Synergy Part 6 - Pronunciation

Discussion in 'The language learning methods of Big_Dog' started by Big_Dog, Mar 8, 2014.

  1. Big_Dog

    Big_Dog Administrator Staff Member

    Joined:
    Jan 11, 2014
    Messages:
    1,039
    Native Language:
    English
    Advanced Languages:
    Spanish
    Intermediate Languages:
    French, Japanese, Mandarin, Russian, Swahili, Thai
    Basic Languages:
    Korean
    I have already expanded on the four basic language learning skills. There are also three of what I call “sub-skills”; pronunciation, vocabulary and grammar. This post is about how pronunciation fits into Synergy.

    There is quite a lot of discussion on the internet about the importance of pronunciation. It’s not necessary to have native-like pronunciation, but your pronunciation should be good enough so that a native speaker will understand you easily. I recommend really trying to sound like a native, within the time allotted for pronunciation in a well balanced language plan, and leaving it at that if you are easily understood. But try to be very honest with yourself regarding how well your pronunciation is understood. Recording yourself and sharing with others is a brutal but very good way to check your understandability. Fossilized pronunciation errors can be very painful to fix, so work hard on it in the beginning.

    My experiences with Thai pronunciation. I’m going to talk about Thai now because I made lots of mistakes with it, and learned a lot about pronunciation in general in the process. I started using a textbook. The textbook explained pronunciation thoroughly, which is impressive because Thai is a tonal language with varying vowel lengths. They defined and used their own transliteration system too. The text came with audio, which was also pretty good. The text introduced the writing system gradually. I did the bare minimum regarding reading and writing. I didn’t even do the reading exercises in the back of the book. I had never learned a script before, and was very intimidated. I thought the transliteration was easy though. But I pretty much ignored tones and vowel length. When I listened to the audio, I didn’t really like the way the natives pronounced, so I just repeated the sentences in my own way. I went to Thailand for a long vacation, and was not understood at all. I was shocked – I’d been learning Thai for 9 months, and it was a total waste of time. This was all due to bad pronunciation.

    I came home and decided to hire a Thai tutor. She helped me pronounce one vowel I was having trouble with. She stressed the importance of tones, and I was able to make three distinct tones without much effort. Unfortunately, there are five tones, and I wasn’t distinguishing between long and short vowels. This was not the teacher’s fault. I still didn’t believe these things were so important, so I didn’t want to make the effort to fix the problem. The next trip to Thailand I was understood to a degree, but still frustrated. I could produce vocabulary and grammar correctly and fluidly most of the time, but I was only understood about half the time.

    I returned home again. Over the next few years I got to where I was producing four tones, and doing a little better with my vowel length. I was understood most of the time now, but still had some very frustrating moments where I would say a sentence or even just a word that I knew was right, but got blank stares. This was not because my face is non-Asian. After analyzing things when I cooled off, it was usually a vowel length problem, but sometimes that missing fifth tone.

    It wasn’t until the last 2 or 3 years that I decided to make a serious effort to get my tones and vowel length straight. I stopped reading for speed. Now I read as fast as I can with correct pronunciation. I write now, because Thai is sort of phonetic, which means if I can remember how to spell it I can remember how to pronounce it. Why am I struggling so much? Because I’ve been learning Thai for over 10 years now. I’m roughly at the B2 level, and know thousands of words. So I’ve probably been mispronouncing well over 1000 words for many years. That’s called fossilization, and it’s time consuming and tedious to fix. But it’s not impossible. Last time I went to Thailand, I was well understood. I still have problems, but my pronunciation is much better.

    So please work on pronunciation from the start. Avoid fossilization. I get this feeling of regret when I realize how good my Thai pronunciation could have been now if I had just cared in the beginning. When you do your pronunciation, follow the steps in Synergy, and make an effort to copy every single aspect of the sound. Of course you won’t have perfect pronunciation after these exercises, but caring about it from the beginning will make a huge difference down the road.


    Chinese & Japanese pronunciation – an exception to Step 1. During your pre-learning research for Chinese or Japanese, you are bound to find out that you will have to learn thousands of characters. This is the type of “unique aspect” I was talking about that requires additional thought, or a different approach. Step 1 of Synergy requires you to learn the orthography while you learn the pronunciation of words. That would be very time consuming for these languages. I don’t recommend putting your other studies on hold until you learn all the characters and their pronunciations. Instead, use transliteration (pinyin for Mandarin, kana for Japanese, etc). This is one of the few cases where transliteration is ok. Remember that, if the language has them, learning tones and their sandhi are absolutely required at this stage. Other than that, just follow Step 1. How you eventually integrate actual characters into your study plan will be left for another post.


    Thai pronunciation – another exception to step 1 or 2? Thai doesn’t have an alphabet; it has an abugida. But that’s just a technicality. It’s fairly phonetic, so based on that it would seem logical that it wouldn’t take much time to learn. Unfortunately, that’s not the case. It’s unlike any other alphabet I’ve worked with. It consistently tops the list of the world’s hardest alphabets in the forum discussions I see. It’s unlikely that you will be able to read correctly pronounced single words with confidence in less than 50 hours of study. And reading sentences often requires that you know the individual words, due to the lack of spaces between words. I think a person who has successfully learned a language with a different script before could spend under 100 hours up front with Thai script and avoid using transliteration. But for all others, I recommend using transliteration at first, and weaning themselves off of it as they learn the script, vocabulary and grammar in parallel.


    Does delaying speech lead to better pronunciation? As I’ve mentioned in other posts, you absolutely must listen to the correct native pronunciation before repeating it in Steps 1 and 2. Reading before listening is probably the main cause of incorrect pronunciation. So delaying speech until after you hear the pronunciation is critical. But I occasionally hear that you should just listen to the language, for several dozen to several hundred hours depending on the pundit, before doing anything else. And one of the benefits they claim the learner will get is native like pronunciation. I remember when I started learning Mandarin several years ago, several forum members tried to convince me that I needed to listen for at least 700 hours before doing anything else, otherwise my pronunciation would be terrible. Well, I didn’t do it, and my pronunciation is fine. Most of the people I read about who try this live to regret it. The few I have read about who do it and still believe it’s a good idea don’t claim to have significantly better pronunciation. Therefore, due to all its other benefits, I believe starting speech early, as spelled out in Synergy, is a better method.


    Don’t model after songs or non-natives. When you are in the early stages of learning a language, you want to use standard native audio to model your pronunciation on. Although I have read lots of people recommend you start learning languages by listening to songs, I think it’s a bad idea. If singing is the main source of your pronunciation, then you can very easily develop the same kind of non-standard speech that you hear in the songs. This is especially true of tonal languages - singers change tones to match the surrounding lyrics and it's understood in the song, but people won't understand in conversation. Of course it’s ok to listen to target language music; just don’t use it to model your pronunciation on. Also, don’t mimic non-native or non-standard speakers in your formative stages if your goal is standard speech.


    Learn the linguistics of pronunciation. I admit to knowing almost nothing about linguistics. In the past few years, having studied so many languages, I’ve begun to realize how helpful it would have been to know the linguistics of pronunciation. There are times when I’ve practiced a sound until I’m blue in the face, and still get it wrong. The idea that knowing the linguistic terms for a given sound, along with the audio, will allow the linguist to produce the sound correctly is very appealing. For example, I know what the term “aspirated” means. After producing and aspirated consonant, I can feel a puff of air if I put my hand in front of my mouth. I can’t after an unaspirated consonant. I learned that Thai consonants at the end of a word aren’t aspirated, and knowing that helped my pronunciation. Many Thais have trouble pronouncing English words that end in an aspirated consonant. When I explain aspirated to them, their pronunciation improves. I believe knowing some linguistics would be very useful to the aspiring (no pun intended) polyglot. Although I wish I learned some earlier, I hope to study it in the future.
  2. BAnna

    BAnna Active Member VIP member

    Joined:
    Apr 17, 2014
    Messages:
    104
    Native Language:
    English
    Intermediate Languages:
    German, Spanish
    Basic Languages:
    Russian, Turkish
    luke and biTsar like this.
  3. biTsar

    biTsar Active Member VIP member

    Joined:
    Apr 22, 2014
    Messages:
    237
    Big_Dog likes this.
  4. Cainntear

    Cainntear Active Member VIP member

    Joined:
    Apr 29, 2014
    Messages:
    343
    Native Language:
    English
    Advanced Languages:
    Catalan, French, Italian, Scottish_Gaelic, Spanish
    Intermediate Languages:
    Corsican
    Basic Languages:
    Dutch, German, Irish, Polish, Russian, Welsh, Sicilian
    I seem to recall recent studies finding that aural minimal pair practise made no difference in performance - ie the test subjects got better at doing minimal pair tests, but showed no improvement in pronunciation or in distinguishing the phonemes in fluent speech.
  5. Peregrinus

    Peregrinus Active Member

    Joined:
    May 27, 2014
    Messages:
    613
    Native Language:
    English
    Intermediate Languages:
    German
    Basic Languages:
    Spanish

    So gleefully I will review the concept, but not the book since I am unwilling to pay for something that likely consists of nothing much new, and a relabeling of known concepts to make it seem like the author invented something heretofore unknown. From the blurb on Amazon it looks like his system rests on minimal pair training and SRS.

    So the author guy ( Gabriel Wyner) went to kickstarter for funding for pronunciation trainers. He has several youtube vids for French pronunciation, and also one regarding the research behind his big concept: The Research behind the Fluent Forever Pronunciation Trainers. Helpfully, one of the commentators there linked to the study cited (McClelland et al., 2002), regarding differentiation of "L" and "R" sounds by Japanese listeners. The big result was that training plus immediate feedback, as opposed to no feedback, raised differentiation from 50% to 80%, which the study authors viewed as almost a failure, but a result that Wyner says, likely rightly, that most language learners would be thrilled with.

    When I learned Mandarin decades ago, minimal pair training was, and likely still is, standard. I found it helpful. The question though as Cainntear alludes to, is whether lab training results equate to results in the wild. I am not willing to read that study cited, but it makes a big difference if they only worked on individual word pairs in isolation, or with longer utterances as well. Also I wonder if they taught tongue/mouth position along with the other training. Learning the proper position of the tongue/mouth (plus other factors like vowel length) is key to me for learning to distinguish many difficult to produce and understand sounds. If a learner does not know how to produce a sound, how can he/she be expected to understand it?
    Big_Dog likes this.
  6. luke

    luke Member VIP member

    Joined:
    Apr 27, 2014
    Messages:
    66
    Native Language:
    English
    Intermediate Languages:
    Esperanto, French, Spanish
    FSI Basic Spanish starts out with minimal pair training in the first two units. I found it helpful. Note that this was towards the beginning of my Spanish studies. It may be more effective if minimal pairs are understood early on. Also, there is the variability amongst students. What I would wonder is not if some have studied minimal pairs and not made good progress in pronunciation, but rather of those who have decent pronunciation and began learning as adults, how many of them had minimal pair training.
    Big_Dog likes this.
  7. Cainntear

    Cainntear Active Member VIP member

    Joined:
    Apr 29, 2014
    Messages:
    343
    Native Language:
    English
    Advanced Languages:
    Catalan, French, Italian, Scottish_Gaelic, Spanish
    Intermediate Languages:
    Corsican
    Basic Languages:
    Dutch, German, Irish, Polish, Russian, Welsh, Sicilian
    ...only if they didn't realise that it was 80% on two pairs of words from one speaker. The researchers did a very restricted test and are in a far better position to judge the effectiveness of their results than your average uninformed language learning hopeful.
    Minimal pair training almost always means words in isolation and that's what the book author proposes. It would be very hard to find natural-sounding sentences that used minimal pairs. How many sentences can you think of that would allow you to contrast "chip" and "cheap", for example?

    I totally agree with you on pronunciation training. I learned to physically pronounce sounds that I couldn't initially hear the difference between, and with time (and no conscious effort) I found myself hearing the difference. Production first, I say.
    Last edited: Jul 10, 2014
  8. Peregrinus

    Peregrinus Active Member

    Joined:
    May 27, 2014
    Messages:
    613
    Native Language:
    English
    Intermediate Languages:
    German
    Basic Languages:
    Spanish
    Sentences don't have to fully make sense, or be natural utterances, for purposes of demonstration, and many think humorous examples as with the original Assimil courses help learning.

    To wit:
    "The bird cheated me out of my crumb and then chirped about it."
    "My brother only buys cheap chips."
    "Don't be so cheeky to the man from Chile."

    It only takes a couple examples for each pair.
  9. Cainntear

    Cainntear Active Member VIP member

    Joined:
    Apr 29, 2014
    Messages:
    343
    Native Language:
    English
    Advanced Languages:
    Catalan, French, Italian, Scottish_Gaelic, Spanish
    Intermediate Languages:
    Corsican
    Basic Languages:
    Dutch, German, Irish, Polish, Russian, Welsh, Sicilian
    But that's no longer minimal pairs. The point of minimal pairs is that there should be no contextual cues that allow you to disambiguate, forcing you to rely on a single phonetic difference to distinguish the two. In "my brother only buys cheap chips", you have "cheap" shown as an adjective thanks to word order, and "chip" must be a noun because of word order and plural marking. You don't need to hear the vowel difference to distinguish the two sentences, so there's no impetus for the brain to learn.
  10. Big_Dog

    Big_Dog Administrator Staff Member

    Joined:
    Jan 11, 2014
    Messages:
    1,039
    Native Language:
    English
    Advanced Languages:
    Spanish
    Intermediate Languages:
    French, Japanese, Mandarin, Russian, Swahili, Thai
    Basic Languages:
    Korean
    Can you give an example of a valid minimal pair exercise? I'm totally ignorant.
  11. Cainntear

    Cainntear Active Member VIP member

    Joined:
    Apr 29, 2014
    Messages:
    343
    Native Language:
    English
    Advanced Languages:
    Catalan, French, Italian, Scottish_Gaelic, Spanish
    Intermediate Languages:
    Corsican
    Basic Languages:
    Dutch, German, Irish, Polish, Russian, Welsh, Sicilian
    Most minimal pair exercises are just isolated words.

    For example, you hear the word "sheep" and are given the choice between the written words "sheep" and "cheap". Here, it's testing the minimal difference between SH and CH (=TSH) phonemes.

    Or you hear "zip" and have to choose between "zip" and "sip", testing your discernment of voiced vs unvoiced.

    You can check out shiporsheep.com for more English examples.

    The thing with minimal pairs tests is that they're designed for measurability - it's labwork, not schoolwork. It's probably another case of teachers trying to be "cutting edge" and doing what the scientists do without stopping to think about why they do it.
    Big_Dog likes this.
  12. Cainntear

    Cainntear Active Member VIP member

    Joined:
    Apr 29, 2014
    Messages:
    343
    Native Language:
    English
    Advanced Languages:
    Catalan, French, Italian, Scottish_Gaelic, Spanish
    Intermediate Languages:
    Corsican
    Basic Languages:
    Dutch, German, Irish, Polish, Russian, Welsh, Sicilian
    Actually, it sounds as though Wyner didn't even bother to read the study properly. In order to get the students to learn to make the distinction, they started with exaggerated differences, en reduced the exaggeration gradually towards natural speech. The study showed that if you didn't use adaptive learning, the students never learned. Wyner's page doesn't talk about exaggeration, so it's not doing what the study considered necessary for learning.

    Another case of blindly citing a study that doesn't support your stance.
  13. biTsar

    biTsar Active Member VIP member

    Joined:
    Apr 22, 2014
    Messages:
    237
    I had pre-ordered it, but have canceled. There's a sucker bored every minute, or something like that.
    Big_Dog likes this.
  14. Elexi

    Elexi New Member

    Joined:
    May 31, 2014
    Messages:
    10
    Native Language:
    English
    Intermediate Languages:
    French
    Basic Languages:
    German
    As posted by Daristani on HTLAL, Routledge have put up the entire Colloquial series in downloadable audio. I have found the Colloquial pronunciation sections at the start of the course to be among the best exercises for what Polydog advocates - and now usable even if one doesn't care for the course in general - here's the link: http://www.routledgetextbooks.com/textbooks/colloquial/defau lt.php#languages
    Big_Dog likes this.

Share This Page