I know you’re busy so you might not see this, but I have a clear understanding of how you handle recognition-acquisition but not use-acqusition. I’m very happy with my workflow for the former, which has intuitively converged on ideas close to yours and is giving me good results, but I’ve never taken the latter to the same level in any language I’ve studied. I’m experimenting with how to do that, but if you have a second and can give me a super quick description of what activities in your study session specifically target use-acquisition I’d be grateful.
1 Like
FORK 110/400 WPM
I’ve just read this here: https://debatrix.com/en/speech-calculator/
The fork in that page is 110/170 wpm in English language.
In this more accurate article https://www.write-out-loud.com/speech-rate.html:
The fork is 110/400 wpm
-
Slow speech is usually regarded as less than 110 wpm, or words per minute.
-
Conversational speech generally falls between 120 wpm at the slow end, to 160 - 200 wpm in the fast range.
-
People who read books for radio or podcasts are often asked to speak at 150-160 wpm.
-
Auctioneers or commentators who practice speed speech are usually in the 250 to 400 wpm range.
Where exactly am I ultra reading/listening? 1.5x is irrelevant. We consume all types of different contents, from audiobooks, podcasts and plenty of Youtube videos from different sources.
But maybe they are all opinions, who cares anymore, right?
He doesn’t talk about failure, neither did I. The point is a reliable method (which I’m genuinely interested) not an emotion.
Those people are not generic people but the ones that are interacting and reading here in LingQ. They are language learners and quite intelligent. I assure you they understand.
Btw, now I go to have lunch and I’ll listen to an American podcast where there are 2 people. One of the two is painfully slow, I mean, seriously slow talking. I cannot speed him up to have a “normal” speed for me so I will have to zen myself to digest it. But I like their topic so I’ll go through the pain.
Fortunately, today, I don’t have to call my car mechanic otherwise, I will go from that slow American talker to an insane fast talker. My mechanic talks very fast, my father understand half of what he says, I struggle to understand 100%. He doesn’t talk in dialect, he’s just super fast.
@davideroccato The only relevant categories here are ‘conversational speech’ (120 - 200 wpm) and ‘podcasts’ (150-160 wpm). For ‘conversational speech’, just because they say that people are speaking at 120 wpm and 200 wpm doesn’t mean it’s common. The majority of people are probably speaking something similar to that of ‘podcasts’. Sure, it’s possible to encounter people who speak very slow or very fast, it’s just not that common.
The reason why talking about audio speed as a multiple of normal speed (i.e. 1.25x, etc.) is reasonable is simply because it’s more practical. The vast majority of devices and software record it this way - LingQ, YouTube, your podcast players, VLC Player, Spotify, WhatsApp, Netflix, etc. To tell someone they should be listening at 175 wpm to 225 wpm (which obviously changes with each language!), they have no easy way of do this without manual calculations. It’s just easier to say listen between 1.25x to 1.5x, so whatever player you are using, you can easily do it.
Also, regarding my above question, which one of these would you back? These are based on my real reading speeds: ~240 wpm extensive reading while listening, ~150 wpm using Language Reactor, ~55 wpm using LingQ.
1 Like
@nfera I totally understand what you are saying, it just becomes very unreliable if “we” want to standardise a method based on speed. What’s normal speed, what’s ultra reading/listening speed and so on. We cannot assume that we know what other people are listening or consuming as material. You can definitely consider that even the fork 110/200 wpm is not so tiny and you thought it was 145/155 before. Although we consume also Youtube videos that could be very fast. For example if I want to listen to Spanish football commentators.
If I read/listen at 110wpm at 1.5x I might think I’m ultra reading/listening because I see 1.5x. When maybe you are consuming material at 200wpm 0.9x and you might have the feeling that you are not doing enough. But the reality is that we are listening at the same speed. It is just the perception that is different.
The only parameter that we can really use to understand if there is any value in increasing speed to attain some faster result at the same acquired knowledge is wpm. But at the same time, as you pointed out and I’m well aware about it, the most common softwares we are using don’t allow that option. That is why I have the “feeling” that it is all quite anecdotal and based on the “perception” of speed and not actual measurement.
Don’t get me wrong, I appreciate the analysis and statistics you all are doing but I would definitely like to eliminate those “perception”, “feeling” and “opinions” from those numbers and get a reliable starting point that eventually we could all use.
Of course, if this is done just for you it’s ok but if you are trying to develop a method for other learners it becomes different.
However, it becomes important even for you if you want to be sure that you are keeping the same speed for a period of time and check your results and improvement without falling in the trap of distorted perception.
I answer to the other question in a separate post.
1 Like
@nfera There are indeed many variables. The comparison for me would more be like reading while listening on LingQ at ~55 wpm, reading while listening on YouTube with Language Reactor at ~150 wpm (I just tested it with the Harry Potter audiobook on YouTube with auto-subs at 1.3x, but having to repeat sentences every now and again) or true extensive reading while listening at 1.65x/~240 wpm. In this case, using my reading while listening speeds, we would be talking about 10M words read as extensive reading while listening, ~6M with Language Reactor and ~2.5M with LingQ. As mentioned, the LingQ reading speed is so slow because of the amount of time it requires to get a decent definition (which often includes pausing the audio to open up a dictionary and write a definition). With the translation under the subtitle on Language Reactor (aka bilingual text), you can merely glance down often to get a good definition, without ever having to click anything, hence the increased reading speed. Which of these three options would you back?
@nfera the fact is that I embrace now the philosophy to use one method, one software, and get the most out of it. This is to avoid wasting time to learn too many different softwares, too many different methods and so on.
In this case, I would definitely use LingQ as my primary source because I learn more when I stop and look-up words, even if I read/listen at the same time, or if I listen to a video/podcast before and then I upload the text on LingQ and scan through the words.
I would use Language Reactor only if I would like to improve fine tuning listening and repeating/speaking. Which means that I would probably use it after having reached a high level with LingQ, as an extension or plugin to LingQ. Or if I would like to solidify a certain level I want to reach and I would test myself if I have it.
I don’t know about extensive reading/listening without looking up words because I’ve never done it intentionally. I should probably use another software like Kindle/Whispering to test it as I wouldn’t like to skip words on LingQ (blue/yellow words and so on).
I don’t understand this approach as I don’t have room for it.
When I read articles online or listen to audiobooks or radios and so on, without looking up words or paying attention to the language by slowing down things a bit, I don’t see much progress compared to LingQ. I pay attention and enjoy the content but it is very difficult to switch where my mind is focused (the language? the content? the sentence structure? the pronunciation? and so on).
I don’t know about your statistics. The fork seems to be very wide from 2.5M to 10M. However, I’m not an advocate of number of words read, I don’t use that parameter. As always, it’s not the quantity but the quality. In this case, the relationship between both is very difficult to measure.
Interesting topic by the way, definitely for geeks. 
2 Likes
I think my experiences echo those of many others in this thread.
To start, I would define intensive reading as taking the time to re-read pieces of content, actively look up definitions, stopping to try to remember definitions of words while reading, etc. while extensive reading would be akin to the experience of reading a physical book where you don’t look up words and just consume as much content as possible. Obviously this can also be done in LingQ, but a physical book is a good way to visualize what I would consider to be the process of extensive reading.
It seems to me that at the beginning stages of a language, intensive reading is better since intensive reading activities are going to give you more opportunities to see the same words over and over. On the other hand, once you’re at the intermediate/advanced stages, extensive reading seems to be more effective unless you want to actively try to learn some special vocabulary. For example, even though I’m advanced in Italian, if I want to learn vocabulary specific to an area of interest (say, computers), I might re-read an article about the latest graphics cards over and over so that I can acquire that specialized vocabulary easier than I would otherwise.
All together, this means that I think intensive reading is simply useful for when there is a need to actively learn vocabulary (at any stage of learning), while extensive reading is useful for learning some vocab at a potentially slower pace, but is much more useful for solidifying existing knowledge and building the skill of reading in general.
3 Likes
I am not @PeterBormann 
, but for reference I started well below 100 WPM and now read comfortably between 250-350 WPM depending on the language, content and if I am R+L’ing.
5 Likes
.
Nobody’s perfect :>
8 Likes
Yes, I can talk about it in a few days.
Unfortunately, I’m in a hurry now…
1 Like
Hi Davide,
Just a few quick answers:
“I thought you were creating a method about Ultra reading/listening”
I didn’t have time to read most of the new comments, but
my impression is that your understanding of a “method” is that
it has to be mathematically precise and foolproof.
That’s not possible in the context of SLA.
Instead, you always want to be highly flexible and have a toolbox with many
approaches / practices and digital / non-digital tools.
Focused exposure time in timeboxing approaches should
be enough for RWL. Just increase the number of Pomodoros -
If you want to do more.
That said, if someone has done 1-2 RWL-Pomodoros and now
wants to read in / listen to his / her L2 for several hours? Sure, why
not. And you don’t need Pomodoros for that either
To be clear here:
I’m not interested in increasing the number of hours more and more
and more, but in the “minimal effective dose” (MED) so that busy bees
(professionals with a day job, moms with kids, etc.) can use this approach.
I’d I have the wrong impression?
Book: yes.
SW: yes.
Unfortunately, I don’t have time for this now bc. my dad died
at the end of 2022. My sister and I have to wind down his company.
This plus my day job takes most of the time now.
I hope I can finish the book till dec. bc. I’ve still
got 5 weeks of vacation this year.
Have to go now.
Nice day
Peter
5 Likes
@davideroccato At the end of the day, it’s all anecdotal and opinion based, because it’s based off experience or we see ourselves as applying logic. That is, increasing the audio speed has two benefits: increases your exposure to the language (which we think is a good thing, because we assume our experience and the experience of others means exposure = learning the language), and increases your comfort with listening to fast talking (also personal experience and recounted experience of others).
I’m not interested in making a method for anyone else, but personally, I have had very positive experiences with increasing audio speed.
Ah okay. I agree that lots of different tools can be distracting, that’s why I try and keep them to a minimum. But I’m completley okay for using more than one tool. But each to their own.
2 Likes
2 Likes
To me, LingQ is an incredible time and effort-saver. At a low-to-intermediate level, it takes me MUCH longer to look up words than it takes to click the word on LingQ, even if I need to read through a few non-ideal definitions, or even if I need to click through to one of the dictionaries.
But you’re right, at some point as you become advanced, that flips around. It would be ridiculous of me, e.g., to use LingQ for English. I need to look up maybe a word or two per month, and usually that’s not because I didn’t understand it so much as to double-check. It would be an incredible waste of time trying to do all that through LingQ.
1 Like
@fabbol For sure, it is a lot faster than using a paper dictionary. It is also faster than having a second tab open and copying and pasting the word into there every time you want a definition. We all agree on that. The question is how does it compare with other methods of looking up definitions? I want my time-waste to be an absolute minimum.
For me with Italian, as a B2+, I still need to look up a lot of words. I’m just saying that, on LingQ, my experience of wanting to know the definition of a word still requires a lot of time-wasting faff. The comparison of my wpm on LingQ and on Language Reactor (55 wpm vs 150 wpm) is just an example of how much my use of LingQ is actually not studying the language, but rather navigating the software.
For me, this comparison was a wake up call (almost 3x slower!). I never really thought about it, so I thought going through a lesson was studying. But it turns out, no. Over half my time is dealing with the software, clicking here, clicking there, opening dictionaries, and writing definitions. It took a comparison with a different software to realise this.
This is just my use of the software though. Perhaps it’s different for others. Especially those who are beginner or lower intermediate in popular languages, because can actually be good Community Definitions. Though, it still requires two clicks per word though, which means you have to be fast, if you don’t want to pause the audio. But maybe others have a better LingQ practice than me, so waste less time.
1 Like
I definitely agree with that and LingQ could definitely help a lot. If they would care about this, they could definitely help with an better UI that would reduce a lot of clicks and waste of time. But I suppose we all know that and sometimes we just get angry about it or ZEN!
If LingQ would take this seriously, I bet your wpm would increase.
The fact is that you are the first on saying that you are not really sure on what you have gained after your extensive reading experience (vocabulary wise). I wouldn’t be sure either and I would have a sort of psychological block about it. It is like if processing with LingQ gives me more confidence that I’m actually doing something more than just R+L without looking up words and focusing on learning. Different thing is that we just focus on content.
However, you could definitely increase your wpm if you reprocess the same material a second time with LingQ. Because in this scenario, you don’t have to stop looking up blue words, and you can focus only on converting yellow words.
In these cases, I only stop here and there to check some yellow words, the ones that jump on my mind with more curiosity. I bet in these cases the wpm would be similar to Language Reactor. With the advantage to have already my own dictionary on those yellow words.
I don’t remember if you wrote about it already but I’m curious to know, how do you use Language Reactor in those circumstances? What is the exact process and material you are using? Could you make an example?
With this extensive reading/listening, I wouldn’t use it on LingQ with blue words, but it might be worth trying with another software.
1 Like
Well, you shouldn’t be me.
I was kicked out of OpenAI’s ChatGPT permanently bc. I provided
some ideas about the optimizations of xenomorphs and terminators
(btw., esp. the hybrid “Xenomorph - Mickey Mouse” was very nice) for
the AI superintelligence to come (let’s call it: SkyNetGPT).
So, I’m probably on the hit list of SkyNetGPT as well:
“Knock, knock. Hey, Peter! Thanks for your optimization
ideas. There are some creatures that like to pay you a
visit” 
In short, not being me is probably better for your future health.
BTW, how do you like my little monster? (“Man, thanks to Craiyon, I feel
like I’m seven again. Ergo, it’s a fountain of youth for the mind” ).
1 Like
I’d say that’s way too inflexible in all kinds of skill acquisition processes (sports, SLA, math, programming, you name it).
This “simplification” is just another expression for the “law of the instrument” (Maslow’s hammer: " “if all you have is a hammer, everything looks like a nail”).
In general, skill acquisition practitioners are much more successful when they adapt to their environment and the specific problems they encounter. In short, you want to have a box with several approaches and tools - and then you try to use (or even invent) the “right approaches / tools for the job”. Examples:
- (Art) craftsmen have been practicing this flexible adaptation for centuries.
- The Chinese culture has had a similar view on “strategy” for thousands of years. See: F. Jullien, " A Treatise on Efficacy: Between Western and Chinese Thinking",
- And you see the same flexible “logic” applied in Coach Wade’s books about “progressive calisthenics”.
etc. pp.
BTW, people who practice this “flexible adjustment” usually don’t stay on a plateau, because before they do that for a longer period of time, they change what they do …
1 Like
The 1st lever you have here is: the selection of the material (known - unknown topics,
fiction vs non fiction, poetry vs prose, long vs short, contemporary vs non contemporary,
etc.).
1 Like
Monster: you are right @PeterBormann , I like way too flexible meat but I taste many, no problem with that! Btw, I received a call from a baby SkyNetGPT, it was talking about knocking someone’s door. Knock, knock!
1 Like
@PeterBormann So the vast majority of my study at the moment (Harry Potter series test aside) is YouTube. One of the channels, which I watch, is Nova Lectio, where he essentially makes mini documentaries mainly on modern or historical geopolitics. He uses a large range of vocabulary, so there are many words I would like to learn. According to my LingQ stats, a 25 minute video generally may have between 120-160 New Words in it (8-11%). With Language Reactor, I just open the video, increase the speed to 1.25x, and then press play.
The reason I can listen without much stopping on Language Reactor is because it doesn’t require a minimum of two clicks to get the meaning of the New Word (or more, if I need a dictionary), which often means I have to pause the audio. On Language Reactor, I can quickly glance down at the English translation to understand the word. This auto-translation works better, because Google Translation / DeepL (whatever they use) has the context from the sentence to correctly guess the word, unlike LingQ, which translates the individual word. If I’m still unsure, I hover the mouse over the word and it shows me several definitions, no click required. At my level, because there is often only one unknown word per sentence, the English sentence is often enough to understand the definition of the word - that is, scan for the single word I don’t know. If there are several unknown words in a sentence (at 1.25x), I generally can’t get all definitions. In those cases, hovering on a definition may suffice, as it auto-pauses the playing at the end of the subtitle timestamp (unpausing also requires no click, but merely moving the mouse off the subtitle), or pressing the hotkey to repeat the subtitle. Playing at 1.25x, even with my repeating sentences and momentarily pausing, I spend much more time with the language than on LingQ.
The main issue hindering my LingQ effective reading speed is the dictionary issue. If I could just buy a bilingual dictionary and easily import it into LingQ, it would massively reduce my time-waste on LingQ. Secondarily, Language Reactor has applied things to really reduce the number of clicks (English subtitles, hovering over the word to get the definition and auto-pausing while looking up a word, etc.).
Good point.
4 Likes