Mastern the Tones in Asian Languages

Master the tones of Mandarin, Thai, Cantonese, Taiwanese, Vietnamese and switch between them with ease.

• Introduction
  • Terminology
    • Language Families
    • Pronunciation Terminology
  • The Sounds You Hear
  • Syllable Structure
  • Tone Category
  • Tone Class
  • Combine Tone Category and Class
  • Writing Tones: Start With Mandarin!
    • IPA Style Tone Writing
    • The Best Way to Write Tones
  • Review the Basic Concepts
• The Universal Tone Table
• Traditional Chinese System
  • Mandarin
  • Taiwanese Hokkien
  • Cantonese
  • Thai
  • Lao
  • Vietnamese
  • Shanghainese
  • Fuzhou
  • Hakka: Sixian
• Large Tone Table
• Further Reading

Introduction

The tones in each of the Asian languages present a different set of challenges for anybody who is learning them, regardless whether your own language has tones or not.

But there is actually one simple system that unifies the layout of tones in all languages, which I will present to you here, then get into the actual tones of each language.

This summary covers the following languages: 12 Chinese languages (including Mandarin and Cantonese), Vietnamese, Thai, and Lao. There are many other smaller tonal languages in the Tai-Kadai and Hmong-Mien families but they should also fall into a similar pattern. If you have data on these languages you can send them to me, and I can add them to the Large Tone Table at the end of this article.

This summary doesn't cover: Japanese, Korean, Mongolian, Uyghur, Cambodian, Tagalog, Indonesian, or Malay. For one reason: they don't have tones. 😃 If you're tone averse, these are for you!

If you intend to learn Cambodian, you'll still want to learn the basics here as this will help you get up and running on being able to read Cambodian vowels, which depends on the class of consonants they appear with. This guide will help you decipher consonant class.

This summary doesn't cover Tibetan or Burmese either because tones are still developing in those languages, and they have few minimal pairs and tones are determined more or less by the syllable structure.

As full disclosure and to answer common questions I get, yes, I have studied all of the languages presented here. No, I have not investigated tone in African languages yet.

The objective of this article is several fold:

1. If you're a learner of Mandarin, put your tones on a mental map so that you can pick up any other Chinese language easier, and get the tone right 90% of the time in Cantonese, Hakka, Shanghainese, you name it.
2. Thai and Chinese tonal systems seem alien to each other on the surface. This guide will unify them in your mind. Mandarin tones are very straightforward, but any learner of Thai has a great disadvantage to the Mandarin learner. This just comes down to different ways the two languages label and count the tones, which are based on the same framework.
3. Vietnamese has a large number of Chinese loanwords. This guide will unify the Chinese/Cantonese tone system so that you can match your Vietnamese tones to Chinese with ease.

Terminology

Language Families

Sino-Tibetan 漢藏語系: this family of 488 languages contains Burmese, Tibetan, and Chinese languages and dialects. Chinese languages that have more tones have a higher percentage of single-syllable words (like Cantonese). Chinese languages that have fewer tones have a higher percentage of two- or three-syllable words (like Mandarin).

Kra-Dai 壯侗語系: this family of 94 languages includes the languages on the borders between Thailand, China, Laos, and (the northern head of) Vietnam. It includes Lao and Thai which are closely related to each other.

Hmong-Mien 苗瑤語系: this family of 39 languages is spoken on the borders of China, Laos, and Vietnam and made up of many smaller languages.

Austroasiatic 南亞語系: this family of 159 languages consists of the Mon-Khmer and Viet-Muong languages of Cambodian (Khmer) and Vietnamese. They are generally non-tonal. Vietnamese went through a period of tonogenesis at one point in its history, adopted a large number of Chinese loanwords from a period of Middle Chinese that predates the earliest of any of the modern Chinese languages. As such, the tones on Chinese words have been kept consistent in their adoption into Vietnamese. But since this happened over a very long period of time, Vietnamese now tends to have multiple pronunciations for most Chinese characters and a lot of these may be exceptions to the general rules we lay out here.

Japanese and Korean: technically these are language isolates, meaning they don't belong to any family of languages. They seem similar to each other, but there is not enough phylogenetic evidence available to prove that they are related. Japanese and Korean have pitch accent (fading in Korean), which is not like tone at all.

To put things into perspective for the western reader, the languages of English, Russian, Greek, and Hindi are all closely related in the same family, while Arabic and Hebrew are closely related in a different family. This article attempts to explain the tonal patterns of vastly different languages across several language families.

Pronunciation Terminology

Onset 音節首/韻首: this is the first sound in a syllable, like the d- in dog, or the pl- in play. We will learn about 4 kinds of onsets.

Class 組別: tone categories can be split up into different classes. Generally speaking onsets determine tone class.

Coda 音節尾/韻尾: this is the last sound in a syllable, like the -g in dog, or the -ay in play. We will learn about 2 kinds of codas.

Category 聲類: this refers to a category of tone. Generally speaking codas determine tone category.

Voicing 清濁音: there are two kinds of consonant voicing, voiced and unvoiced.

• unvoiced /t/ and voiced /d/
• unvoiced /k/ and voiced /ɡ/
• unvoiced /p/ and voiced /b/

Aspiration 氣音: there are two kinds of aspiration, aspirated and unaspirated.

• aspirated /tʰ/ and unaspirated /t⁼/
• aspirated /kʰ/ and unaspirated /k⁼/
• aspirated /pʰ/ and unaspirated /p⁼/
• It's also possible for voiced consonants to have aspiration: bʰ dʰ ɡʰ, but this only survives in some languages of India. But we'll need to remember these in a few minutes, because historically they have affected modern tone systems.

Contour 調値: this is the actual surface sound of the tone, such as high, low, rising, falling, etc. The underlying value of the tone across languages should be the same, but the actual contour will differ. The citation contour and the sandhi contour is usually different! The contour spoken in one city is sometimes different from the contour in another city! But the underlying tone class and category is the same.

Citation 本調: this is the dictionary entry for a tone. Usually the underlying base tone. In most dictionaries this is given with a number, but it is not historically accurate. I will touch on briefly how citations got their numbers, but in order to understand the macro-picture, we need to start from an even deeper place.

Sandhi 變調: tones change when they come in contact with other words. This is a confusing concept, but since it is very language specific, and relies on very surface situations, it is beyond the scope of this article.

The Sounds That You Hear

The sounds that you hear flowing through the air to your ear are what we call the most surface version. These sounds have gone through lots of changes and evolution, and interaction with the words around them to produce the sounds now traveling through the air.

You then interpret these sounds and derive meaning from them. But you can usually attach a meta-tag to these sounds. Even though a tone has changed into a new sandhi contour, you can probably go to a dictionary and look it up by its citation tone. This is META LEVEL ONE.

What this article teaches you is where these meta-level tones came from. In other words META LEVEL TWO. We'll start from meta-level 2 to show you how to get to meta-level 1, your dictionary citation forms. But we'll leave the sandhi surface forms for another article.

Another way to think of it is like evolution. Sounds evolve over time and they end up in vastly different places. When different writing systems try to write the tones, things get obscured and you can't see the relation between them. By analog, this is like showing you the cues to understand how elephants and whales and mice are all mammals.

Syllable Structure 音節架構

The syllable is made up of three parts:

Onset 韻首: This is the first consonant in a syllable. Sometimes it's a cluster like pl-, str-, cr-. The onset affects the tone class. When a syllable starts with a vowel, it has an empty onset usually as a glottal stop. In many writing systems, this is actually written. Since it is empty, not surprisingly it usually looks similar to a zero: Korean ㅇ, Thai อ, Lao ອ. Thai is not as curvy as say, Burmese, so this happens to be the most circular shape in Thai. For comparison, the same onset in Cambodian is shaped significantly different: អ.

Core 音節核: This part of the syllable consists of a core vowel that can have a w-/y- glide in front of it, or a -w/-y glide behind it. These glides are sometimes written as -i-/-o-/-u- depending on the language and spelling customs. They have little bearing on determining tone though and we will largely ignore this part of the syllable.

Coda 韻尾: The end of the consonant affects the tone if it ends in a stop consonant. This puts the whole syllable into a separate category.

Tone Category 聲調類別

In the following table, I use the vowel -a- for simplification. In fact, any vowel can take its place.

Continuous Coda			Stopped Coda
Category 1 -a, -am, -an, -anh, -ang	Category 2 -a, -am, -an, -anh, -ang	Category 3 -a, -am, -an, -anh, -ang	Category 4 -ap, -at, -ak

Why do we care about category?

We care because words that share the same stopped endings all end up with the same tone patterns. Words without stopped endings, all share 3 kinds of other tone patterns. Not only that, but if you follow a single word into a different language or dialect and observe how its pronunciation changes, you can easily bet that all of the words in the same category all have the exact same pronunciation change in the other language. In other words, you can now guess the pronunciation of hundreds or thousands of words with a single observation.

Tone Class 聲調組別

The combination of voicing and aspiration gives us four possibilities:

HIGH unvoiced 清音	aspirated 次	pʰ, tʰ, cʰ, kʰ
	unaspirated 全	p, t, c, k
LOW voiced 濁音	aspirated 全	bʰ, dʰ, jʰ, ɡʰ
	unaspirated 次	b, d, j, ɡ m, n, ny, nɡ, l, r, w, y

If a glottal stop appeared before w- or -y, they could move up to the HIGH class and this is how Mandarin words starting with w- or y- can appear in either class. In Modern Thai, any of the letters in a LOW class can move to HIGH class by adding h- before the consonant.

Since these sounds have changed over time, what we find now in the modern languages is that bʰ, dʰ, jʰ, ɡʰ no longer exist. Instead each of them have become: pʰ, tʰ, cʰ, kʰ.

In Modern Mandarin, all the voiced sounds have become unvoiced. They are still spelled b, d, z/j, g in pinyin though, but these sounds represent the IPA unvoiced p, t, c, k. Meanwhile, pinyin spelling prefers to drop the small ʰ from the aspirated consonants, simply as p, t, c/q, k.

Modern Mandarin 現代漢語(官話):

HIGH unvoiced 清音	aspirated 次	p, t, c/q
	unaspirated 濁	b, d, z/j, g, k, w, y
LOW voiced 濁音	aspirated 濁	p, t, c/q
	unaspirated 次	m, n, l, r, w, y

In Modern Mandarin there is only one tone that belongs to the LOW class (we'll get to it), so this means that words that start b-, d-, z-, j-, g-, k- (with just one exception k-) just won't show up in this LOW class tone. The only ones that are, are those that migrated from another spot (and I'll show you how that happened).

The last group "unaspirated" is what I call the "mnl group". I'll use this term again with the table of many Chinese dialects.

I know, I know, nobody ever referred to high or low class sounds in Mandarin, so why am I doing this? This is in the grander scheme of things and I'm not making this up. It's actually there. It's just an unnecessary thing to learn for anybody just focused on Mandarin. I got fluent in Mandarin without having to know it. But I can't keep things straight when I start learning more languages. I need a framework to help me. Otherwise, and in case you're like me, I have a tendency to just give up on it too easily. This article is to help me just as much as it is to help you.

Notice in the Modern Thai table below, how the unaspirated high class is called "MID" in Thai. Doesn't really matter what name you want to call it, but this is what teachers call it and this is where it fits into the scheme of things. I've split the table up into point of articulation from lips to throat (front to back), followed by the fricatives f, s, h.

HIGH unvoiced 清音	high class	pʰ ผ hm หม	tʰ ถ ฐ hn หน hl หล	cʰ ฉ	kʰ ข ฃ hng หง hw หว	f ฝ	s ส ศ ษ	h ห
	mid class	b บป	d ด ฎ t ต ฏ	j จ	g ก ʔ/y อ
LOW voiced 濁音	low class	pʰ ภพ	tʰ ท ธ ฑ ฒ	cʰ ช ฌ	kʰ ค ฆ ฅ	f ฟ	s ซ	h ฮ
	low class	m ม	n น ณ l ล ฬ (ฦ) r ร (ฤ)	y ย ญ	ng ง w ว

As demonstrated above, you can move a sonorant to a high class by just adding ห h- in front of it. This doesn't actually change the pronunciation of the consonant, but will now let these voiced sonorants adopt high class tones. The same rules apply to Lao (notice the newly created combinations ໜ and ຫຼ). Generally speaking, Lao is a simplified version of Thai spelling. Since the alphabet originates from India, Lao simply discards a lot of traditional letters that Thai keeps from the original Indian spelling which results in a lot fewer letters. Once you learn how to read Thai, it will only take you a few minutes to learn how to read Lao.

Modern Lao:

HIGH unvoiced 清音	high class	pʰ ຜ hm ໝ	tʰ ຖ hn ໜ hl ຫຼ	hny ຫຍ	kʰ ຂ hng ຫງ hw ຫວ	f ຝ	s ສ	h ຫ
	mid class	b ບ p ປ	d ດ t ຕ	j ຈ	g ກ ʔ/y ອ
LOW voiced 濁音	low class	pʰ ພ	tʰ ທ	y ຢ	kʰ ຄ	f ຟ	s ຊ	h ຮ
	low class	m ມ	n ນ l ລ r ຣ	ny ຍ	ng ງ ຫງ v/w ວ

This orthographic complexity just doesn't exist for Mandarin. If we want to write a word like 呢 in a high tone (which would be a phonological exception), we just write the tone over the letter as it's pronounced: nē whereas Thai and Lao would require prefixing it with h-.

Why do we care about class?

We care for the same reasons that we care about category. If you follow one word into another language and observe what's different, chances are that all words that match will have the exact same change. By tracking all of these changes between languages, we can lay out the minimum number of things to track: category times class, or 4x4. Actually instead of 16, we can actually cut that into half. I'll show you how to keep track of 8 class categories with a maximum of 2 exceptions depending on the language.

Follow the link in "Further Reading" to view the article this map appears in.

Combine Tone Category and Class

Note that languages may not express each of the possible syllables in the following table. The -nh ending (borrowed from Vietnamese spelling) is like Spanish ñ, representing the nasal version of a c or j ^{IPA: /ɟ/} (which are halfway between ch and k, dge and ɡ, respectively by voicing).

Class		Continuous Coda			Stopped Coda
		Category 1	Category 2	Category 3	Category 4
HIGH unvoiced 清音	aspirated	pʰam, tʰan, cʰanh, kʰang	pʰam, tʰan, cʰanh, kʰang	pʰam, tʰan, cʰanh, kʰang	pʰap, tʰat, cʰat, kʰak
	unaspirated	pam, tan, canh, kang	pam, tan, canh, kang	pam, tan, canh, kang	pap, tat, cat, kak
LOW voiced 濁音	aspirated (now unvoiced)	pʰam, tʰan, cʰanh, kʰang	pʰam, tʰan, cʰanh, kʰang	pʰam, tʰan, cʰanh, kʰang	pʰap, tʰat, cʰat, kʰak
	unaspirated	mam, nan, nyan, nɡang, lan, ran, wang, yang	mam, nan, nyan, nɡang, lan, ran, wang, yang	mam, nan, nyan, nɡang, lan, ran, wang, yang	map, nat, nyat, nɡak, lat, rat, wak, yak

Now, let's simplify the table and add on the Chinese names for all the headings. In order not to crowd the table, I'll just use the bilabials (p and m) as placeholders from each set for all related sounds.

Class	Category 1 平	Category 2 上	Category 3 去	Category 4 入
HIGH CLASS Unvoiced 陰	pʰ...m p...m	pʰ...m p...m	pʰ...m p...m	pʰ...p p...p
LOW CLASS Voiced 陽	pʰ...m m...m	pʰ...m m...m	pʰ...m m...m	pʰ...p m...p

Writing Tones: Start With Mandarin!

The easiest of the tonal languages, and by far the most popular and most widely spoken language is Mandarin Chinese. It is like the English of the east. It has a simple tonal system, a simple grammar, and a simple sound system. It is different from western languages, but it is simple in structure.

We write Mandarin out phonetically using the pinyin alphabet. The tones written in pinyin resemble the actual tone. Since it visually represents the tone, it's a very easy system to learn and use.

Familiarize yourself with the following, if you haven't already:

MANDARIN TONE NUMBER	CONTOUR	PINYIN SAMPLE	DESCRIPTION	TONE BARS
1	→	ā	high flat	˥ [⁵⁵]
2	↗	á	rising	˧˥ [³⁵]
3	↘↗	ǎ	dipping	˨˩˦ [²¹⁴]
4	↘	à	falling	˥˩ [⁵¹]

IPA Style Tone Writing

Note that some authors do not use the visual method of Mandarin to represent tones and instead use the IPA standard ^{(International Phonetic Alphabet)}. This is especially true for most Thai romanization systems. Consult the following chart to understand how this works. Moving tones that change pitch simply combine a high or low together. If this is hard to read for you, just focus on the last half of the symbol: the last half of ^ is falling and the last half of \ is low.

CONTOUR	IPA NOTATION	DESCRIPTION	TONE BARS
→	á	high flat	˥ ˦ [⁵⁵ ⁴⁴]
→	ā	mid flat	˧ [³³]
→	à	low flat	˩ ˨ [¹¹ ²²]
↗	ǎ	rising	˦˥ ˧˥ ˩˥ ˧˦ ˨˦ ˨˧ ˩˧ ˩˨ [⁴⁵ ³⁵ ¹⁵ ³⁴ ²⁴ ²³ ¹³ ¹²]
↘	â	falling	˥˦ ˥˧ ˥˩ ˦˧ ˦˨ ˧˨ ˧˩ ˨˩ [⁵⁴ ⁵³ ⁵¹ ⁴³ ⁴² ³² ³¹ ²¹]

The disadvantage of this system is that it can't properly represent movements to or from a half height tone, what we call high rising, low falling, etc. Though this can be accomplished through extended sets of Unicode character combinations, few if any authors actually know or use them: a᷄, a᷅, a᷆, a᷇, a᷈, a᷉, a᷃. So not only can you not type them, they are admittedly not that easy to read either. As can see from the table above, the other disadvantage is that you cannot be precise with in-between tones without using tone bars or tone numbers.

The Best Way to Write Tones

When working across languages and mapping things out, in my personal opinion, the best way is to simply use the tone bars unless you're working exclusively within a certain language.

With Unicode support on your computer/device, they should appear linked together. Although some complex tones like this Mandarin 3rd dipping tone which uses 3 bars in a row (˨˩˦) may appear to have a broken link between the 2nd and 3rd bar. This really depends on your device and how current the system that you're using. I assume that smartphones just aren't going to have full support for rarely used sections of the Unicode character set since it is time consuming and consumes memory.

I have a spreadsheet called "Unicode" where I keep all my characters, one per cell, and I copy and paste in order to type. This just wouldn't be possible on a handheld device though, since speed and productivity largely dictate what device I choose.

As you read on, you'll discover why you'll want to avoid "tone counters" (like Tone 1, Tone 2, Tone 3).

I can only think of one instance where tone counters are beneficial: if that's the one and only language you'll ever be using (that has tones). If you are exposed or working with more than one language, then it's a bad idea.

Review the Basic Concepts

1. A tone class is something called "high" or "low" based on the first consonant of a syllable.
2. There are three tone categories that do not end in a stop consonant. They can all have "high" or "low" tone classes.
3. There is one tone category that ends in a stop consonant. This tone category can also have "high" or "low" tone classes. And there is one upcoming exception that you will learn for specific languages.
4. Tones can be written by "tone number", "tone contour over the letter" in either visual format or IPA format, or as a "tone bar". We're going to stick with tone bars because they're easiest to understand at first glance, that is, 一目了然 yímù liǎorán.

So now that you have the basics down, let's get into how everything fits together.

Word Pronunciation vs Tone Pronunciation

You may be wondering how you can transfer the sounds of the whole word to another language like we're learning for tones. Is it possible? To a certain extent, yes, but often times, no. It is very similar to learning how I can learn that an Italian -uo- transfers to a Spanish -ue-, or a Spanish h- transfers to a Portuguese f-.

The tones are probably the hardest part. So once you understand this system, it puts you on level footing for switching between European languages (how individual consonants and vowels transfer), which leaves you with just focusing on learning the exceptions.

It's an extensive topic and I'll assemble all my loose notes into a forthcoming article to cover the non-tonal aspects of the syllable.

The Universal Tone Table

Traditional Chinese System

	平 píng	上 shǎng	去 qù	入 rùp
🌜 陰 High	陰平	陰上	陰去	陰入
🌞 陽 Low	陽平	陽上	陽去	陽入

Any character or word can show up in one of the cells. Any Chinese character that appears there, pretty much "belongs" there. Let's use animals as an example. Examples of animal words in each cell:

	平 píng	上 shǎng	去 qù	入 rùp
🌜 陰 High	🐔 雞 chicken kei	🐶 狗 dog kəu	🐰 兔 rabbit tʰu	🐤 雀 sparrow tsewk
🌞 陽 Low	🐮 牛 cow ngyuw	🐴 馬 horse ma	🐘 象 elephant zyang	🐝 蜜 bee mit

Notice how the Middle Chinese pronunciation (~2000 years ago) of chicken, dog, and sparrow (kei, kəu, tsewk) do not have aspiration. This means that they could be considered "middle class" when we need to make that distinction. For most languages, we don't need to. We'll only need to make small adjustments in the following situations:

• Thai: recall the "middle class" for the first column
• Lao: recall the "middle class" for the third column
• Vietnamese: treat the "nasals/liquids" in "low class" as "high class" including the word for "cow" here, but not other consonants.

We'll get back to each of these points in turn later on.

This table is supposed to be the control constant for translating tones between languages. Every time you learn a new word, you should put it on this table and learn its location on the table.

Very rarely do things get moved to other cells in the table. Mandarin is the only language that does it to a large extent because it has lost so many tones. Some Chinese languages have lost the two middle Low tones and so there tends to be some migration from there. But for the general rule of thumb, this is the table you memorize and keep with you forever.

So when you learn "elephant 象" you should put it into the 去 Column as a Low Class. You'll get the tone exactly right in any Chinese language. Since the word for elephant happens to be cognate across language families into Thai (ช้าง), you'll learn that this table allows the word to fall into place perfectly, giving you both its tone contour and its spelling.

Mandarin

A lot of things have happened to arrive to Modern Mandarin:

1. All the stop endings have disappeared.
2. The words that originally had those stop endings in 入 have merged into 1 of the other 4 remaining tones.
3. Two other tones have disappeared by merging: Low 上 has merged with 去, but a few words have merged with High 上. Low 去 has merged with High 去.

Counting Tones in Mandarin

In Modern Mandarin a lot of tones have disappeared over time. This means there are no syllables that end in -p, -t, -k, or -m.

	平	上	去	入
🌜 陰 High	Tone 1	Tone 3	Tone 4	x
🌞 陽 Low	Tone 2	x	x	x

Notice how Mandarin doesn't count the tones across in a row. Mandarin first counts the first category high and low, then the next two categories. Since all other categories have disappeared, the words in those categories have migrated. What happened to them?

• A lot of consonants in Mandarin only have High Class: b, d, g, j, k, s, sh, z, zh; which means they rarely show up in Tone 2.
• A lot of 入 Category words have migrated into those Tone 2 positions absent from the high class sounds in the point above, and all of these are listed in the next section below. For example, dī and dǐ and dì are all normal High class words in Mandarin. But dí is not! Any word in Mandarin pronounced dí is a migration from 入 Category, which means it was originally pronounced dip, dit, or dik.

Tone Contours in Mandarin

	平	上	去	入
🌜 陰 High	ā ˥	ǎ ˨˩˦	à ˥˩	x
🌞 陽 Low	á ˧˥	x	x	x

The Animal Names in Mandarin

	平	上	去	入
🌜 陰 High	🐔 雞 jī ˥	🐶 狗 gŏu ˨˩˦	🐰 兔 tù ˥˩	🐤 雀 què ˥˩
🌞 陽 Low	🐮 牛 niú ˧˥	🐴 馬 mǎ ˨˩˦	🐘 象 xiàng ˥˩	🐝 蜜 mì ˥˩

Notice how the tones match in the 上 category, the 去 category, and how the 入 category matches the 去 category. This last point is not a general rule you can depend on. The tones that 入 words end up with in Mandarin is quite random. Though the next section on Tone Manifestation in Mandarin can help you determine many of them.

So let's revise our table:

	平	上	去	入
🌜 陰 High	🐔 雞 jī ˥	🐶 狗 gŏu ˨˩˦ 🐴 馬 mǎ ˨˩˦	🐰 兔 tù ˥˩ 🐘 象 xiàng ˥˩ 🐤 雀 què ˥˩ 🐝 蜜 mì ˥˩	x
🌞 陽 Low	🐮 牛 niú ˧˥	x	x	x

Examples of Tone Manifestation in Mandarin

If you're only learning Chinese, you can skip this section. This section is supposed to give a tie-in to the other Southeast Asian languages.

If you have or are learning Thai/Lao and want to eventually learn Mandarin in the future, this section will be of great value to you. Come back in a few days to get our PDF Download.

The following Mandarin words have no 2nd Tone (they are all High Category):

• ban, bang, ben, bian, biao, bin, bing, bu
• che, cui
• dai, dan, dang, dao, dian, diao, ding, dong, dou, duan, dun
• gan, gang, gao, gen, geng, gong, gou, gu, gua, guai, guan, guang, gui, gun
• ha, hong, hun, jian, jiang, jun
• kai, kan, kao, ken, kong, kou, ku, kua, kuai, kuan, kuang, kun
• sa, sai, san, sang, sen, si, song, sou, suan, sun, suo
• shai, shan, shang, shou, shua, shuai, shuan, shuang
• wai, wo, xin, xiu, yong
• zai, zang, zeng, zi, zong, zou, zuan, zui, zun
• zhan, zhang, zhen, zheng, zhong, zhua, zhuan, zhuang, zhui, zhun

Other than the individual words che and cui, I noticed that the k- group is the only aspirated group here.

Missing from the set above are those words that do have 2nd Tone. But the reason they exist at all, is because they migrated from a 入 Category! If you compare any of the following words against Cantonese, Taiwanese, Vietnamese, Japanese, and Korean, they'll all still have their p/t/k endings ^{(p/t/c/ch/k in Vietnamese, p/l/k in Korean, u/t/k in Japanese)}:

• bá, bái, báo, bié, bó, dá, dé, dí, dié, gé, guó, huó, jí, jiá, jiáo, jié, jú, jué, ké, páo, shí (all except 時), shú, sú, xí, xié, zhé, zhí, zhú, zuó

The following Mandarin words have no 1st Tone (they are all Low Category). Notice how l, m, n, r are all voiced sounds:

• hai, hang, hao
• lai, lan, lang, leng, li, lian, liang, liao, lin, ling, long, lu, lü, luan, lun
• mai, man, mang, mei, meng, mian, miao, min, ming, mou, mu
• na, nai, nan, nang, nao, nei, neng, neng, ni, nian, niang, niao, nong, nu, nü, nuan, nuo
• ran, rang, ren, rong, rou, ru, ruan, rui, run

Missing from the set above are those words that do have 1st Tone, because they migrated from a 入 Category:

• chī, chū, chuō, hē, hēi, lēi, pāi

Some words that appear in first tone are onomatopeic in nature, so they don't follow the rules of the rest of the language (the equivalent of adding h- in Thai):

• cū, lāo, mā, māo, mēn, mī, mō, niū, pīng & pāng 🏓, rāng, rēng

Special mention:

• Most words beginning with f- and ending in a vowel migrated from a 入 Category
• All of the following words migrated from a 入 Category and in fact make up the only tone that exist for its particular pronunciation: cā, cè, cù, jué, lè, liè, lüè, miè, niē, niè, nüè, piē, qià, quē, què, rì, ruò, sè, shuō, shuò, tè, xue (all tones), yue (all tones), zé, zéi.

Finally these are the left over words that appear in both 1st and 2nd Tone (both High and Low). What is apparent to me is that all of these are aspirated (except z, zh, w, y):

• cai, can, cang, cao, cong, cun
• cha, chai, chan, chang, chao, chen, cheng, chong, chou, chuan, chuang, chui, chun
• pan, pao, pei, pen, peng, pi, pian, piao, po, pu
• qi, qian, qiang, qiao, qin, qing, qiu, qu, quan
• shao, she, shen, sheng, shui, sui
• tai, tan, tang, tao, tian, tiao, ting, tong, tou, tu, tui, tuo
• wan, wang, wei, wen, wu, ya, yan, yang, yao, ye, yi, yin, ying, you, yu, yuan, yun
• xia, xian, xiang, xiao, xing, xiong, xu, xuan, xun
• za, zan, zao,
• zha, zhai, zhao, zhou, zhuo

Learn to speak Chinese (Beijing) or Chinese (Taiwan) today!

Taiwanese Hokkien (Southern Min) (臺灣閩南話)

Counting Tones in Taiwanese

	平	上	去	入
🌜 陰 High	1	2	3	4
🌞 陽 Low	5	x	7	8

Tone Contours in Taiwanese

	平	上	去	入
🌜 陰 High	a ˥	á ˥˧	à ˨˩	ak ˧˨
🌞 陽 Low	â ˧˥	x	ā ˧	a̍k ˦

How to Memorize Taiwanese Tones in Order

	平 1 (5)	上 2	去 3	去 7	入 4	入 8
Citation	→ (↗)	↘	⇘	⇒	↓	↑
Sandhi	⇒	→	↘	⇘	↑ ↘	↓ ⇘

Animal Names in Taiwanese

	平	上	去	入
🌜 陰 High	🐔 雞 gue ˦	🐶 狗 gáu ˥˧ 🐴 馬 bhé ˥˧	🐰 兔 tòo ˨˩	🐤 雀 qiok ˧˨
🌞 陽 Low	🐮 牛 ghû ˨˦	x	🐘 象 qiūnn ˧	🐝 蜜 bhi̍t ˦

Learn to speak Taiwanese Hokkien today!

Cantonese: Hong Kong (粵語香港話)

Counting Tones in Cantonese

	平	上	去	入
🌜 陰 High	1 ↘	2 ↗	3 →	1 or 7 → 3 or 8 →
🌞 陽 Low	4 ↘	5 ↗	6 →	6 or 9 →

Tone Contours in Cantonese

	平	上	去	入
🌜 陰 High	˥˧	˧˥	˧	˥ ˧
🌞 陽 Low	˨˩	˩˧	˨	˨

How to Memorize Cantonese Tones

	平	上	去	入
Both Classes	↘	↗	→	→

Animal Names in Cantonese

	平	上	去	入 short / 入 long
🌜 陰 High	🐔 雞 gai ˥˧	🐶 狗 gau ˧˥	🐰 兔 tou ˧	🐤 雀 dzoek ˧
🌞 陽 Low	🐮 牛 ngau ˨˩	🐴 馬 maa ˩˧	🐘 象 dzoeng ˨	🐝 蜜 mat ˨

Learn to speak Cantonese today!

Thai

Counting Tones in Thai

	平	去	上	入 short / 入 long
🌜 陰 High 🌗 Middle 🌞 陽 Low	unmarked	◌่ Tone "1" written above the onset	◌้ Tone "2" written above the onset	unmarked

Thai actually has a lot of tones. But only 2 are counted. Thai counts tones at the Meta level rather than at the contour level. This is why it is confusing for students. But once you understand the Meta level categories, then the tones should be easier to master. The tones are written with tiny Indic numerals above the letter, ◌่ for 1, and ◌้ for 2. These tone categories do not align with the traditional Chinese categories in the same order: the 2nd category ◌้ refers to 上, and the 1st category ◌่ refers to 去.

It's worth mentioning that Thai also has a 3rd tone written above the onset in Thai as ◌๊ and a fourth tone written as ◌๋. These fall outside the normal Thai tonal system and are only used for writing words of foreign origin.

Tone Contours in Thai

	平 ◌	去 ◌่	上 ◌้	入 short ◌ก / 入 long ◌าก
🌜 陰 High 🌗 Middle	ข ˧˥ ก ˧	ข่/ก่ ˩	ข้/ก้ ˥˧	ขก ˩ / กาก ˩
🌞 陽 Low	ค ˧	ค่ ˧˩	ค้ ˦˥	คก ˦˥ / คาก ˧˩

Remember the letters that govern each class:

HIGH unvoiced	high class	pʰ ผ hm หม	tʰ ถ ฐ hn หน hl หล	cʰ ฉ	kʰ ข ฃ hng หง hw หว	f ฝ	s ส ศ ษ	h ห
	mid class	b บ p ป	d ด ฎ t ต ฏ	c จ	k ก ʔ/y อ
LOW voiced	low class	pʰ ภพ	tʰ ท ธ ฑ ฒ	cʰ ช ฌ	kʰ ค ฆ ฅ	f ฟ	s ซ	h ฮ
	low class	m ม	n น ณ l ล ฬ (ฦ) r ร (ฤ)	y ย ญ	ng ง w ว

One method for learning the Thai tones and their proper spellings is to listen to the recordings on Glossika (start a Thai course, and just start with audio-only mode). While looking at the tone chart above, listen to the rising and falling tones, and make an educated guess about the tone that's written. If you hear a high falling tone then it's always written with a 2nd tone. If the first consonant is also nasal or liquid, you'll need to add a preceding h- to the syllable to change the tone to high class.

Likewise, if you hear a low falling tone, the first consonant should aspirated or nasal or liquid and the tone written should be a 1st tone. Also notice that although teachers and students refer to a "high" tone in Thai, the contour is actually a high-rising tone.

By using the charts above, you should be able to transcribe what you hear directly back into properly spelled Thai, and at the same time vastly improve your reading ability in Thai. In fact, this is the same way native speakers learn, because they already are familiar with the sounds before they learn to read and write. Conversely, if you learn to read Thai directly from the writing system, you'll just need to remember so many more rules, not to mention you're doing it backwards from native speakers' experiences. But we do have some tips from the greatest foreign master of the Thai language below.

How to Memorize Thai Tones

World-class polyglot Stu Jay Raj claims that one of the easiest ways to remember Thai tones is to think of the High, Middle, and Low class as classes of society. Here is my retelling of his story:

The High Class people are always rising higher (rising tone), while the Middle and Low class stay where they are (flat tone).

When the first tone ◌่ falls down on them, it's like calamity striking out of the sky. The High and Middle classes are actually quite fragile and fall straight to hell (low tone), while the Low classes don't fall that hard (falling tone).

When Robin Hood arrives with the feather in his hat ◌้, he changes social class. He takes from the rich and gives to the poor, causing the High and Middle classes to fall, but the Low class to rise.

Get Stu Jay's Book Here

How to Memorize Thai Letters

Again, Stu Jay Raj gives us insight. He says that the High Class people are inward looking rather than outward looking. Likewise, several Thai letters show us this inward-looking High Class letters:

ผ ถ ฝ

And outward-looking Low Class letters (the 3 pronunciations happen to match the 3 above):

พภ ท ฟ

Also, to tell the difference between ด/d and ค/kʰ is the upward or downward looking loop inside the letter. The downward looking loop is Low Class. The upward looking is not an aspirate, so it is not high, but rather Mid Class which is higher than Low Class.

Animal Names in Thai

Thai is in the Tai-Kadai language family, so vocabulary is just going to be different, that is, some of these words are not cognate with Chinese. Despite this, we can fill in any kind of vocabulary and be able to see how the writing system dictates the pronounced tones. 🐘 is cognate and appears in the original 上 category before it moved in most Chinese dialects to 去.

	平	去	上	入
🌜 陰 High 🌗 Middle	🐶 หมา mǎa	🦢 ห่าน hàan(雁) 🐔 ไก่ kài(雞)	🐝 ผึ้ง phʉ̂ng(蜂)	🐸 กบ kòp
🌞 陽 Low	🐒 ลิง ling		🐘 ช้าง cháang(象) 🐴 ม้า máa(馬)	🐤 กระจอก kràjɔ̀ɔk(雀) 🐊 เงือก ngʉ̂ʉak(鱷)

Learn to speak Thai today!

Lao

Tone Contours in Lao

	平	去	上	入short/入long
🌜 陰 High 🌗 Middle	ຂ ˨˦	ຂ່ ˧	ຂ້ ˧˩ ກ້ ˥˧	ຫັກ ˦ / ຫາກ ˧˩
🌞 陽 Low	ຄາ ˦	ຄ່າ ˧	ຄ້າ ˥˧	ຮັກ ˧ / ຮາກ ˥˩

Remember the letters that govern each class:

HIGH unvoiced	high class	pʰ ຜ hm ໝ	tʰ ຖ hn ໜ hl ຫຼ	hny ຫຍ	kʰ ຂ hng ຫງ hw ຫວ	f ຝ	s ສ	h ຫ
	mid class	b ບ p ປ	d ດ t ຕ	j ຈ	g ກ ʔ/y ອ
LOW voiced	low class	pʰ ພ	tʰ ທ	y ຢ	kʰ ຄ	f ຟ	s ຊ	h ຮ
	low class	m ມ	n ນ l ລ r ຣ	ny ຍ	ng ງ ຫງ v/w ວ

Animal Names in Lao

	平	去	上	入
🌜 陰 High	🐶 ຫມາ mǎ	🐔 ໄກ່ kai	🐝 ເຜິ້ງ pho᷅eng	🐸 ກົບ kób
🌞 陽 Low	🐒 ລິງ líng		🐘 ຊ້າງ sa᷆ng 🐴 ມ້າ ma᷆	🐤 ກະຈອກ káchôk

Vietnamese: Hànôi

Counting Tones in Vietnamese of Hànôi

	平	上	去	入
🌜 陰 High	ngang 🌞 陽 Low mnl	hỏi	sắc	sắc
🌞 陽 Low	huyền	ngã	nặng	nặng

Onsets with mnl in the first tone category move up to High. Since the onset for 🐮 in the Chinese borrowed word is pronounced with ng (an mnl sound), it moves up to the High category. The Vietnamese native word bò is a completely unrelated root.

Tone Contours in Vietnamese of Hànôi

	平	上	去	入
🌜 陰 High	a ˧	ả ˧˨˧	á ˨˦	ák ˧˥
🌞 陽 Low	à ˨˩	ã ˧'˥	ạ ˧'˨	ạ ˨˩

Notice that in Vietnamese the following are not tone marks but rather different kinds of vowels: ă â ê ô ơ ư

They are combined with tone marks to make the following letters: ấ ầ ẩ ẫ ậ ắ ằ ẳ ẵ ặ ế ề ể ễ ệ ố ồ ổ ỗ ộ ớ ờ ở ỡ ợ ứ ừ ử ữ ự

Animal Names in Vietnamese of Hànôi

	平	上	去	入
🌜 陰 High	🐔 雞 kê 🐮 牛 ngưu	🐶 狗 cẩu	🐰 兔 thố	🐤 雀 tước
🌞 陽 Low		🐴 馬 mã	🐘 象 tượng	🐝 蜜 mật

Learn to speak Vietnamese (Northern) or Vietnamese (Southern) today!

Wu: Shanghai (吳語上海話)

Counting Tones in Shanghainese

	平	上	去	入
🌜 陰 High	1	2		4
🌞 陽 Low	3			5

Tone Contours in Shanghainese

	平	上	去	入
🌜 陰 High	tá ˥˧	ta ˧˦		ta' ˥
🌞 陽 Low	na ˨˧			na' ˩˨

All tones in Shanghainese can be determined by the onset consonant of the word. The only minimal pair tone that needs to be marked is the first tone.

Animal Names in Shanghainese

	平	上	去	入
🌜 陰 High	🐔 雞 ji ˥˧	🐶 狗 geu ˧˦	🐰 兔 tu ˧˦	🐤 雀 jia' ˥
🌞 陽 Low	🐮 牛 nieu ˨˧	🐴 馬 ma ˨˧	🐘 象 xhiann ˨˧	🐝 蜜 miî' ˩˨

We don't have Shanghainese yet, but you can learn to speak Wenzhounese today!

Eastern Min: Fuzhou (閩東福州話)

Tone Contours in Fuzhou

	平	上	去	入
🌜 陰 High	˥	˧	˨˩˧	˨˦
🌞 陽 Low	˥˧	˧	˨˦˨	˥

Animal Names in Fuzhou

	平	上	去	入
🌜 陰 High	🐔 雞 gie ˥	🐶 狗 gɛu ˧	🐰 兔 tou ˨˩˧	🐤 雀 cuɔ' ˨˦
🌞 陽 Low	🐮 牛 ngu ˥˧	🐴 馬 ma ˧	🐘 象 cuɔng ˨˦˨	🐝 蜜 mi' ˥

Hakka: Sixian (客家四縣話)

Counting Tones in Sixian Hakka

	平	上	去	入
🌜 陰 High	1	3	4	5
🌞 陽 Low	2	x	x	6

Tone Contours in Sixian Hakka

	平	上	去	入
🌜 陰 High	˧˥	˥˧	˥	˨
🌞 陽 Low	˩	x	x	˥

Animal Names in Sixian Hakka

	平	上	去	入
🌜 陰 High	🐔 雞 gié ˨˦	🐶 狗 gièu ˧˩ 🐴 馬 ma ˧˩	🐰 兔 tū ˥ 🐘 象 xiong ˥	🐤 雀 jiok ˨
🌞 陽 Low	🐮 牛 ngiu̠ ˩	x	x	🐝 蜜 mēt ˥

Learn to speak Hakka (Sixian) or Hakka (Hailu) today!

Large Tone Table

In the following table I use contour numbers instead of tone bars as it's easier on the eyes to determine the exact height of each tone when comparing data. Note that in non-standard languages and dialects, pitches can often differ from person to person by as much as one number, in other words, person A may say '32' ˧˨ but person B may say '21' ˨˩.

Southeast Asia		🌜				🌞
Language	Location	陰平	陰上	陰去	陰入 / long	陽平	陽上	陽去	陽入 / long
Thai ไทย	Bangkok กรุงเทพ	35 3	53	11	1	33	45	31	45, 31
Lao ລາວ	Vientiane ວຽງຈັນ	24	31 53	33	4, 31	44	53	33	33, 51
Tai Lü ᦅᧄᦺᦑᦟᦹᧉ (傣仂, ไทลื้อ, Lự)	Xishuangbanna 西雙版納, Myanmar, Thailand, Laos, Vietnam	55	35ᧈ	13ᧉ	55 / 35ᧈ	51	33ᧈ	11ᧉ	33
Hmong (Hmoob)	China, Laos final tone letter in superscript	5b	3	34v	1s	53j	31g	1s	1m, 213d
Vietnamese tiếng Việt	Hà Nội	33	323	24	35	21 33	3'5	3'2	21

Chinese		🌜				🌞
Language Dialect	Location	陰平	陰上	陰去	陰入	陽平	陽上	陽去	陽入
Mandarin 官話	Běijīng 北京	55	214	51	55, 35, 214, 51	35	51 214	51	51 35
	Hārbīn 哈爾濱	44	213	53		24	53 213	53
	Dàlián 大連	312	213	53	213	35	53 213	53	53
	Jǐnán 濟南	213	55	21	213	42	21 55	21	21 42
	Xúzhōu 徐州	313	35	51	313	55	51 35	51	55
	Kāifēng 開封	24	55	31	24	41	31 55	31	41 24
	Xī'ān 西安	21	53	55	21	35	55 53	55	35
	Lánzhōu 蘭州	21	33	24	24	53	24 33	24	53
SW Mandarin 西南官話	Xīníng 西寧	44	53	213	44	24	213 53	213	24
	Chéngdū 成都	55	53	213	55	21	213 53	213	21
	Guìyáng 貴陽	55	42	13	55	31	13 42	13	55
	Guìlín 桂林	44	54	213	44	21	213 54	213	44
	Kùnmíng 昆明	44	53	212	31	31	212 53	212	31
	Chóngqìng 重慶	55	42	24	11	11	24 42	24	11
	Wŭhàn 武漢	55	42	35	213	213	35 42	35	213
Jianghuai Mandarin 江淮官話	Nánjing 南京	31	22	44	5	24	44 22	44	24
	Yángzhōu 揚州	31	42	55	4	34	55 42	55	4
	Héféi 合肥	212	34	42	4	55	42 34	42	55
Jìn 晉	Tàiyuán 太原	11	53	45	2	11	45 53	45	54
	Hohhot 呼和浩特	31	53	55	43	31	55 53	55	43
Xiāng 湘	Chángshā 長沙	33	41	45	24	13	21 41	21	24
	Shuāngfēng 雙峰	55	21	35	5	23	33 21	33	55
	Shàoyáng 邵陽	55	42	35	5	12	42	24	5
Gàn 贑	Nánchāng 南昌	42	213	35	5	24 55	21 213	21	2
	Líchuān 藜川	22	44	53	13	35	13 44	13	5 3, 5
Huī 徽	Wùyuán 婺源	44	21	35	54	22	31	54	54
Wú 吳	Shànghǎi 上海	53	34	34	5	23	23	23	12
	Sūzhōu 蘇州	44	52	412	4	24	31	31	23
	Hángzhōu 杭州	33	53	55	5	213	53	13	2
	Níngbō 寧波	51	355	355	5	255	355	223, 44	23
	Wēnzhōu 溫州	44	45	42	323	31	34	22	212
Shàojiāng Mǐn 邵將閩	Shàowǔ 邵武	21	55	213	53	33	213 55	213	35 53, 35
N Mǐn 閩北	Jiàn'ōu 建歐	54	21	33	24	33, 21	44 21	44	42
E Mǐn 閩東	Fúzhōu 福州	55	33	213	24	53	242 33	242	5
Púxiān Mǐn 莆仙閩	Pútián 莆田	533	453	42	21	13	11 453	11	4
S Mǐn 閩南	Xiàmén 廈門	55	53	21	1	35	11 53	11	5
	Taiwanese 臺語	44	53	21	32	24	33 53	33	4
S Mǐn 閩南	Cháozhōu 潮州	33	53	213	21	55	35 53	11	4
S Mǐn 閩南	Shàntóu 汕頭	35	53	11	1	55	13 53	31	5
S Mǐn 閩南	Léizhōu 雷州	24	31	21	5	11	33 31	55	1
S Mǐn 閩南	Hǎikǒu 海口	24	213	35	5	21	33 213	33	5
Hakka 客家	Sìxiàn 四縣	24	31	55	2	11	55 31	55	5
Hakka 客家	Méixiàn 梅縣	44	31	52	21	11	52, 44 31, 44	52	5 1, 5
Hakka 客家	Hǎilù 海陸	53	13	11	5	55	33 13	33	32
Hakka 客家	Hong Kong 香港	34	41	52	32	21	52, 34 41, 34	52	5 32, 5
Yuè 粵	Guǎngzhōu 廣州	55, 53	35	33	5, 3	21	13	22	2
	Táishān 台山	33	45	31	5, 3	22	21	55	21
	Kāipíng 開平	33	55	31	5, 3	22	21	55	21
	Nánníng 南寧	55	35	33	5, 3	21	24	22	2
Píng 平	Nánníng 南寧	53	33	55	5, 3	21	24	22	23, 2
Tǔ 土話	Liánzhōu 連州	22	53	11	24	55	33 24	33	22
Mán 蠻話	Wēnzhōu 溫州	44	45	41	4	213	45	22	2

Burman Languages		🌜				🌞
Language	Location	陰平	陰上	陰去	陰入 / long	陽平	陽上	陽去	陽入 / long
Burmese	Yangon ရန်ကုန်	àː22	á̤ː55	á̰ˀ53	ăʔ44	àː22	á̤ː55	á̰ˀ53	ăʔ44
Achangacn ŋa˨˩tʂhaŋ˨˩	Myanmar, China	44	31	35	55	44	31	35	55
Nusunuf nu˧su˧	China 泸水县, 福贡县	33	55	31	53	33	55	31	53
Zaiwaatb tsau˧˩va˥˩	China - Burma border	51	21	55	55	51	21	55	21
Hanihni Haqniqdoq /xa˧˩ɲi˧˩/	China - Burma - Laos - Vietnam border	55	31	33	32 / 44	55	31	33	32 / 44

Further Reading

Register now on Glossika to start your listening and speaking fluency training in any of the following languages. Glossika provides at least 1000 reps for free in any language of your choice, and dialects like Hakka, Taiwanese, Wenzhounese have unlimited reps for free.

• Beijing Mandarin
• Taiwan Mandarin
• Taiwanese Hokkien
• Cantonese
• Thai
• Northern Vietnamese
• Southern Vietnamese
• Wenzhounese
• Hakka (Sixian)
• Hakka (Hailu)

Cracking Thai Fundamentals: A Thai Operating System for Your Mind, Stuart Jay Raj

Change in the Standard Thai High Tone: An Acoustic Study
http://www.manusya.journals.chula.ac.th/files/essay/Phanintra_p.34-44.pdf

Illustration of the IPA: Vietnamese (Hanoi), James P. Kirby
http://www.lel.ed.ac.uk/~jkirby/docs/kirby2011vietnamese.pdf

The Complex Tones of East/Southeast Asian Languages: Current Challenges for Typology and Modelling, Alexis Michaud
https://halshs.archives-ouvertes.fr/halshs-00676251/document

Tone and Phonation in Southeast Asian Languages
http://www.lel.ed.ac.uk/~jkirby/docs/brunelle2016tone-preprint.pdf

Layers of Chinese Loanwords in Proto-Southwestern Tai as Evidence for the Dating of the Spread of Southwestern Tai
https://brill.com/downloadpdf/journals/mnya/17/3/article-p47_4.pdf