that so far has consisted solely of expensive commercial courses such as MT and Pimsleur, so it's natural that there is a high demand for FIGS courses as well.

Community buy-in leads people to make excuses that they wouldn't accept from a commercial outfit. If I pay for something, I want to get it. I am not paying for maybe getting something, but more likely having my money spent on buying something someone else wants.

I've already bought MT German, Italian and Spanish. I see no real need to make any of the FIGS on LT -- anyone who wants to can get MT from a bookshop, a library, ebay or a torrent site. Opening things to the world is great, but an LT course based on English isn't much good to a Mongolian -- it mostly serves people who can afford commercial materials, and commercial materials are one important way for language professionals to make a living.
I will not pay money for someone to undercut the living of my colleagues, friends and acquaintances in order to serve people primarily in wealthy countries by replicating what is already available, just for free.

But Expugnator's defence is... interesting. It boils down to "pay him for something you don't want so that he finishes it so that he might do the one you want later"... but then you're in an everlasting loop of paying in in the hope that at the next vote you'll get what you want, but not doing, and being encouraged to keep paying in so that you can have your say next time. And you never get what you want.

[Random Review]

...

Seneca asked me my opinions of LT recently, so I had a quick look at a couple of lessons to try and get an initial impression. First up, I tried the German course, as that's a language on my started-but-never-mastered list. I bailed almost immediately when he pronounced "lernen" almost as though it was an English word, that is with ER rendered as a single vowel phoneme, rather than a vowel plus a consonant. (I'm aware that ER occurs in German as a single phoneme as well, but that's in word-final position, and it's a radically different phoneme from the English one, so this is doubly wrong.)

Michel Thomas may have had a horrible accent in all of his courses, but accent is only one component of pronunciation, and a superficial one at that.

I have long been a fan of this guy, I did his early Greek course, which is basically Michel Thomas with a few minor improvements. That in itself is no small thing! Everyone else who tried to continue what MT was doing (whether Hodder & Stoughton, Harold Goodman, Paul Noble, etc) missed the point completely!

He still makes some of the classic errors though -- he starts the Spanish course, for example, with an overly long section on cognates, leading to a pretty scattergun selection of unrelated vocabulary before starting to form phrases and sentences. I call this a "classic" because pretty much every post-MT MT course seemed to start with this. (With the exception of the Chinese one which started in an even less MT way -- coloured thumbs and saying "dog" with different tones...)

Among the small improvements were Mihailis' own command of the languages he teaches (better than MT and with a much better accent), only having one student (I don't know why that proved to be an improvement, but it did) and smarter editing (unless I am mistaken, he sometimes left a correction in but removed the mistake that prompted it, thus keeping the student feedback element without the listener having to hear every single mistake made by the student on the audio).

Well, in the Spanish course, his "student" appears to have a better command of the language than he does -- her accent is pretty damn good. And then after the better speaker says the answer, he repeats it in a worse accent, and the listener will incorrectly assume that his model is the one to follow, rather than hers.

The other thing about this is that it calls into question the validity of any errors and correction. When someone pronounces "fatal" like a native, right down to the point of articulation of the /l/ phoneme, then erroneously includes a y-glide in "natural", you can be pretty sure it was a purposefully included "error". The effectiveness of mistakes in MT's course derived from the fact that learners' mistakes are actually pretty systematic, and a learner following an MT course is likely to make similar mistakes to the students on the recording (in fact, I remember being stunned at how frequently I made the same mistake as the guy on the Spanish recordings to start off with) and so you need to be dealing with genuine mistakes, not scripted ones.

Think of it this way: if students tend to make the same errors, there's something missing in your teaching. If you can identify it, you should eliminate it, meaning no need for correction. If you can't identify it, how do you know what the optimal place is for a fake error and correction?

To do all that and provide these courses free is already really good,

There's a problem with free: it takes time and money to make a good course, and you just can't put that in when you're working for free, and even Patreon donations can't change that. I'm very dubious of the ability to create a coherent single course if you're doing it in stages, and he's looking for Patreon funds even now to finish his Greek course, which should be the easiest one for him.

The reason that this is an issue for me is that if he hasn't finished it, it means he can't have tested it, and therefore he's committing to tape* something he can't know works. (*Yeah, I know, not really tape, but "committing to digital file" doesn't really sound right.) Once it's recorded, it's fixed, and the opportunity for improvement goes out the window.

Another example of this is SaySomethingInWelsh who've been "working on" an improved version of their original course for almost a decade, but still haven't finished the new version, so the original course is still the main product. They're struggling for cash, and they're not even entirely free. What's happened with them is that they've moved on to a subscription system and people are paying more over time than they would for a similar one-off-purchase product. They're happy to do this, because they see themselves as contributing to the future of the language learner community, but in the end, even with all of them paying over the odds, and some of them offering free labour to the company, it's still not enough to get the additional material funded, completed and out the door.

LT's Patreon campaign and community model has unintended consequences in terms of buy-in:

Let's see it on the positive side. Once they are done with those FIGS there will be mostly cooler languages to be added. The sooner they release those FIGS courses, the sooner we'll have unique material. We should bear in mind that Language Transfer fills a marketing segment that so far has consisted solely of expensive commercial courses such as MT and Pimsleur, so it's natural that there is a high demand for FIGS courses as well.

Community buy-in leads people to make excuses that they wouldn't accept from a commercial outfit. If I pay for something, I want to get it. I am not paying for maybe getting something, but more likely having my money spent on buying something someone else wants.

I've already bought MT German, Italian and Spanish. I see no real need to make any of the FIGS on LT -- anyone who wants to can get MT from a bookshop, a library, ebay or a torrent site. Opening things to the world is great, but an LT course based on English isn't much good to a Mongolian -- it mostly serves people who can afford commercial materials, and commercial materials are one important way for language professionals to make a living.
I will not pay money for someone to undercut the living of my colleagues, friends and acquaintances in order to serve people primarily in wealthy countries by replicating what is already available, just for free.

But Expugnator's defence is... interesting. It boils down to "pay him for something you don't want so that he finishes it so that he might do the one you want later"... but then you're in an everlasting loop of paying in in the hope that at the next vote you'll get what you want, but not doing, and being encouraged to keep paying in so that you can have your say next time. And you never get what you want.

There are a number of misconceptions in this critique.

1) He may seem to you to be committing "classic" errors that everyone emulating MT commits, but he is no longer emulating MT, he has moved on from that. The way the new Greek course works is based throughout on building connections between the building blocks of words and I imagine he will have been doing the same in his German course (in fact German is the perfect language for that!). His approach to cognates is not the usual one of looking at correspondences; instead he draws you into seeing the connections between words (both within the TL and between the TL and English). In contrast the old Greek course was basically an attempt to continue where MT left off and contains far fewer cognates.

2) You are wrong that he is not testing his courses. It is true that his funding model leads to the less-than-ideal need to create courses in stages; but each stage is tested with real students and then revised before the final version is recorded. He asks for volunteers for this.

3) I do not believe the errors are scripted. Unlike MT, most of the students are not native English speakers and so they may indeed nail some parts that are very difficult for English speakers and find other parts difficult. I'd actually be more suspicious if they didn't.

4) I checked out a little bit of his German course after reading your review. You are fine to go ahead with it IMO. His pronunciation is not perfect, but it is at least as good as MT's Spanish pronunciation TBH. All the audios are checked by native speakers before being published and any important errors corrected (I tried a bit of the Turkish course and this was sometimes quite amusing). They must have decided that this error is not a bad one. The wonderful thing about the nature of his community model is that if I am wrong and it is an important error, however, you can contact him and enable him to put it right!

5) Regarding his funding model. You know I think you are wrong about it only benefiting people in rich countries, it is truly amazing how many members of the middle class in developing countries get themselves some level of English by hook or by crook. They can benefit from this. Also poor people in rich countries can benefit. I know at first hand just how little disposable income you have when you are washing dishes, cleaning offices, etc., because I did it for years (I say "little" and not "zero" because I was a single guy with no kids). I don't really like the fact that you are dismissing the fact that this project can provide access to information for people like I was TBH.

Now the poor people in developing countries, who are of course the ones who need it most, won't benefit. About that you are surely right.
If you have any better ideas that would help them, you should (of course) do so. God knows that no-one else is. Meanwhile I don't think you should hold that against Language Transfer TBH. It's a great wee project, but no one is claiming it's a panacea for the language ills of the world. Everyone who contributes to his funding campaign knows how it works and does so of their own free choice.

[Seneca]

Cainntear, thanks for taking the time to share your thoughts. Looks like you are still on Team MT in this comparison

I am beginning the active wave of Assimil right around when I finish up a current vacation I am on. I have been rather lazy in doing much beyond the daily Assimil lesson since starting due to the pace of life at the moment I definitely think doing either LT or MT for Spanish right when I get home and trying to get into speaking with people immediately would be a good plan for "activating" my skills.