Relying romanization (Pinyin) is probably the only way at the beginning. I don’t think there is a way around it, learning the characters takes a long time.
Well, it depends on your definition of “possible”. If it means being able to focus your eyes on the currently spoken word; then I have been doing it for about a year, progress is noticeable but slow. The main stumbling block is the comprehension.
Personally, the idea of speeding up a Chinese language podcast sounds almost frightening. I can only speak for myself, but the mental load is already high enough. As an example: I’m listening to this podcast 无聊斋免费在线收听-喜马拉雅 , it’s listed under “comedy”, so my current goal in life is figuring out what these people are laughing about, I’m a couple episodes in but I’m still in the dark…
Frankly, the only reason I have been able to keep my sanity is that this podcast has the words DON’T PANIC inscribed in large friendly letters on its cover.
That sounds about what I would imagine. I have read at least one billion words in English. That familiarity with the English alphabet means there is no resistance to seeing it, even when the word is in an unfamiliar language.
All learners I know want to ditch pinyin and romaji as quickly as possible (or never use them). But I now wonder if there might be a benefit in R+L a lot longer than is expected. I would obviously 100% recommend some intensive reading without and study of characters for languages that use logograms, but if romanization can make (much) more audio accessible faster that seems like it could be a good thing.
Unfortunately the only scenario I would test it myself is if someone paid me to make it my job.
This does not work well for multiple speakers, but for a podcast with a single speaker (at least at once) I used to simply open a Google doc and set it to voice-to-text with the audio out being the source. That worked better than HappyScribe for the “En Svensk Tiger” and “Mytologier” podcasts once upon a time.
Who are you referring to when you say that there is an extreme school of thought that claims immersion is the only necessary component? Who advocates this?
Maybe the early Khatz (from AJATT), but I’ve never heard anyone claim that we shouldn’t speak at all. That doesn’t make any sense.
As I wrote in another comment, the question is rather
“outputting earlier or later” (and how “late” is late?)
On the one hand, I don’t think it matters if you want to master all four language skills.
On the other hand, “hyperfocusing on conversational fluency”, as Toby would say, has definitely its merits. That is, achieving conversational fluency early on can be a huge motivational boost, esp. for inexperienced language learners. For experienced language learners, it’s probably less interesting.
This is too widespread blame on individuals. There are large numbers of people online I come across all the time who believe all they have to do is sit and immerse and the speaking part will come easily and quickly when they start outputting months or years down the line.
If you doubt this scroll down to the comments below by @TofuMeow who writes “I totally agree at this point that immersion alone won’t get a person to outputting.” The use of the words “at this point” suggests at a previous point this learner had to some degree fallen for the immersion delusion - as many learners (including msyelf) have also done.
This is often due to misreadings of what influencers they follow actually think about language learning. For example, when asked by Olly Richards what the biggest misconception people had about him was, Matt Vs Japan answered that many people believed the transition from input to output happened overnight as if by magic when in fact it didn’t. Matt also covered this when I interviewed him for my podcast.
“I’m kinda curious if massive reading is more effective for speaking for languages like Spanish, German etc.” (@TofuMeow)
Massive reading definitely helps for speaking Germanic and Romance languages, i.e.:
your vocabulary and syntactic structures will become more sophisticated,
you’ll internalize more collocations so you can speak more naturally,
And if you combine reading with listening (using LingQ, ReadLang, etc. plus Audible, YT, etc.), you can read much faster and with more focused attention (it’s like a highly intense “word shower”).
However, if you want to learn to speak well, reading/listening to (audio) books is less effective. It’s better to read the scripts / subtitles of YT videos, TV series, audio drama, etc. while listening to them.
That said, practicing speaking (sooner or later) is still a must for anyone who wants to speak well.
Mutatis mutandis, it should be a similar experience for non-Indo-European languages…
But imitating young children 1:1 as adult learners makes little sense because mass immersion of native speakers, while effective, is also extremely slow and (time-)inefficient.
My favorite example, at least at the moment, is the following in this context (because I’m trying to reach an advanced level in Br. Portuguese):
"Let’s say a baby is born in a Portuguese speaking environment. As an educated and adult native speaker of German who wants to learn Portuguese from scratch, I could learn Portuguese
in ca. 1.5 years (investing about 2000-2500 hours = ca. 4-4.5 h per day ) to reach a C1 level in all four language skills
start studying computer science (or another subject I’m interested in) full time at a university in Portugal or Brasil
and get a degree
before the child as a native speaker is 6 years old and has a very limited vocabulary of about 7000 words!
I really love children, but when it comes to learning effectively and efficiently they’re out of the competition: not only in language learning, but in learning in general!" (PB quote).
Maybe we can rephrase Michilini’s “immersion delusion” here:
The problem is the thesis of Krashen, Steve & Co
that SLA is primarily unconscious learning, not conscious learning and
that conscious learning should be avoided because it’s ineffective.
However, this thesis is highly controversial in SLA research and not convincing with regard to artificial SRS (if it’s used with audio. cloze tests for everyday collocations, active recall, etc.).
Moreover, conscious learning based on SRS is much more time-efficient than natural SRSing by reading and / or listening.
The main benefits of reading and/or listening, on the other hand, are:
They’re much more interesting than using SRS.
There’s more context info available.
The combo “reading + listening” results in a higher degree of focused attention.
Ergo, hybridization of both approaches (i.e. natural SRS + artificial SRS) makes the most sense to me because that way I can get the best of both worlds.
“But it’s a bad idea to believe that more input alone will solve the problem as this doesn’t seem to be the case.”
I agree.
For example, if it’s a psychological issue (shyness, a perfectionist attitude, etc.), then it’s never the right time to start speaking. More and more and more input won’t solve this problem.
However, speaking early, making lots of mistakes and experiencing that it’s perfectly ok to make mistakes is definitely a more successful strategy in this context.
“they wanted a campaign to wage war on traditional methods. Anyone who questioned these ideas or supported methods became the enemy.”
I’m sorry, Michilini, you lost me here.
The common credo among probably all polyglots is:
“There are many ways to make an omelette!” (OrientalPearl).
But I’ve never met anyone online or offline who said that
there’s only one way in language learning and
“pure inputism” is the only way to language paradise.
Yes, there is Krashen’s approach, which emphasizes the importance of “unconscious” learning by focusing on input. But in SLA research, this claim is highly controversial.
By the way, Krashen has a great sense of humor.
And I can’t see an enemy here. In contrast, I like him a lot!
“Matt Vs Japan answered that many people believed the transition from input to output happened overnight”.
Digesting lots of compelling and comprehensible content definitely reduces the “collocation problem” that many L2 learners face.
In other words. when they spend hundreds or thousands of hours inputting, they get a pretty good sense of what correct L2 sentences should sound and look like!
This also means that they’ll find speaking much easier than learners who start speaking earlier.
However, if they don’t practice speaking, they will sink much faster than the Titanic in a fast-paced conversation with native speakers.
And if the avoidance of speaking is based on psychological issues (e.g., shyness, perfectionism, etc.), then more and more and more … input will never be enough to solve the underlying problem (that’s just a typical case of “avoidance behavior”!).
Re “the immersion delusion - as many learners (including msyelf) have also done.”
Sorry, mate, I can’t resist:
If you had read more SLA research then you would have known that Krashen’s ideas are highly contested
Apart from that, the “speaking first / early” approach is neither revolutionary nor new. Benny, for example, has been preaching it online for almost 15 years
And, finally, there are “no enemies” here (esp. not Krashen!)…
Matt said that Khalifa was a “solid B2”.
However, we don’t have enough background info to verify that.
Matt doesn’t speak Spanish (as far as I know) so he can’t judge.
It’s not that hard to practice “an introduction of yourself” and be pretty good at it. That’s what Benny has been preaching for years. Therefore, the significance with regard to the actual language level is pretty low.
Talking to your tutors isn’t a realistic “test” because tutors usually speak quite slow with a very clear pronunciation.
As I wrote before a few times, the “bar / restaurant, etc. test” is a much more realistic test scenario. That is:
If learners pass the “bar / restaurant, etc. test” (with background noise, contractions, slang, fast-paced conversations, jokes, unclear pronunciation, unpredictable topics coming up, etc. for at least 30-60 minutes) then they’re conversationally fluent.
But if they struggle here, they aren’t fluent.
B2 depends on the skill dimension:
Khalifa has been hyperfocusing on “fluency first” choosing the subtitles of YT videos and Netflix shows. Similar to Will, that’s an excellent strategy, esp. for inexperienced language learners.
But he’s probably not B2 in reading / writing comprehension. For example, a word count of only 23k words (on LingQ) is way too low to digest Stephen King’s “It” without problems. And even with >= 40k words (on LingQ) he would still have problems with “It” (that’s my case in Br. Portuguese).
But a “solid B2” should have no problems with reading / listening to “It”.
Anyway, I like the idea of “hyperfocusing on fluency first” more and more (since our in-depth discussion with xxdb). It’s definitely a strategy that all language learners should have in their toolboxes!
Right. Babies essentially spend 16 hours scratching their ass and listening to their mother tongue. Rest of us, we got bills to pay.
Why we need to find a more efficient way to force the neural pathways to burn in.
Numerically speaking, to get to the same understanding as a six year old would have looks like at least 10,000 hours.
Most adults can probably find an hour a day.
So we need at least a 20X efficiency gain.
Just focusing on vocabulary acquisition alone that children acquire around 5-10 words a day. So at the low end that’s 10,000 words.
To do this in a year you’d need to accumulate 27 words per day.
How much watching TV would you have to do to accumulate those 27 words?
I can’t find the research article quickly, but the gist of it is that you’d need at least 4 hours per day of watching gradually laddered up levels of difficulty to do it.
IMO, therefore, SRS is a must, at least until you have accumulated enough words to be able to learn from accurate guessing (the research says something like 5,000-7,000 words gives you that ability).
podcasts are super helpful.
I typically “ladder up” by starting with TPRS (the easiest kind of podcasters) on youtube.
Then when I understand pretty well (a couple months) I’ll ladder up to the next level (slow spoken schoolteacher style podcasters).
Then in a month more I’ll ladder up to news.
Then in another month I’ll ladder up to native podcasters.
Then in another month I’ll ladder up to easy TV shows (telenovela types).
Finally I’ll start on netflix shows.
So there’s kind of a straw man here though.
One of the very loud commentors is arguing that input only doesn’t work.
It clearly does work, for comprehension…
But if you move the goalposts and ask if it works for output my guess is the answer is mostly “probably not”.
I have a friend whose parents are slovenian. He can understand perfectly his parents but can’t speak it.
I have no means of saying whether he’s B2 or not. He speaks very well, makes some grammar mistakes as expected, but an extremely impressive level after 10 months.
Pronunciation is good, very clear - more towards Mexican Spanish which I find harder to evaluate than accents from Spain.
It’s a strawman argument, no question about it.
No rational person is saying that listening is a one-stop shop.
Listening DOES provide incredible potential and the building blocks/raw materials to speak but doesn’t guarantee the listener will become a phenomenal speaker. You still have to put in the time having conversations to develop that skill.
I’ve seen similar arguments in the gym. People who don’t do deadlifts or squats use the excuse that just doing squats or deadlifts won’t lead to the gains they want, so they don’t do either of them.
Who said to just do squats or deadlifts? Do them plus other high-yield exercises!
Here’s another: Language apps and word/sentence SRS apps by themselves won’t make you fluent. So, they are a complete waste of time to use at any stage of the language learning process and for any length of time. Umm, okay.
Nobody said listening was a standalone cure-all.
But it’s an insanely important skill to develop; don’t ignore it!
If I can’t find one activity by itself that makes me completely fluent and completely literate I won’t do any of them and you can’t make me; nanny nanny boo boo!
