I’ve been a Skritter user for about 3 weeks now and I’m loving it.
But… does anyone have strategies they’d like to share for remembering tones?
With characters I find I can “chunk” together the radicals, usually with some kind of mnemonic, and muscle-memory helps as well. Eventually I will “get” the character… after failing and trying again enough times, it finally sinks in.
For tones, I’m finding it much, much harder. Looking at my stats, I seem to have hit some kind of plateau when I reached 250 characters/60 words.
There are a few that I remember using mnemonics (like 学 being rising tone because kids at school are growing up; 川 is high tone because water stays level), or sometimes from the shape of the strokes (for example, the last stroke of 是 and 大 suggests a falling tone).
But apart from that I’m really bombing at tones all of a sudden. My retention has been dropping for the past few days. My words-tones score is suddenly abysmal—it’s even harder to get 2 or more right. The audio is helpful, sure, but I find recalling a sound way more difficult than a visual or logical mnemonic.
I’d love to hear your tactics and strategies! Thanks.
What level are you at? If you are fairly new to listening / speaking I would think it is going to be a challenge to hear the correct tone in your head. If that is the case I think time and practice with aural Chinese is only solution. One feature that I use is tone colors. Pleco supports this and so does Skritter 1.0 on the web. A standard combination is red = first tone, orange= second tone, green = third tone, blue = fourth tone, black or grey is a neutral tone. Tone colors are in development (I think) for other versions of Skritter, but not yet available. Finally, full disclosure: I’ve been at this for years and I would still describe my tones as “abysmal.” I’m not currently testing them, so I can’t give you an objective score number, but lets just say it would’t be pretty. I can say that they have at least improved some.
Level: I’ve been learning for 6 months on and off, but only just stepped up a gear in the last 2 months, and now having 2 skype lessons a week via italki. Scored 16/20 and 20/20 in HSK1 mock test. I find listening and pronunciation way harder than reading and writing.
Ah, the colours thing sounds interesting—I was using MDBG today and was wondering what the colours were for. I guess the theory is that you remember the colour along with the character? I have Pleco on my phone too.
I’m not sure how objective/accurate it is (maybe @skritterolle can comment as I found it via his blog post), but I tested myself on the wordswing tone training test after I’d been learning for about a month and scored 97%. That’s pure listening, divorced from any context or worrying about radicals, stroke order etc. Maybe I should go back and actually do the course now, as I feel like I’m going backwards!
I have to say my teacher is really great and is always encouraging me to get the tones right… it just requires some serious re-wiring of the brain, right? (And I’ve been speaking English for 50 years.) I still have a tendency to go up in pitch when asking questions, or even just reading something out loud that I’m not sure of.
This is really a separate question, but maybe I’m just trying to go too quickly on Skritter? Is there a rule of thumb for how many new words/characters it’s sensible to add per day, so you don’t get swamped with reviews?
The tone colors are a substitute for tone marks or sticking numbers at the end of pinyin syllables. Instead of a separate mark the character itself is colored. Of course numbers or colors are both arbitrary, but my theory is that colors would map through a different part of the brain than numbers.
As for how fast to go, I can’t answer that but you should be aware that the speed of adding new words is adjustable in the settings. You might want to try tweaking that. As your vocabulary grows you will find reviews due may tend to pile up. You can always ban words that have little incremental value in memorizing relative to the value of other words that are also in progress or upcoming.
Colours makes sense… I used to make revision notes using different coloured pens in high school and university, probably from something I read by Tony Buzan or similar. It does seem to make things more memorable.
I’m using Android Beta 2.3.4 and web version 2.0, and the auto-add feature is currently disabled, so I’m having to add manually as I go.
I recently tried to speed up to match the chapter of my textbook, and I’m wondering if I started to go a bit too fast. I’m wary of switching between 1.0 and 2.0, given the threads I’ve been reading about unpredictability of the SRS algorithm.
Ah, a fellow former Tony Buzan disciple. And I still do mind maps. I noticed a lot of forum comments about mismatches across platforms on Skritter. Eventually everything is supposed to be on a unified underlying platform, but who knows when that will be, or when they will restore the colors option. Can’t blame you for being cautious.
I think most college courses expect the student to learn fairly long vocabulary lists that probably aren’t well retained by the students. Skritter helps, but ultimately the learning may be a bit superficial if one is just drilling and cramming for tests instead of really learning word usage.
Another way to speed up Skritter is to lower the target retention rate. It seems counter-intuitive (isn’t the purpose of Skritter to ensure that I never forget anything I ever learnt?) but it’s a case of not letting the perfect be the enemy of the good. If you go for an overly ambitious retention rate you will never get to the end of your lists.
Speaking of never forgetting anything, I haven’t done a regular study session on Skritter for over a year now, so my queue is well into the thousands. I decided to sacrifice the goal of retaining all that vocabulary for the expediency of just focusing on some specialized vocabulary for whatever I am working on currently. I probably should “nuke” all my progress stats and go back to the 500 ~ 1,000 most common words to make sure I at least have them in working memory. I think at your level it’s best to see how much you can build your vocabulary and “know” at least a few thousand characters, but eventually the time required to maintain and grow your Skritter active vocabulary will probably become a constraint.
I’m not doing a college course, just 1:1 classes with a teacher on italki, who’s great. We’ve been using the Integrated Chinese Level 1 Part 1 textbook on and off, and I’m trying to read the dialogues from characters in class rather than the pinyin, hence why I’m trying to get the vocabulary down for each chapter. I’m just finding it hard to keep up.
I’ve also been doing some vocab learning on memrise, and writing answers using the pinyin input method is way easier and quicker. I’m a believer in using multiple methods at the same time to try and cover some of the gaps. I like the way memrise breaks sessions up into manageable chunks, but the quality of some of the lists is not great. I guess I could make my own.
A review queue of “thousands” sounds terrifying! Hopefully the SRS spacing will eventually settle down… I think I’ll just stay in review mode for a while, and maybe use memrise for learning vocab more quickly.
I also use Memrise, also partly on the theory that “multiple methods at the same time” may have value. However, if you believe that one or the other method has a nearly optimal SRS schedule (assuming that you use it faithfully) then using two methods concurrently could in theory actually degrade the effectiveness. Here is what I mean: The idea behind SRS is not “frequent reviews aid learning.” That would simply be “R”, not SR. SRS works by getting you to go through the effort of remembering something that you were on the verge of forgetting. It is the effort of digging something out of nearly-gone short term memory that places it into long term memory. If you ace your reviews on Skritter because you just reviewed the same list on Memrise (or vice versa) then the second SRS software that you used is going to have a false record of how well you have really memorized the words, and will therefore push the next reviews further into the future than they probably should be. That, at least is my understanding of how these algorithms work.
Of course we are not studying in order to maximize our review stats (at least I assume not) but rather we want to improve our “real world” functioning in the language. If we are reading / listening in the target language regularly then this vocabulary is going to get reinforced organically, and that’s a good thing. One habit I have changed in my use of Skritter is to really focus on the sample sentences: try to read them, change them if I don’t like them, or make my own. If you are comfortable doing a bit of hacking you might look into using the Memrise Chinese Examples script developed by cooljingle (check the Memrise forum)
Even though Memrise reviews seem to go by “quicker” (because many of the answers are just point-click or hit a number on the keyboard) I’m not convinced that it will move me toward literacy any sooner. I started using Skritter in the first place instead of flash card programs thinking that being forced to write the character would keep me more honest about whether I really “know” a word. I think Skritter has also helped me with building a base of etymology with which to relate new words to existing vocabulary. I don’t have anything to offer by way of empirical evidence. English is my native language, and I have never encountered anyone who could read English but not write it (although some people are terrible at spelling). On the other hand, there are Chinese people who have lived overseas for years who have problems remembering how to write many words but who probably can read just fine.
Coming back to the original question on tones, have you tried the tone pair drills from John Pasden / Sinosplice?
Sorry, didn’t see this post earlier for some reason! I can repsond to the question about the tone course, since I designed most of it (but Kevin did the coding). It’s accurate and helpful, but it only covers single syllables at the moment, which is a limiting factor since most words are actually two syllables and therefore undergo tone changes and is in general harder to both hear and pronounce. In other words, having no problems with the course tells you that you have basic tones down, but it doesn’t tell you you have no problems with normal listening, partly because that is about a lot more than just tones, and partly because polysyllabic words are common and not quite the same as monosyllabic words. Hope this helps!
Olle,
Are you (or anyone else) familiar with the speakgoodchinese computer program? I couldn’t find it at speakgoodchinese.org because the University of Amsterdam department’s server where it’s hosted is down for maintenance for a while, but I found a version of it at github: https://robvanson.github.io/sgc3/
I have the impression that with this program (and Rosetta Stone too) the user has to match the pace of a referenced sample. Speak a word more slowly or faster than the program’s reference sample and you will be marked wrong, even if the overall shape of the contour of your pitches is highly similar to what the program would consider correct. Ideally the program should just be looking at the overall shape of the pitches for (e.g.) two spoken syllables, just like a bar code scanner is looking at ratios, not absolute spatial distances between bars (at least that’s my understanding of how bar codes work). You could test this by getting a native speaker to pronounce the words at the same pace as the program’s sample, then at a faster pace and then a slower pace and see if the feedback differs. Also, I believe the programs are not “hearing” pronunciation, just pitches. If you just whistled or sang “la-la” where a spoken word is required you might get marked “correct” if the sounds you make match the tempo and pitch contour of the “correct” reference item. I should think with the current state of the art in speech recognition these types of programs could be refined much further.
I haven’t checked out that particular program, but I have tried similar products and they are generally very limited in the way they describe, i.e. they don’t really check your pronunciation, they check how close it is to the model audio. In other words, you can pronounce things perfectly and still fail. We tried such a program when I studied in Taiwan and many of my native speaking classmates who had very standardised teacher-style pronunciation still failed, not because they’re doing it wrong, but because their voices, pacing or whatever wasn’t similar to the original. As I said, though, I haven’t tried this particular program, but my guess is that it suffers from the same problem as other programs of this kind that I have checked out in the past.
Some recent research did not seem to endorse tone colors as a replacement for other systems, at least not for beginners:
“Results from pretests as well as immediate and delayed posttests indicated that multimodal training aided L2 learners’ tone perception, with a small, practical advantage for pitch contours and numbers over color coding.”
Hmm, without being able to read the full paper (paywall), from the abstract it appears to say that using multiple modalities (including colour) is slightly better than just one (including colour).
Practically speaking I’m finding that I’m s-l-o-w-l-y starting to remember tones mainly from hearing the audio — and my teacher — over and over again, although I’m tending to mix up 2nd ´_ and 3rd ˇ_ quite a bit.
Many thanks for the pointer to the Sinosplice content, @podster. I’m playing with it now… what’s the best way of using it do you think?
To be honest I have never done any methodical drilling on tones, so you are on your own here. Please let us know how it goes. I would suggest another strategy too, (which I haven’t really tried myself either): Get hold of a native speaker, either in person or on line. Give them a text that you will then read aloud. It could be as long or short a reading selection as you are comfortable with. Make sure that you already know the characters, if not what the tones are “supposed” to be. Have the native speaker make note of where your problems are but don’t interrupt during your recitation. Then review the problem words and combinations to see if you can reproduce the native speaker’s reading of the problem areas.
I don’t think it should be necessary that the native speaker be a trained teacher. Actually I haven’t seen many teachers really focusing on helping students sort out the tones in normal spoken speech (as opposed to word by word.)
Well, I do have a teacher via italki.com and we have one or two 1-hour lessons on Skype each week.
I’ve discussed with her (native speaker) and she says I’m doing OK but that generally it takes time to get fully comfortable. I’m definitely not ignoring tones, but her suggestion was to not get majorly hung-up on them right now, because it will come with time.
Thanks for your suggestion – I will discuss with her next time tones come up.
Of course the teacher is right, progress will come with time and experience. My own experience is that I seem to have gotten habituated into using the wrong tones for some fairly common words. I think what can often happen when working with a teacher is that conversations are highly structured and have very predictable vocabulary, so your interlocutor can easily guess what you meant to say, despite errant tones, etc. (I think Olle wrote an article about this once, probably on Hacking Chinese). I would also agree that its not worth stressing out over in the early stages, as long as you are making a conscious effort to check on your tones from time to time. The fact that I did not do this is probably why its hard for me to unlearn my wrong pronunciation of quite a few words.
It has multiple (and somewhat familiar) voices with audio for everything. I’m finding it’s helping me quite a bit with remembering tones, particular with words that might otherwise get confused such as 卖 and 买. Just hearing the pronunciations over and over again while reviewing quickly is very useful.
I learn in a fairly similar way, and while I can remember the tones they teach us, often visually or with a mnemonic, I can’t recall it and produce it fast enough to speak it in a sentence.
Still, I just practice by vocalizing the words in pairs and trios, and when I score myself on pronunciation I include the tone. I have an idea of what sounds I want to verify and improve on when I can converse with a fluent speaker…
In summary, I do skritter to build my foundation and vocabulary and confidence, I don’t think this will enable me to speak at all, honestly, not in complete sentences. But I’m gonna use it as a launching point for more listening comprehension, and just the all important pronunciation and conversation practice whenever I get the opportunity.
Yes well, for me there certainly isn’t time/brain capacity to recall tones while speaking. I find that by not worrying about this it “just comes” after a while. And I can definitely hear some of those words from memrise in my head if I do need to actively recall.
(Interesting to look back at this one year later, thanks for the bump!)
I’ve asked my teacher to pick me up and correct me on incorrect tones during our lessons, which she does. When we work together through texts in the book she asks me to read all the way through, then we go through again one sentence at a time with her speaking and me copying. This has helped too.
I passed HSK 3 in July and am now working through the HSK 4 textbook. FWIW I abandoned using Skritter to learn vocabulary because I wasn’t able to keep up with the required rate of learning new words. I’ve been using memrise instead, and occasionally StickyStudy for revision.
I’m still looking to re-introduce Skritter into my learning regime because I would like to be able to write the Hanzi, but I’m not sure yet how best to squeeze it back in. I think it has worked best for me when revising characters I already know, then I can focus purely on writing and stroke order.
(And I’ve been speaking English for 50 years.) I still have a tendency to go up in pitch when asking questions, or even just reading something out loud that I’m not sure of.