Page 1 of 3

A Million Sentences: Cenwalh's 2020 Conquista

Posted: Sun Jan 12, 2020 10:50 am
by Cenwalh
Welcome to my belated Spanish 2020 and beyond log.

Previous log here.

I had no intentions of starting a new log for each year when I started my previous log, but I have done so because when people open up my old log, they are greeted with someone mainly obsessed with Anki. This no longer reflects my learning style or beliefs about methodology as I am attempting to learn Spanish solely via input.

I'm also changing my medium/long term goal from one of hours to one of sentences consumed. The reason I have done this is that I was finding myself watching things on Youtube search as gaming streams that whilst they are important at a level lower than mine because of their comprehensibility, they have quite a low amount of actual speaking in them. Therefore an hour of content consumed in one medium does not have the same benefit as in another medium. Having started this input only Spanish learning in July 2019, a "back of a napkin" calculation says I have reached almost 400,000 sentences. I plan to calculate this more accurately at some point - I am just really busy at the moment. This also gives me freedom to speed up or slow down content as necessary and not feel guilty that the time spent consuming isn't the same as the number I write down.

Current level: I've done a few tests on the internet recently to get a rough estimate of my level. I know they're not DELE or SIELE, but they're what I have access to. They all put me as either B2 or C1. That's a great level, but it's worth noting that I can't actually speak Spanish, so that's a level for understanding and being able to recognise what grammar "feels right" to me in a grammar type test (just like in English I suppose).

Goals for 2020: I want to hit 1,000,000 sentences. For reference 1,000,000 sentences is about 5 super challenges* if the book part is an adult novel without pictures and the film part is high density audiobooks, or about 7 super challenges if the sentence density is half that - and a lot of content is lower density still.

Why one million? I've seen the number mentioned a few times such as on the Antimoon website, this inspiring video, and other places. Plus it's a nice round number. 100,000 seems too few, and 10 million seems unachievable, so a million is a nice middle ground. Who knows what it'll bring?

Curve ball: I've started up Catalan again as a non-serious side language. As I mentioned in my previous log, my other half started learning Spanish, and the start of learning a language is pretty fun, so I'm having a go with Catalan. This is also partly to motivate her as we're about the same level, although I have a significant leg up having learnt quite a bit of Spanish.

*Number of SCs = 1,000,000 sentences/(100 books x 50 pages each x 250 words per page/17.5 words per sentence + 100 films x 1.5 hours each x 1,000 sentences per hour in high density audio)

Re: A Million Sentences: Cenwalh's 2020 Conquista

Posted: Sun Jan 12, 2020 11:52 am
by Cenwalh

Re: A Million Sentences: Cenwalh's 2020 Conquista

Posted: Mon Jan 20, 2020 9:48 pm
by Cenwalh
I had already written an update in this log, but I hadn't checked on it and it appears that it's not here... Maybe I closed the tab without posting or maybe there was a technical error. I'll try to fill in the gaps.

I started my first adult audiobook (Reina Roja by Juan Gómez-Jurado) - and promptly gave up. It was too hard, and I'm not about that. I decided instead to listen to the first two Harry Potter books. JK Rowling is such a good writer, and the translator really does her credit in my opinion. The books were pretty easy for me and I got through them in under a week each.

What next then? Well back to the adult audiobook. This time I stuck with it and at about 7 hours in I'm really enjoying it. Sometimes descriptions go over my head. I'm not sure if it's because I find it hard to pay attention when I'm tired or if I just haven't heard the words, but sometimes they fly past me and I don't catch much. Still, I persevere. My comprehension feels about as good in this book as it did in the first few young adult books I listened to which hopefully means I'll quickly catch up to adult books as I did there. I think I'll stick with this author for another book to get used to his idiolect.

In other news, the series I watch, Cuéntame cómo pasó, is much easier now. I don't have to try to understand what everyone's saying, although there remain a few problem characters whose parts I sometimes miss. Such is life. I don't only watch Cuéntame, I love watching documentaries especially a series called El cazador de cerebros which I find fairly easy.

I'm continuing Catalan on Duolingo. Note that I don't think it's the best way to learn it - it's more a side hobby I can do on my phone when I have a moment or when others are scrolling down social media which never interests me. I don't wish to go too deep into Catalan as I'd be worried about coming out with Spatalan! I frequently accidentally slip in Spanish words when I should be typing Catalan, although fortunately not the other way around.

Other half learning Spanish
This continues also through Duolingo. She finds the heart system frustrating and frequently runs out of lives without having done as much as she wants for the day. We have discussed flash cards and she has watched some Dreaming Spanish, but unfortunately doesn't find it very entertaining.

It makes me think that after all this time learning Spanish, I really don't actually know how to do it!


I enjoyed seeing this new log with a similar idea to mine, and I wish them the best of luck! I wish there was some science being done on people who do input only versus "studying"but I guess that's super hard to control and get good volunteers for.

Re: A Million Sentences: Cenwalh's 2020 Conquista

Posted: Wed Feb 05, 2020 10:51 pm
by Cenwalh
I think a sentence is a great measurement of consumption for several reasons:
  • A sentence unlike an individual word has grammar in it, and conveys a meaningful message
  • Sentences that are comprehensible to you get longer and harder as you get better. This means people at a beginning level can consume roughly the same number as someone more advanced, just like levelling up with XP in a computer game
  • Being mindful of sentences keeps you on track with keeping input high density.

However, it's becoming clear to me that some of my calculations are waaayyy off. I'm listening to the third Harry Potter audiobook at the moment. It's roughly 12 hours long and has roughly 110,000 words in it, that makes it roughly 150 words per minute. If I have assumed that a sentence is 17. 5 words as I did above, that would make it ~1,900 hours (24 months at my current rate) of high density audiobook listening to reach my goal! :o Clearly somewhere along the line I have made poor assumptions to make this a target for 2020, as that is just not achievable. Still a cool title though...

Maths issues aside, I continue with a heavy Spanish input approach. I've got into a habit of posting charts, and no doubt I'll do that for January at some point, but it was the usual ~80 hour affair so nothing too interesting.

It is amazing how much I am enjoying literature in another language. I'm getting to a level where I can appreciate a particularly well written or well narrated sentence. Sometimes I pause a book or rewind just to appreciate the writing. I'm glad I waited (what is almost a year now) to read Harry Potter because I can really feel the author and translator's work coming through. I feel like it wouldn't have been as well appreciated when I first considered reading them.

It's always interesting to see competing views on language learning methodologies in this forum and elsewhere. It's a topic I enjoy reading about. Truly I don't think we'll ever know the "correct" path, but I'll stick to mine for a while longer since I can feel it working for the purposes I want it to. Somehow I just manage to absorb words and grammar in an enjoyable way, and that's great. It's also great if other people take a different path and get similar or better results.

In Catalan I've just about scraped past checkpoint 2 on the Duolingo tree, so it's the home stretch to the golden owl for me which comprises of roughly the same amount of content as I've done thus far. I've watched a few videos in Catalan, and propped up by knowledge of Spanish I have a considerable amount of comprehension which is exciting!

An input only method for Catalan isn't in my sights though because it's very much a "when I get a minute" language which isn't compatible with my approach to Spanish.

Better half
The other half continues learning Spanish on Duo. She's got some of her family in it, so we have quite the little group going. I continue to not know where we should go from here. With textbooks and supposedly dull input that's comprehensible out of the question, it seems like a flashcard approach is where we might end up. In my opinion, that is more boring than the input, but I did it too so who am I to talk?

Re: A Million Sentences: Cenwalh's 2020 Conquista

Posted: Sun Feb 16, 2020 1:24 pm
by Cenwalh
After reading some posts in slowmoon's Reading 3.000.000 words in German log about why slowmoon didn't find a big increase in words known after 1 million words, and writing some thoughts myself on the matter, I've been getting a good enough mix of language.

How can one know if one is consuming enough variety? I certainly try to consume variety. I listen to audiobooks, I watch videos on various topics on YouTube, I watch fictional and non-fictional TV shows, I listen to learners podcasts where they try to fill in oft unspoken topics in other media etc etc. But is it enough? Is there some big list of topics that I can consume content about to get all the common words? I don't know, but I press on...

I've done some early experimenting with using Google Docs' voice typing to transcribe podcast audio so I can get word counts. It seems to be pretty accurate, and even though it sometimes gets words wrong, it tends to be that a word said is a word recognised (even if it's the wrong word) which is all I really need. I'm supposing that a few hours transcribed from different sources will give me a good start to my calculations, and I might be able to get YouTube subtitles from their API.

I've been very much relaxing on consuming solely content from Spain. Readers of my previous log may remember that that was a firm rule for me; if it's not from Spain, it doesn't get consumed. My reasoning is that content from other places can be fun, and if I think a video looks good on YouTube, I shouldn't not watch it just because it's not European. I also think I have a pretty good base in the language now, and I'm unlikely to be swayed into a different dialect if I don't do that much of it - and who cares if I am?

On the Catalan Duolingo front, I'm one league down from diamond (having been promoted every week), and I'm desperately trying to get promoted so I can relax a bit. I'm supposing staying in diamond will be a lot easier than getting here. After that, no plan... But I'm hoping I'll have an OK A1 level by the time I next go to Catalan speaking areas so I can do all the basic interactions just for the novelty, then revert to Spanish for actual communication.

Re: A Million Sentences: Cenwalh's 2020 Conquista

Posted: Thu Feb 20, 2020 7:20 pm
by Cenwalh
I have done some calculations on sentences consumed to date. The best reference I have is books because I can properly see where sentences are. I estimate that from September to the end of January I have consumed:
  • 140,000 sentences from audiobooks
  • 100,000 sentences from podcasts
  • 70,000 sentences from YouTube
  • 30,000 sentences from TV
  • 340,000 sentences in total

I have calculated these numbers by totting up as many sentences as possible in the ebooks of the audiobooks I've listened to, and extrapolating out across them. From this I found an average sentence length of 10 words - I can only assume this is down to young adult fiction being limited in sentence size anyway, and dialogue often being short. I then used YouTube's API to download the subtitles to all of the videos I've recorded (14% failed to give me anything), then I used the average sentence size from the books to tell me how many sentences this makes. I estimated that podcasts were 10% less efficient than books - this is consistent with limited recordings in Google Docs. Finally I estimated that TV is less a bit less word heavy than YouTube.

All in all it's not bad, and better than I thought it would be, but it's still unlikely that I'll hit 1,000,000 by the end of the year. Although contrary to what I thought before, it does actually seem in reach. I suspect that reading books might be quicker than listening to audio, but dyslexia might have something to say about that...

Re: A Million Sentences: Cenwalh's 2020 Conquista

Posted: Wed Apr 08, 2020 7:29 pm
by Cenwalh
I seem to have skipped a few entries, but they're all similar so what does it matter?

I tallied up ~160 hours of (active) input in Feb and March which is pretty standard for me. Since COVID-19 lockdown started in the UK I've had a bit of a challenge getting the Spanish in as I'm so used to mainly doing it on my 2 hour daily commute which I'm no longer doing. Furthermore I've now got a puppy to care for which means when I am doing Spanish it has to be audio so I can look after her, but it's so worth it :D

Feb and March 2020.PNG
160 hours in Feb and March by input type
Feb and March 2020.PNG (15.25 KiB) Viewed 1203 times

I like to sometimes do quizzes on the internet to assess my level. I first did the Kwiziq test in January and did fairly well all things considered, but I redid it today, inspired by Bex's log, (had to make a new account, was a right faff), and seem to have improved quite a bit. I left one question blank because I wasn't sure but got every other question correct. Unfortunately it doesn't seem to test C1 skills, so I had to do a C1 specific test and got 80% on that which is pretty good.

January results:
Kwiziq JAnuary 2020.PNG
Kwiziq JAnuary 2020.PNG (28.69 KiB) Viewed 1203 times

April results:
Kwiziq April 2020.PNG
Kwiziq April 2020.PNG (24.9 KiB) Viewed 1203 times

If anyone knows why it doesn't give 100% for things I got 100% on, I'd be interested to know...

I don't bother to discuss methodologies with people anymore because it doesn't lead anywhere, but it's clear to me that my method works for me and I enjoy it, so I'd never consider anything else.

It recently occurred to me that it's been a long time since I chose to consume something in Spanish and didn't understand it. I suspect this is because what I'm interested in is limited and I need to branch out a bit, but given my summer holiday to Spain is probably not going ahead now, I might as well stick to what I enjoy until I'm forced to change out of necessity!

Very limited progress here. I'm pretty much 75% of the way down the tree, but I've been there for almost a month now, and am just practicing instead of learning due to time constraints. I'm not bothered though because as long as I'm doing something, I'm not forgetting, so I can come back to it seriously when I like. I continue to be in the diamond league, although I've had a few close shaves!

Other half learning Spanish
This continues unabated, but is still mainly limited to Duolingo. Recently there's been TV consumption in Spanish with English subtitles which I think is a reasonably good idea if she'd be watching TV anyway and it's enjoyable because I think it tunes the ear a bit. Next steps aren't immediately clear, and there's still a long way to go before she'll be able to consume continuous content like a learner's podcast or Dreaming Spanish intermediate level videos.

Re: A Million Sentences: Cenwalh's 2020 Conquista

Posted: Mon Apr 13, 2020 3:13 pm
by Cenwalh
The Super Challenge is coming up, and I'm not sure what I should do about it. I didn't join last time because I didn't think I could do the reading, but I think perhaps I should start to read now. How can I ever get properly literate without either reading lots or learning the rules of grammar from scratch? The challenge entails 150 hours of audio/visual content, and 5,000 pages of reading. The former I could do in two months at my current rate, but how long would the latter take? As I've mentioned before, I'm dyslexic, and a large part of that is a slow reading speed. I'd like to do the double challenge for Spanish if my reading speed is fast enough, but 10,000 pages of reading is fairly daunting to me in any language.

I suppose I would consider a reasonable reading speed one where I can fit in the reading part of the challenge in 30 minutes daily which would mean about 17 pages daily, so a page every ~2 minutes which I think is a reading speed of 100 - 150 words per minute. I think that might be achievable with simple enough texts with a significant amount of practice, but for the moment seems unlikely to be achievable, so if I do go for the double SC, I'd have to supplement daily reading with longer stints at the weekend/on holiday.

Another consideration is if I want to do anything with Catalan. Perhaps I should go for the half super challenge so I have something to aim for, or perhaps I should do a double challenge for the Romance family to get them both in. Either way I need to decide sharpish because it starts soon!

Re: A Million Sentences: Cenwalh's 2020 Conquista

Posted: Wed Apr 15, 2020 9:55 am
by Cenwalh
What is it about wretched numbers that makes them so bloody difficult? One would have thought at this point (where I can casually start watching a TV show I've never seen before and understand it well) that numbers in Spanish would be a breeze. The reality could not be further from the truth.

1 - 20 are fine (perhaps even 1-100), and catching the first part of a number is fine too, so if the number is 1734.89 I'll know that it's around 1700, but I just never catch the detail because they take so long to process. These numbers fly by, and for months now I've just been ignoring them because I need to hear the rest of the sentence. One type I can get is years starting with 19 or 20, presumably because I hear them so much.

When will it get easier? Does it even matter?

Re: A Million Sentences: Cenwalh's 2020 Conquista

Posted: Wed Apr 15, 2020 12:30 pm
by cjareck
Perhaps you are making the mistake of translating the number in your head instead of just visualizing it? It could explain the reason why you need extra time to process them and can't catch up with the rest of the sentence.