How does one learn to speak a foreign language with a native or near-native accent? Described below is what has worked for me in the past. Hopefully it helps you too.
Learn the IPA
I've written about this before, so I'll keep it simple. Make sure you can look at this chart and understand what's going on. Backness, roundedness, vowel height, and so on. Visualize that the IPA vowel chart as a cross-section of someone's mouth, as if you were looking through their left cheek. Each vowel is produced by a different tongue position; for example, to produce the sound /i/, your tongue should be at the very upper front of your mouth. For /ɒ/, it should be as far back and down as possible, with your lips rounded, and so on. Be sure to understand the principles behind how vowel sounds are formed - it's all about tongue position and lip roundedness.
The same goes for the consonant chart. I've gone over this before as well, but in a nutshell, consonants are formed by the tongue touching (or almost touching) various parts of the mouth. The left-most side of the chart is for sounds formed at the front of the mouth (lips, teeth), and the right side is for the back of the mouth (soft palate, throat).
Understand how these two charts work; it will be immensely helpful, I promise. You need to achieve a finer degree of control over your speech organs than most sane people ever do in their lifetimes.
Applying the IPA to your target language
Open up "language phonology" on Wikipedia, where language is your target language's name, and scroll down to the consonant and vowel charts. Note every symbol that is used in these, and practice them using the following resources:
- the IPA chart with audio. Keep in mind that the audio here is for the cardinal vowels, which might be slightly different than how those vowels are realized in your specific language (i.e. /ə/ in French is not the same as /ə/ in English). These are still a very good reference point, however. This just means you'll have to adjust things a bit for particular languages.
- Forvo, WordReference, and Wiktionary (if it has audio recordings for the language) for native speaker recordings of words - more on how to effectively use these in a moment.
Now, how does one go about actually learning all the various sounds in a language? It's all well and good to know the IPA, but actually applying it to a language is the really important step.
I like to do vowels first, because those are usually more difficult than the consonants. Errors in vowel pronunciation also tend to be more audible in my experience.
Compile minimal pairs
Here's what you need to do: compile a list of minimal pairs (link to Wikipedia), starting with the vowel phonemes. An example (taken from Wikipedia's "French phonology" page) would be:
- /i/: si
- /e/: fée
- /ɛ/: fait
- /ə/: ce
- /œ/: sœur
- /ø/: ceux
- /y/: su
- /u/: sous
- /o/: sot
- /ɔ/: sort
- /a/: sa
- /ɑ̃/: sans
- /ɔ̃/: son
- /ɛ̃/: brin
Note how these are all very short, simple words, but each has a different vowel. The surrounding consonants are also very similar. This allows you to ignore the consonants and really focus on the different vowel qualities. Find recordings for each word on Forvo, WordReference, Wiktionary or anywhere you like. Save these audio files (or just keep them open in separate tabs).
Go through the list in order, and repeat each word, using your knowledge of phonetics and the IPA to place your tongue in approximately the right starting position. For example, if I was learning French and I saw the word ceux /sø/, here's what I would think:
/ø/ is a mid-high front rounded vowel. Breaking that down:
- Mid-high means my tongue should be somewhere between the sounds in get /gɛt/ and feet /fit/, as this is the rounded counterpart of /e/ .
- Front means my tongue should be in the front of the mouth, like in get /gɛt/, feet /fit/, cat /kæt/, fate /fe͡ɪt/, and so on, as opposed to words like boat /boʊt/ or lot /lɑt/ or /lɔt/.
- Rounded means my lips are rounded, like in goose /gus/ or boat /boʊt/.
Once you combine all these qualities into one sound, you should end up with /ø/. Now, naturally, you probably aren't going to pronounce exactly this sound on your very first time. That's what the recording is for: repeat after the recording as many times as you can manage until what comes out of your mouth sounds as close as possible to the recording. While doing this, have a mental picture in your mind of where your tongue is – as you try to hone in on the right sound, update your mental picture so that you have an accurate idea of where your tongue is for future reference.
Important note: only move on to the next sound when you can pronounce the previous words correctly. If you go through the list too quickly, you won't have time to accurately learn all the pronunciations, and you'll start with bad habits. You need to take your time with this stage; I've done it for 5 or 6 languages, now, and it still takes me at least a few weeks before I am satisfied with my pronunciation. Good foundations for pronunciation are absolutely vital if you want to develop a good accent, so be ready to invest some time into this stage.
Things to watch out for with vowels
Diphthongs in English
I can't really cover all the possible differences in vowel pronunciations between languages, but as this is an English-speaking blog, I wanted to point out that English is particularly rich in diphthongs, compared to other European languages. In particular, /e͡ɪ/ and /o͡ʊ/ are sneaky saboteurs when it comes to learning /e/ and /o/ in other languages.
Listen to this:
- English "day" /de͡ɪ/:
- French "des" /de/:
- German "dee" /deː/:
- English "so" /so͡ʊ/:
- German "so" /zoː/:
- French "saut" /so/:
In English, there is a noticeable change in vowel quality as the vowel progresses. In French and German, that is not the case. Keep this in mind.
The other thing is that languages vary in how much they reduce vowels. In English, pretty much every unstressed vowel just turns into a sort of schwa sound, like in the first sound of "about" or "believe" .
In French, German, Italian, Swedish, and so on, vowels are not reduced in unstressed position (with the exception of the /ɛ/ vowel, which tends to reduce to a schwa in many languages I've encountered). Every vowel is pronounced the same regardless of whether or not it is stressed.
Reducing vowels when you shouldn't (the way English or Russian speakers, for example, tend to do in other languages) or pronouncing vowels clearly when you shouldn't (like French or German speakers) is also something to watch out for.
/a/ is often pronounced slightly differently in various languages, with some pronouncing it more fronted and some more backed.
- English "sock":
- French "sac":
In French, the /a/ vowel is more fronted, so watch out for small differences like this.
Consonants are generally more straightforward, if you understand the consonant chart and the IPA.
Find the consonant chart on the Wikipedia page of your language's phonology, and make sure you understand what each symbol means. If you don't, click on the symbol and do some research. For instance, if you don't know what /ʒ/ means, click on the link in the table (or just paste it into google), and you should arrive at a page that looks like this. Read through the page, listen to the recordings, and make sure you can reproduce it.
Once you can comfortably produce each consonant, you can compile a minimal pair list, like you did for the vowels. Find words with the different phonemes (Wikipedia often provides a list for you on "language Phonology"), find recordings, and go through them as you did for the vowels.
Alternatively, if you feel like taking a bit of a shortcut, just read through the descriptions underneath the consonant chart. These are often very detailed and should be enough to get you started with the right pronunciations.
Things to watch out for with consonants
Aspiration is a puff of air after voiceless consonants. In English (and most other Germanic languages with the notable exception of Dutch), voiceless consonants are aspirated in word-initial and stressed syllables. This means there is a puff of air after the consonant, rather than an immediate vowel; there is a period of what's essentially whispering before the vocal chords actually start vibrating. I can tell you right off the bat that aspirating your consonants when you shouldn't be (or vice-versa) is one of the most noticeable traits of a foreign accent. If your native language is in any other language group in Europe besides the Germanic languages (e.g. Russian, French, Greek), you probably don't aspirate your voiceless stops. Check your language's phonology page on Wikipedia to be sure, or just hold up a candle in front of your mouth and say a word with a /p/, /t/, or /k/ sound in it.
Wikipedia will often mention if voiceless stops are aspirated. If it doesn't, google it and do some research - it's extra work to sift through google results and academic papers, but sometimes that's what you have to do to get the answers you need. Remember, this extra pronunciation work is an investment that will pay itself off a hundred times over when people confuse you for a native speaker.
If a voiceless consonant is non-aspirated, vocal chords start vibrating immediately after the consonant is over. To native English speakers, this often sounds like a voiced consonant, since we rely on the aspiration cue to distinguish voice pairs like "t" vs. "d" or "b" vs. "p".
To illustrate what I mean, here are some Forvo recordings of voiced vs. unvoiced pairs:
- Russian: "ta" vs. "da" Only distinguished by voicing.
- English: "tea" vs. "d"
Here, the presence or absence of aspiration is by far the most salient feature of the distinction.
There are roughly two main pronunciations for the dental sounds (/t/, /d/, /n/), at least in European languages. One is the laminal pronunciation, found in most non-Germanic languages such as Spanish, Italian, Russian, Greek, and so on. This is the pronunciation that doesn't sound as "hissy". Find some recordings on Forvo of any Spanish words with the letter "t", e.g. "terra" to hear what I mean. Note how in "terra", the /t/ is also non-aspirated - it almost sounds like a /d/, right?
- English (laminal) "d":
- Italian (apical) "di":
I bet that if you've ever heard somebody speaking English with an accent from most of the Western European languages (French, Spanish, German), you've noticed the different way they pronounce the "l" sound.
The distinction, broadly, is that of velarized vs. non-velarized. In American English (and many Slavic languages), /l/ is velarized in all positions. This means that when pronouncing /l/, the back of the tongue is touching the velum, or soft palate. This results in what's commonly referred to as the "dark l" sound.
In other languages, such as French or German, the /l/ is "light", or non-velarized. This means that only the tip of the tongue is making contact with the roof of the mouth while producing the sound.
- English (velarized l):
- Italian (light l):
Rhotics refer to the "r" sounds. These can be grouped into three main categories:
English-style approximant: /ɹ/:
The famous English "r" sound, by far the most noticeable feature of an English accent in basically every other language. If you are a native English speaker, be aware that this is a dead giveaway that you're from an English-speaking land.
- Dutch (some dialects):
Alveolar trill/tap: /r/:
Think Spanish, Italian, or Russian - I'm sure you've heard this sound before. It's formed by repeatedly tapping your tongue against your hard palate (in the front of your mouth). This takes some practice to learn, but it is totally doable.
French and German are famous for this group of sounds. There are actually several sub-types; if you listen to French, German, and Danish, you'll note that they all pronounce it slightly differently, but the principle is the same: it's formed by tightening the back of the through and bringing the back of your tongue to your velum (where /k/ and /g/ are formed).
- German (uvular trill):
- French (uvular fricative):
- Danish (uvular/pharyngeal approximant):
Pro-tip: this sound is often devoiced to /χ/ when adjacent to voiceless consonants. For instance, prier /pʁie/ "pray" is actually realized as [pχie] .
These tend to be very similar or the same as in English, as there's not so much room for variation (aside from degree of voicedness and aspiration, as covered above). I do want to point out, however, that they are often palatalized or retracted to some extent in different languages.
In French and Swedish, for example, they are strongly palatalized in front-vowel environments:
- French "car":
- French "garage":
- Swedish "kat":
In Swedish, "k" and "g" are actually palatalized by default, and only pronounced "normally" (without palatalization) in back vowel environments. Listen to how /k/ is realized in word-final environments in these three languages:
- English "talk" [tʰɑk]
- Russian "так" [t̪äk]:
- Swedish "tack" (strongly palatalized k) [t̪ʰakʲ]:
Keep an eye (ear?) out for these things.
How much of an accent you will have in your language of study depends entirely on you. If you follow the guidelines I've described in this post, and really put the time and effort into digesting all of this material, you can learn to speak with a native or near-native accent.
I've used a lot of technical linguistic vocabulary in this post, so make sure you google anything you don't understand and take the time to learn. These words exist for a reason: they accurately describe properties of sounds that most people who aren't familiar with linguistics don't even think about. If you want to learn a foreign language's pronunciation, learn these words and concepts. Like I've said before, this is an investment that will pay itself off immensely.
I hope this guide will help you on your language-learning journey.