There's the CD 'An Irish Lullaby/Suantraí' by Pádraigín Ní Uallacháin which you can now download on the net, apparently it's still available as a CD from Ossian USA and there's her other CD 'A Stór is a Stóirín' which you get directly from her own website http://www.irishsong.com

This one was written by Rúairí Ó hUallacháin -

Suantraí Sí

Suantraí sí a linbhín luasc go mall sa chliabhán
Lú lá luí a linbhín dún do shúil a naíonáin

Seoithín seoithín seoithín seó
Seoithín seoithín seoithín seó
Seoithín seoithín seoithín seó
Suantraí sí a linbhín a thaiscidh a stór

Luí go socair luí go ciúin codladh sámh a ghrághil
Fan id' shuan a thaiscidh buan go n-éirí tá ar maidin.

PS I think there must be a mistake there - shouldn't it be - go n-éirí tú ar maidin ?

You say that as if there are these two things and nothing else, and you're contrasting "activeness" with "rote", which is too narrow a perspective. (And you're also suggesting that task-based activities aren't games.)

Tasks are useful if and when they promote meaningful engagement with the language, but rote vs meaningful is not "digital", there's a spectrum, and most tasks aren't 100% meaningful. If the task is "get X" when I say "hurdi-gurdi-hurdi-gurdi X", there is some meaningfulness in the task, but I don't know whether hurdi-gurdi-hurdi-gurdi is a polite request, an absolute order, or anything in between.

In the end, such a long phrase would be memorised by rote, and there would be little variation in the language.

And that's the key word in language learning: variation. If you use a language feature in one way only, you learn it by rote. To learn it meaningfully, you need to learn it in various forms and circumstances, understanding how it influences the meaning.

Duolingo are kind of close to the right path on this. They have a reasonable level of variation -- in fact, in the languages that use text-to-speech for voices, they have potentially unlimited variation. Unfortunately, I think they've let their data lead them astray. They've no doubt spotted that people get more right answers if they repeat the same example phrases, and therefore concluded that they're "learning" better. No, all they've proven is that their teaching wasn't good enough to beat rote in the short-to-medium term.

(I believe DuoLingo's problem is their reluctance to explicitly present the rules, relying instead on learners intuiting them. But you can only intuit from variation, and if variation isn't working for your students and you're forced to fall back on rote, this means that intuiting rules doesn't work.)

_________________
A language belongs to its native speakers, and when you speak it, you are a guest in their homes.
If you are not a good guest, you have no right to complain about receiving poor hospitality.

Implicit teaching is fine, as research shows people can intuit rules by picking out statistical features and reading lots helps great in this buthumans are conscious creatures, so there needs to some explicit teaching as well as making clear the context of use.

For example, with the subjunctive in fan go dtaga mé that it is/was used to show the will of the speaker in a wish that is somewhat beyond the power of the said speaker (a curse or a blessing or a request like above that would require an intervening event or state)

Knowing when (social context), the function (curse, blessing, restricted command), plus the form (Verb 1 + Verb 2), and how to generate it (bare Verb 1) + copula + (eclipsed Verb 2 and final schwa) would help make it stick, and then the learner could be shown examples of it in action

_________________
__̴ı̴̴̡̡̡ ̡͌l̡̡̡ ̡͌l̡*̡̡ ̴̡ı̴̴̡ ̡̡͡|̲̲̲͡͡͡ ̲▫̲͡ ̲̲̲͡͡π̲̲͡͡ ̲̲͡▫̲̲͡͡ ̲|̡̡̡ ̡ ̴̡ı̴̡̡ ̡͌l̡̡̡̡.___

I also wonder, as an aside, how long it too children to master traditional Irish sounds when speakers regularly used up to 50 consonants?

_________________
__̴ı̴̴̡̡̡ ̡͌l̡̡̡ ̡͌l̡*̡̡ ̴̡ı̴̴̡ ̡̡͡|̲̲̲͡͡͡ ̲▫̲͡ ̲̲̲͡͡π̲̲͡͡ ̲̲͡▫̲̲͡͡ ̲|̡̡̡ ̡ ̴̡ı̴̡̡ ̡͌l̡̡̡̡.___

Other research says people aren't actually very good at this at all. In fact, in Finland, they're pretty convinced that trying to learn from exposure is a waste of time, because Finnish is so different from other European languages.

Mainstream ideas in Western European language teaching are all coloured by the strong similarities between the usual suspects: English, German, French, Spanish and Italian. Bundle in everything in western Europe, and the only genuinely unusual languages are the Celtic ones and Basque. Intuition/assimilation/exposure works reasonably well in such situations.


Actually, the problem is that we are not always conscious creatures. The brain makes shortcuts, and doesn't notice minor details if the full meaning is understood -- that's how we survive accent and dialectal differences.

I once had a couple I was teaching privately, and we did a listening lesson. The man in the recording was talking about "house prices". They said "prices of houses". I repeated it another three or four times, and every time they were convinced they had heard "prices of houses".

You can't learn from input if your brain alters the input.

_________________
A language belongs to its native speakers, and when you speak it, you are a guest in their homes.
If you are not a good guest, you have no right to complain about receiving poor hospitality.

I have someone like that in one of my classes. I will say "slán," for example, and she will repeat it back to me as "shlán." I'll model the correct sound (even exaggerate it), but as soon as he puts it with the world, the "s" turns slender again.

Redwolf

I had a similar issue myself with /L/ and /l/ and a native speaker trying to correct me. He'd model the pronunciation with /l/ and I would repeat with /L/. It wasn't until he told me, "You're saying /L/ not /l/" and wrote it out in the modified IPA that I got what he was trying to correct. I couldn't be "shown" by him saying it correctly as I was not hearing a difference. I had to be explicitly told what it was. I still could not look back in my memory and say, "Oh yes, I was doing that." I just had to accept that he was right and fix it.

This is getting a bit off topic, but when I'm teaching French children English (only using the target language) I am always careful to make sure they can pick up the basic sounds. For example, if I pronounce a word with a dipthong in it, they don't hear it and of course they can't go on to reproduce it accurately. We always 'decompose' these kinds of sounds at the beginning, so that they can take them in and following on from that, they know that they can pronounce them properly. It boosts their confidence in their language acquisition. This is often a problem when it comes to listening to Irish learning materials - either it's too fast or it's not clear enough, plus you hear people who are supposed to be from one part of Ireland speaking in a dialect from somewhere else, which can also be confusing. Forvo and the on-line dictionaries that give examples of the different dialect prononciations are vital in this respect. You know who's speaking and where they're from.

There is also the famous example of the child who had a toy fish and pronounced fish fisk. And this was from a college introductory linguistics text book, not just an apocryphal story. When adults would correct him he'd insist he was saying it correctly and if anyone else said fisk he would correct them, "It's not a fisk! It's a fisk!"

Yes I'm sure we all know of examples of children having set on their particular idea of how to pronounce or to use certain words in a way they have decided on that doesn't necessarily correspond to 'normal' usage. It's all part of a natural process of trying to make sense of language and creating their own logic about it - eventually they'll change and adapt this according to what they'll hear being spoken around them. Eg - 'I goed' instead of 'I went' or in French 'j'ai prendu' instead of 'j'ai pris' (and in our family - un 'tacski' for un taxi etc)