A Social View on the Chinese Language

by Jerome L. Packard (Author)
Monographs XXIV, 206 Pages

Table Of Content

←xi | xiii→


Figure 2.1Routes of migration into East Asia—the ‘southern’ route (red and white arrows), and the ‘overlapping’ model, which combines the northern (blue arrows) and southern routes

Figure 2.2The Split of the Proto-Sino-Tibetan Group into Proto-Tibeto-Burman and Proto-Chinese

Figure 2.3Proto-Sino-Tibetan language family tree

Figure 3.1Comparison of Tense and Aspect

Figure 3.2Min ‘Tone Sandhi Circle’

Figure 3.3Phonological Distance from Beijing Mandarin based on textual database comparison

Figure 3.4Hierarchical cluster trees for 15 Mandarin dialects and Chinese regionalects, based on (a) subjective intelligibility ratings, and (b) subjective similarity ratings

Figure 4.1Script types in the development of Chinese writing

Figure 4.2Eight basic Chinese character strokes

Figure 4.3The Chinese character ‘forever’ (yŏng), which contains the eight basic strokes

Figure 4.4The character ‘forever’, including stroke types and stroke order

←xiii | xiv→

Figure 4.5The characters zhe, biang and nang

Figure 4.6Four examples of phonetic-semantic compounds

Figure 4.7Phonetic and semantic radical position configuration

Figure 5.1Stages of Chinese reading acquisition

Figure 5.2Counting Errors in 2-, 3-and 4-year-olds in China vs. U.S.

Figure 6.1Left inferior frontal gyrus activated in Chinese morphological processing

Figure 6.2The ‘dual-route’ reading model as applied to Chinese

Figure 6.3Visual Dyslexia (Chinese)

Figure 6.4Surface Dyslexia (Chinese)

Figure 6.5Phonological Dyslexia (Chinese)

Figure 6.6Deep Dyslexia (Chinese)

Figure 8.1Historical nasal attrition in the Chinese regionalects

←xiv | xv→


Table 1.1Cross-language syllable rate, syllable inventory, syllable complexity and information density parameters

Table 2.1Historical Chinese language periods

Table 2.2Modern Mandarin reflexes of the ‘departing tone’ derivational morphology process in Old Chinese

Table 3.1Numbers of Chinese regionalect speakers

Table 3.2Mandarin initial consonants, IPA and pinyin

Table 3.3Mandarin vowels

Table 3.4The syllable ‘mi’ pronounced with each of the four tones

Table 3.5Some phonetic features in a selection of Mandarin dialects

Table 3.6Mandarin syllable structure

Table 3.7Constraints on the Mandarin syllable

Table 3.8Four Morpheme types in Chinese

Table 3.9Chinese text sample—no word parsing information

Table 3.10Chinese text sample—parsed into words

Table 3.11Chinese complex word types

Table 3.12Voicing contrasts in English, Mandarin and Shanghainese bilabial initial consonants

←xv | xvi→

Table 3.13Wu Tones (Shanghai dialect)

Table 3.14Shanghai two-syllable tone sandhi

Table 3.15Tones of different Yue dialects

Table 3.16Verb-to-noun tonal derivation in Yue

Table 3.17Head-modifier word structure in Yue

Table 3.18Min tones

Table 3.19Min tone sandhi

Table 3.20Gan regionalect Nanchang dialect tone system

Table 3.21Xiang regionalect tones (Changsha dialect)

Table 3.22Kejia tones

Table 3.23Compound word morpheme order in Kejia vs. Mandarin

Table 3.24Jin regionalect tones (Pingyao dialect)

Table 3.25Grammatically-conditioned tone sandhi in Jin (Pingyao dialect)

Table 3.26ge (kʌʔ 圪) prefixation in the Jin regionalect (Pingyao dialect)

Table 3.27Disyllabification in the Jin regionalect

Table 3.28Three demonstrative location pronouns in Jin

Table 3.29Use of the – men plural suffix in Jin (Pingyao dialect)

Table 3.30Regionalect and dialect syllable counts

Table 3.31Average intelligibility ratings for stimuli spoken in 15 Chinese regionalects or Mandarin dialects as judged by speakers of the same 15 regionalects/dialects, with listeners on the X-axis and speakers on the Y-axis

Table 3.32Average phonological similarity ratings for stimuli spoken in Chinese regionalects or Mandarin dialects as judged by speakers of the same 15 regionalects/dialects, with raters on the X-axis and recorded participants on the Y-axis

Table 3.33Percentage of correctly categorized isolated target words

Table 3.34Percentage of correctly identified sentence-final target words

Table 4.1Stages in the development of writing

Table 4.2Simple and complex characters

Table 4.3Character labels that are exemplars of their categories

Table 4.4The six traditional character types– the liu shu (六书)

Table 4.5Phonetic-semantic compound character regularity

Table 4.6The 100% consistent 皇 huáng phonetic-semantic compound family

Table 4.7The inconsistent 工 phonetic-semantic compound family

←xvi | xvii→

Table 4.8The consistent but irregular 兒 phonetic-semantic compound family

Table 4.9Number of characters in historical Chinese dictionaries

Table 4.10Number of characters in contemporary Chinese dictionaries

Table 4.11Character recognition by students in China

Table 4.12Character recognition by U.S. college students of Chinese

Table 4.13Methods of character simplification

Table 4.14Mean character recognition rate as a function of character font size and number of strokes

Table 4.15Properties of single-character words that most affect processing

Table 4.16Examples of partially and fully transparent/opaque complex words

Table 5.1Number and percentage of characters by grade

Table 5.2Percentage of characters in China that are phonetic-semantic compounds and are regular, by grade

Table 5.3Mandarin L1 acquisition order: phonology and lexicon

Table 5.4Mandarin L1 acquisition order: syntax and pragmatics

Table 5.5Number of phonemes in numerals 1–10, Chinese vs. English

Table 5.6Spoken numerals 1–10 in Chinese and English

Table 5.7Numerals 11–20 in Chinese and English

Table 5.8Learners of Chinese—Reading pinyin vs. characters

Table 6.1Summary of cognitive measure performance in the different Chinese dyslexia subtypes

Table 7.1English-to-Chinese phonetic loan words

Table 7.2English-to-Chinese phonetic and semantic loan words

Table 7.3Chinese-to-English phonetic loan words

Table 7.4Primary and secondary school grade names in the U.S. and China

Table 7.5Beijing social groups and their speech characteristics

Table 7.6Effects of professional group and gender on Chinese pronunciation variation, by proportion

Table 7.7Percentage of ‘W’ pronounced as ‘V’ by Mandarin proficiency group

Table 7.8Percentage of ‘W’ pronounced as ‘V’ by regionalect area

Table 7.9Kinship terms in Chinese and English

←xvii | xix→


For all acknowledgements, I do not wish to imply that the persons I thank for helping me would necessarily agree with what I have written. I would first like to thank my brother Perry Packard for all of his helpful conceptual and editorial feedback. Thanks also to Neil Kubler, Bill Baxter, Kevin Miller, Richard Sproat, Shengli Feng, Yaching Tsai and San Duanmu for valuable feedback on the manuscript and help with references. I also want to thank my students Pao-Yuan Chen, Judy Mu, Nobuko Chikamatsu, Taiyuan Tseng, Becky Chen, Liwei Gao, Li-jen Kuo, Fuyun Wu, Dora Lu, Chenwei Zhao, Zhijun Wang, Zhiying Audrey Qian, Sun-A Kim, Yun Yao, Hang Zheng, Junghwan Maeng, You Li and Yihan Zhou, and also Kailu Guan and Yuyun Lei, for playing a role in my development as a scholar and for help with Chinese language and feedback on the manuscript.

←xix | xx→

Gratitude goes out to my colleagues at the University of Illinois, especially my EALC colleagues who have been unfailingly kind and supportive over the years: JaHyun Haboush, Ron Toby, Patricia Ebrey, David Goodman, Bill MacDonald, Nancy Abelmann, Chilin Shih (石基琳), Misumi Sadler, Makoto Hayashi, Zong-qi Cai (蔡宗齐), Kai-Wing Chow, Alexander Mayer and Bob Tierney. A very special thanks to my friend, colleague and mentor C.-C. Cheng (郑锦全) for all his help and support over the years. Thanks to my other Linguistics colleagues: Jennifer Cole, Howard Maclay, Chin Woo Kim, José Hualde, James Yoon, Silvina Montrul, Eyamba Bokamba, Molly Mack, Hans Hock, Rajeshwari Pandharipande, Tania Ionin and Melissa Bowles. Many thanks are due to my wonderful Psychology colleagues Gary Dell, Susan Garnsey, Cindy Fisher and Kay Bock, and to my colleagues in Educational Psychology Dick Anderson, Jana Mason, Kiel Christianson and Liz Stein-Morrow.

Thanks to my colleagues in the field Bob Ramsey, Randy LaPolla, Derek Herforth, Yafei Li, Axel Schüssler, Simon Ager, Tim Xie, Honggang Jin, Debao Xu, Jim Tai, Ovid Tzeng, Jim Huang, Audrey Li, Dana Bourgerie, Scott McGinnis,Ted Yao, Richard Chi, Galal Walker, Mike Everson, Chaofen Sun, Jianhua Bai, Yea-Fen Chen, Margie Chan, Tom Ernst, Charles Lin and Jennifer Liu. My heartfelt thanks go out to my teachers Barbara Lust and Claudia Ross, and to my friends and colleagues Suzanne Flynn and Yuchin Chien. I would like to express gratitude to my colleagues in China: Hua Shu (舒华) at Beijing Normal University, and Xiaolin Zhou (周晓林) and Lu Jianming (陆剑明) at Peking University. Much admiration and appreciation goes to Shou-hsin Teng (邓守信), my first teacher of Chinese language and linguistics. I would also like to acknowledge the foundational role of William S.-Y. Wang (王士元) in our field and thank him for his collegiality and support over the years.

I would like to express love and affection to my children Errol, Sam and Eric, to Emilia the mother of my grandchildren, and to my grandchildren Miriam, Freddie and Temi.

Finally—last and certainly most—thanks to Carol Briner Packard for her photography skills, and for all of her conceptual, editorial and emotional support—and without whom I would be hopelessly flailing in the wind.

←xx | xxi→


2ndsecond position




Aspaspect marker

BCEBefore the Common Era (‘BC’)

CECommon Era (‘AD’)




CSLChinese as a Second Language



DOdirect object

ERPEvent-Related Potential

F0 Fundamental Frequency

FTAFace-Threatening Act

GPCgrapheme to-phoneme conversion rules

IMEInput Method Editor

←xxi | xxii→

IOindirect object

IPAInternational Phonetic Alphabet



LDlong distance

L-IFGleft-inferior frontal gyrus

LPCLate Positive Component (ERP)



NPnoun phrase







PBDPrincipal Branching Direction


Plplural marker

PRCPeople’s Republic of China


PSCPhrase Structure Constraint

Qquestion word or particle

Q-prtquestion particle

Q-wordquestion word

RCrelative clause


SFPsentence-final particle



STMShort-term Memory

Sugparticle of suggestion





VPverb phrase

VWFAvisual word form area

←xxii | 1→



This book offers the reader an introduction to the Chinese language that covers a wide range of topics explained simply in a relatively concise form. In addition to addressing the usual areas of general interest in Chinese such as the history of the language, the Chinese ‘dialects’, Chinese characters and the basics of Chinese grammar, the book also offers a scientific view of Chinese from the perspective of human behavior and society. We will examine how different social groups in China use the language differently, how Chinese children and U.S. college students learn Chinese, and how the Chinese language is processed in the brain. We also look into the origins of Chinese language as revealed by the genealogical evidence, learn how language loss is manifested in Chinese and consider some psychological features of Chinese speech perception and production.

←1 | 2→

The basic sounds and grammar of Mandarin Chinese will be covered to be sure, but we will also consider whether speaking Chinese influences the way a Chinese speaker thinks and whether speaking Chinese affects how quickly children learn to count (hint: it does!). We will cover some basic facts about Chinese reading—such as how character complexity affects reading speed—and we will find out what problems occur in the speech of Chinese stroke patients. We will survey the properties of Chinese language ‘tones’ and find out where they came from. The speech characteristics of different social groups in China will also be described and we will consider whether China’s one-child policy might have affected how family members are referred to in Chinese.

The book will outline the difficulties that dyslexic Chinese children experience in learning to read and write and show how those problems resemble those of dyslexic children in the U.S. It turns out that there are both similarities and differences in Chinese and English dyslexia. One major difference is that in alphabetic languages like English, reading problems are usually sound-based, meaning that dyslexic English readers have problems mapping sounds onto written words. In Chinese readers, dyslexia symptoms are more often morpheme-based, meaning that dyslexic Chinese readers have more problems correctly mapping meanings onto words.

To give an example, a Mandarin speaker with a morpheme-based deficit might be unaware that the meaning of xīn in the spoken word xin1zang4 (心脏) ‘heart’ is different from the meaning of xin1 in the spoken word xin1wen2 (新闻) ‘news’ or the meaning of xin1 in the spoken word xin1ku3 (辛苦) ‘work hard.’ To give a comparable example from English, an English speaker with a morpheme-based deficit might be unaware that the meaning of bank in the word banker is different from the meaning of bank in the word riverbank, or that the meaning of corn in the word cornbread is different from the meaning of corn in the word corner.

We will also unpack what the words Chinese and Mandarin actually mean, because of the ambiguity those terms often entail. Generally the word Chinese referring to the Chinese language is translated as Hanyu 汉语, focusing on the Chinese spoken in China in contrast with the minority languages spoken there. But the term Chinese is often translated Zhongwen 中文, referring either to the written language, the standard spoken language, or any of the several regionalects. And the term Mandarin can mean the standard spoken language in present-day China or it can mean the language that government officials used to communicate with each other during the Ming (1368–1644) and Qing (1644–1911) dynasties.

←2 | 3→

My story starts with the belief that Chinese is a most remarkable language. To begin with, the Chinese orthographic system—Chinese characters—represents one of the most noteworthy writing systems on the planet. Chinese writing is highly structured, containing more phonetic information than people generally realize. Most Chinese characters (over 70%) are phonetic-semantic compounds—characters that contain two parts: a part that represents sound and a part that represents meaning. Using ‘phonics’ to teach Chinese reading would therefore involve focusing children’s attention on the phonetic part of the character. While only about 25% of phonetic-semantic compound characters are completely regular—having a pronunciation that exactly matches the pronunciation of the phonetic part of the character—instruction in reading Chinese may usefully exploit the phonetic information that is present, especially when teaching problem readers.

When conversation turns to the topic of Chinese language, one of the first things that usually comes up is that Chinese has many different ‘dialects’, like Cantonese, Shanghainese and Mandarin. But it is actually more correct to consider them different languages, because speakers of Cantonese, Shanghainese and Mandarin do not generally understand each other. In linguistics, the primary feature that distinguishes ‘dialect’ from ‘language’ is whether speakers of the two speech forms understand each other, that is, whether they are mutually intelligible or not. If they understand each other, the two speech forms are usually considered different, mutually intelligible dialects of the same language, and if they are not mutually intelligible they are usually considered different languages.

Languages like Cantonese and Shanghainese are called fangyan 方言 (literally ‘region-speech’) in Chinese, which means ‘language of a region.’ Fangyan is most commonly—but incorrectly—translated ‘dialect’ following traditional convention, because the fangyan are historically related, because they use the same writing system and because they belong to the same political entity (China). But technically, because they are not mutually intelligible, it is not correct to call these regional fangyan forms ‘dialects’. Some scholars of Chinese language use the label Sinitic languages for the term fangyan—terminology that is correct given their status as mutually unintelligible forms.

In this work I follow the lead of DeFrancis1 and refer to the Chinese fangyan as regionalects, because they refer to the languages spoken in a specific region and are mutually unintelligible. In the pages that follow we examine how the regionalects came to be and compare how close they actually are to each other, using methods ranging from the impressionistic to the scientific. We will see why it is often said that Mandarin is the most simple regionalect and Cantonese is said to be the most difficult.

←3 | 4→

There are other superlatives often heard involving Chinese, like ‘Chinese is the hardest language in the world’, or ‘more people speak Chinese than any other language’, or ‘Chinese has the most difficult writing system in the world’ and so on. Consider instead some facts based on scientific research. As seen in Table 1.1, of the seven major languages examined,2 Mandarin has the highest degree of spoken language information density. This fact, which has nothing to do with Chinese writing, means that spoken utterances that convey a fixed, standardized amount of information are on average shorter in Mandarin than they are in most other languages.3

Looking at the number of syllables in Table 1.1, with 1,274 total syllables (including the four tones) Mandarin has one of the smallest syllable inventories of the world’s major languages. Also the average number of sounds per syllable is greater in Mandarin than in the other major languages, making it—on average—one of the more complicated languages in terms of its syllable structure. So if Chinese has a relatively small number of syllables, and if the average number of sounds per syllable is among the highest in the world, and its spoken language information density is higher than other languages, then it stands to reason that the language would have to be spoken more slowly in order for its information content to be adequately understood. That is why, as seen in Table 1.1, the speech rate of Mandarin (the average number of syllables spoken per second) is the lowest, making Mandarin Chinese one of the world’s most slowly spoken languages—a fact that has nothing to do with Chinese writing.

In the pages that follow we will cover the origins and history of the Chinese language, including where the first speakers of Chinese came from and how the language spread from its Urheimat, or homeland. We will discuss the regionalects—especially Mandarin, China’s lingua franca—and how they came to be, delving into the differences among them and the degree to which they vary in complexity, phonology and grammatical structure.

Table 1.1. Cross-language syllable rate, syllable inventory, syllable complexity and information density parameters. Source: Author. Data from Pellegrino et al. (2011) and Coupé et al. (2019).

←4 | 5→

We will cover Chinese reading and writing in depth, including how the written language began and evolved, and discover how the invention and development of Chinese writing followed the same evolutionary path as the other three writing systems that are known to have independently evolved. We will explore the fascinating structure of the Chinese writing system and find out how many characters are in common use in various contexts today. We will discuss the differences between the ‘traditional’ characters—used in Taiwan and Hong Kong—and the ‘simplified’ characters used in the PRC and elsewhere. We will use scientific criteria to examine how reading Chinese is the same and how it is different from reading alphabetic languages like English.

We will examine how Chinese is learned both by babies in China and by adults studying Chinese as a second language. We will journey through the fascinating area of the Chinese language and the brain. There are so many things that have been said about how the Chinese language is processed in the brain—I will provide the reader a more nuanced view by presenting the scientific facts as they are. As it turns out, there are relatively few ways in which spoken Chinese is any different than the other languages of the world. For example, although Chinese tones seem like musical notes and therefore might be thought to be processed in the ‘music’ (right hemisphere) part of the brain, in fact, because the tones function like consonants and vowels in Chinese, they are processed in the brain more like consonant and vowel speech sounds than musical notes.

And so we set off on this brief excursion, examining the facts and taking a detailed look at one of the world’s most interesting languages. I use simplified Chinese characters except where I am making specific reference to the traditional forms, and for alphabetized Chinese I use pinyin romanization with the tones omitted unless they are necessary for clarity, in which case I write the number of the tone following the syllable. A form enclosed in angle brackets <like this> refers specifically to the orthographic—that is, written as script—form of a word or character.

I am writing this book to offer a popular, uncomplicated view of the nature of Chinese and how it works in everyday life. As in all matters of inquiry, there are often deeper issues and controversies that underlie the simple presentation of surface facts. Those who wish to delve into the more controversial issues will find that the endnotes provide references for those areas of debate. Otherwise feel free to skip the notes and enjoy the narrative account of this most amazing and intriguing language.

←5 | 6→


1DeFrancis (1984).

2Pellegrino et al. (2011).

3See Pellegrino et al. (2011) for a definition of information density in spoken language.


Coupé, C., Oh, Y., Dediu, D., and Pellegrino, F. 2019. “Different Languages, Similar Encoding Efficiency: Comparable Information Rates across the Human Communicative Niche.” Scientific Advances 5, eaaw2594.

DeFrancis, J. 1984. The Chinese Language: Fact and Fantasy. Honolulu: University of Hawaii Press.

Pellegrino, F., Coupé, C., and Marsico, E. 2011. “Across-Language Perspective on Speech Information Rate.” Language 87, no. 3: 539–558.

←6 | 7→


Chinese Language Origins

The Origins of Human Language

Human language is thought to have first appeared as early as 1 million years ago in Africa in the hominin species Homo erectus.1 Our earlier ancestors, such as Homo habilis, may have been adept at making and using tools, but according to the fossil evidence they may not have possessed the cranial capacity to accommodate the larger brain needed for advanced cognition. But it is thought that H. ergaster and H. erectus did possess the advanced thinking skills that eventually gave rise to talking. H. ergaster and erectus may have been the first species of homo that were able to speak aloud, or, ‘externalize’ their thoughts—an ability made possible by their larger brains and the need for them to share thoughts with their fellow group members. It is for this reason that Homo erectus descendants such as floresiensis, denisova, neanderthalis, heidelbergensis and rudolfensis—who inhabited different Eurasian sites after Out of Africa II2—most likely already possessed the capacity for human spoken language.

←7 | 8→

Scientists are able to infer the migration patterns of these early hominins based on DNA3 obtained from ancient human remains and present-day populations. When taken together with historical, anthropological, archaeological and linguistic evidence, this genetic information helps us deduce the characteristics of human populations and their movements over time.4 There is general agreement that most of the populations that now occupy East Asia descend directly from a large group of H. sapiens carriers of the maternal DNA haplogroup L3 that is believed to have emerged out of Africa around 80–90 thousand years ago.5

Although many H. sapiens migrations occurred after Out of Africa II around 300,000–200,000 BCE, genetic evidence indicates that the L3 haplogroup migration that occurred around 80,000–90,000 BCE represents the most recent large-scale African exodus by H. sapiens.6 The haplogroups presently found in north Asia, south Asia and Oceania diverged from their ancestor L3 haplogroup approximately 65,000 BCE.7 Let us note in passing that, if these L3 migration data are accurate, then the L3 migrants exiting Africa are not likely to have been speakers of an Afroasiatic language—a language family that includes Arabic, Hebrew, Egyptian, Akkadian and Aramaic—because although Afroasiatic is the earliest attested language family8 and Afroasiatic languages were spoken in the same areas of northern Africa and the Middle East where the L3 haplogroup exodus out of Africa is thought to have occurred,9 nonetheless the date of proto-Afroasiatic was as early as 16,000 BCE, which would have been more than 50,000 years later than the posited L3 haplogroup exodus.10

Out of Africa into Asia

There are two primary routes posited for the H. sapiens L3 haplogroup migration from Africa into Asia and Oceania11 (see Figure 2.1). The first is the Southern Coastal route (Figure 2.1, red arrows), which proposes that migrants crossed the Red Sea into the Arabian peninsula at the Bab-el-Mandeb strait into what is now Yemen, following the southern coastlines of the Arabian peninsula, Persia and India into southeastern Asia and on to Oceania. The second posited route is the Northern route (Figure 2.1, blue arrows), which proposes that migrants left Africa via the Levant, either by going directly across Suez or by crossing the Bab-el-Mandeb strait, moving northward to the Levant and then eastward into central Eurasia. There is also an ‘overlap’ model12 that proposes independent migrations into Asia along both the Southern Coastal and Northern routes at different but overlapping times.

←8 | 9→

The Southern Coastal route migration is believed to have taken place around 70,000–65,000 BCE,13 eventually branching north and south around 60,000–50,000,14 perhaps in the vicinity of what is now southeastern India or southwestern Bangladesh. The northern branch of the Southern Coastal route (Figure 2.1, white arrows) may have travelled along what would eventually become known as the Ancient Tea Horse Road, occupying southwestern Sichuan around 40,000–30,000 BCE. The southern branch of the Southern Coastal route (Figure 2.1, yellow arrows) would have travelled through present-day Myanmar (Burma), moving southward through New Guinea and is thought to have arrived in Australia around 60,000 BCE.15

The northern branch of the route taken by the Southern Coastal group (Figure 2.1, white arrows) is thought to have moved north out of southwest Sichuan during the Paleolithic (around 23,000–19,000 BCE),16 eventually arriving at the upper and middle reaches of the Yellow River around 18,000–16,000 BCE.17 The Northern group is believed to have reached the upper and middle reaches of the Yellow River a bit later, around 16,000–13,000 BCE, travelling east from western Eurasia north of the Tibetan Plateau, during the period of post-glacial warming that followed the Last Glacial Maximum between 29,000–14,000 BCE.18 These two groups are then thought to have intermingled in the area of

Figure 2.1. Routes of migration into East Asia—the ‘southern’ route (red and white arrows), and the ‘overlapping’ model, which combines the northern (blue arrows) and southern routes. Source: Author. Data from Sanchez-Mazas et al. (2011), Disotell (1999), Su et al. (1999) and Clarkson et al. (2017).

←9 | 10→

the Yellow River central plain around 14,000–12,000 BCE, perhaps forming the Di-Qiang ethnic group19—the members of which have been proposed as the first speakers of Proto-Sino-Tibetan.20

After occupying the Yellow River plains territories, the members of the Proto-Sino-Tibetan group are thought to have then diverged at around 5,200 BCE21 (see Figure 2.2), with some members (eventual Proto-Tibeto-Burman speakers, the light blue arrows in Figure 2.2) moving south and some (eventual Proto-Chinese speakers, the green arrows in Figure 2.2) moving eastward.22 The Proto-Tibeto-Burman group would have moved southward along the western-Sichuan/Yunnan migration corridor to eventually (c4,000–700 BCE) arrive in Tibeto-Burman lands and become known as the Tibeto-Burman people.23 The Proto-Chinese group expanded eastward to occupy the middle and lower reaches of the Yellow River plains area (the green arrows in Figure 2.2), eventually developing into the population we now know as the Han Chinese.24 This suggests that a language we may call Proto-Chinese was spoken in the central plains of China—along the middle and lower reaches of the Yellow River—beginning around 4,000 BCE, suggesting that a Proto-Chinese Urheimat, or homeland, existed in the plains of central China along the middle and lower reaches of the Yellow River around the year 4,000 BCE.25

Figure 2.2. The Split of the Proto-Sino-Tibetan Group into Proto-Tibeto-Burman and Proto-Chinese. Source: Author. Data from Li et al. (2017), Wang et al. (2014).

←10 | 11→

The Traditional Chinese Language Historical Periods

The Proto-Chinese language referred to earlier in this chapter was presumed to have been spoken on the central plains of China along the middle and lower reaches of the Yellow River around 4,000 BCE. Proto-Chinese is posited to be a language ancestral to Old Chinese and the present-day Chinese regionalects whose existence is inferred based on pre-historical anthropological and genetic data rather than on written sources.

Based on written historical records, the Chinese language is traditionally divided into the three historical periods26 Old Chinese (1250–221 BCE), Middle Chinese (500–1200 CE) and Modern Chinese (1300 CE– present) based on written materials like the Book of Documents (尚书), the Book of Odes (诗经), the Shuowen Jiezi (说文解字) dictionary and the rhyme books and tables of the 7th–14th centuries. Old Chinese may be further divided into Early Old Chinese (1250–700 BCE), and Late Old Chinese (700 BCE–221 BCE).27 Middle Chinese is usually subdivided into Early Middle Chinese (500–700 CE) and Late Middle Chinese (900–1200 CE)28 and the Modern period may be subdivided into Early Mandarin (1300–1911 CE) and Modern Mandarin (1911– present), as seen in Table 2.1.

Old Chinese

Old Chinese is usually defined using phonological criteria, beginning with the first written evidence produced around 1250 BCE. Old Chinese is reconstructed as having a three-way distinction in obstruent initials: voiceless unaspirated (e.g. *p-*t-*k-), voiceless aspirated (e.g. *ph-*th-*kh-), and voiced (e.g. *b-*d-*g-), and in addition is reconstructed as having morphological nominalizing or transitivizing processes signaled by voicing prefixation or syllable-initial ablaut voicing

Table 2.1. Historical Chinese language periods. Source: Author.

←11 | 12→

processes.29 Old Chinese is thought not to have had tones,30 but is reconstructed as having four different syllable codas (endings)—1. vowel/nasal, 2. glottal stop [ʔ], 3. voiceless fricative [h] 31 and 4. [p] [t] [k] stop consonants—which would eventually become the four Middle Chinese tones level, rising, departing and entering respectively,32 with those four syllable codas believed to contain remnants of earlier phonological processes that had a morphological function.

The maximum possible Old Chinese syllable structure was CCCGVCCC (C= consonant, G = glide, V = vowel),33 in contrast with the present-day maximal Mandarin syllable structure of CGVXT.34 Most words during this period were phonologically complex but consisted of a single syllable, and the language did not begin to combine single-syllable words into compound words until around 1000 BCE,35 with such compounding occurring on a much larger scale beginning in the Han dynasty, 206 BCE–200 CE,36 to compensate for the loss of phonological contrast caused by its simplified syllable structure and the simplification of the phonological system in general.

Middle Chinese

The Middle Chinese of the 4th–11th centuries CE mainly refers to the time period that contained the phonological system underlying the two sources of rhyme information: the Qieyun 切韻, compiled in 601 CE and the Guang Yun 广韻, compiled in 1007 CE. The literary standard represented by the Qieyun and the Guang Yun and their revisions served as the rhyming standard for centuries of poetic composition until modern times. Middle Chinese may be further divided into Early Middle Chinese—the 4th–6th century period preceding the Qie Yun—and Late Middle Chinese, the 8th–11th century period preceding the Guang Yun. Early Middle Chinese had the four tones level, rising, departing and entering, which are thought to have appeared some time between 200 BCE and 400 CE.37

Early Mandarin to Modern Mandarin

←12 | 13→

Early Mandarin represents the language spoken in the capital Chang'an 长安 (today’s Xi’an 西安) around the year 1300 CE as represented by the pronunciation of the words in the rhyme book Zhongyuan Yinyun 中原音韻Rhymes of the Central Plain published in 1324. By that time, Mandarin had lost the final – p, –t and –k consonants characteristic of the entering tone, which had weakened to become a syllable-final glottal stop [ʔ] and been distributed among the three remaining tones level, rising and departing. The level tone was further divided into an upper level (present-day Mandarin first tone) and lower level (present-day Mandarin second tone) tone conditioned by the voicing feature on the syllable-initial consonant.

The early form of Mandarin spoken around 1300 CE still had the final nasal consonant [m] , which would be dropped sometime between 1450–1600.38 That early form of Mandarin had not yet developed the palatal consonant series [t͡ɕ], [t͡ɕʰ] and [ɕ] represented today in pinyin as <j>, <q> and <x> respectively. Those palatal consonants would arise in Beijing Mandarin speakers some time between the beginning of the Qing dynasty in 1644 and the mid-to-late 1700s, when either the apical consonants [t͡s], [t͡sʰ] and [s] (pinyin <z>, <c> and <s>) or the velar consonants [k], [kʰ] and [x] (pinyin <g>, <k> and <h>) were palatalized when they occurred before the high front vowels [i] or [y].39 Modern Mandarin, from the 20th century onwards, had generally lost the final glottal stop on the entering tone (but see Table 3.5 and accompanying discussion) and had essentially the same pronunciation as contemporary Mandarin, defined as the vernacular language spoken at the time of the May 4th movement (May 1919).

More finely-articulated periodizations of Chinese language have been proposed which further subdivide the historical periods, offering differing chronological coverage according to the different types of linguistic evidence used, that is, phonological, textual, lexical, or syntactic.40

From Proto-Chinese to Old Chinese

Earlier in this chapter, Proto-Chinese is proposed as the ancestor language of the present-day Chinese regionalects, spoken in the middle and lower regions of the Yellow River plains around the year 4,000 BCE. What, then, would be the posited relationship between this ‘Proto-Chinese’41 and the ‘Old Chinese’42 described earlier?

←13 | 14→

Proto-Chinese is clearly ancestral to Old Chinese, and it is unlikely that the two would have overlapped in time. Old Chinese is posited to be the ancient form of the Chinese language spoken in the c1000-year period between c1250 to c200 BCE.43 Old Chinese is usually thought of as the reconstructed language that is ancestral to the regionalects, in part because, in addition to other forms of evidence, Old Chinese is reconstructed by applying the comparative method44 to those very Chinese regionalects. We should assume, then, that if there was a form of Proto-Chinese spoken in the central plains around 4,000 BCE, that it would have preceded the Old Chinese spoken between c1250 BCE to c220 BCE by at least a millennium, and that such a language and its precursors would have contained the proliferation of syllable-initial voicing and other phonological processes as well as the plethora of other syllable-final phonological processes posited for Old Chinese (see, e.g., Table 2.2 and accompanying discussion), and that many of those processes are likely to have possessed significant morphological functionality.45

If we are able to infer the morphological complexity of Proto-Chinese and Old Chinese, and we know that between Old Chinese and, for example, Middle Chinese, the language had lost a great deal of this morphological and phonological complexity, then the trajectory of linguistic change at the time of Old Chinese was most likely moving in the direction of phonological and morphological simplification. It stands to reason, then, that projecting from Old Chinese back to the time of Proto-Chinese and Proto-Sino-Tibetan, the language of that time would almost certainly have had an even richer morphological system. This is in fact the explicit claim made by a number of scholars of Chinese linguistics,46 including the sweeping conclusion that ‘… word derivation by prefixes and suffixes constituted one of the most outstanding characteristics of Proto-Chinese.’47

The Origin of Chinese Tones

If we assume that Old Chinese did not have tones, then when did the Chinese tones appear and where did they come from? The tones of Chinese are thought to have originated between late Old and early Middle Chinese—sometime between 200 BCE and 400 CE48—by following a phonological process of tonal evolution known as tonogenesis.49 According to tonogenesis theory, certain consonants at the ends of syllables turned into tones that spread over the entire length of the syllable. According to this theory, syllable-final consonants caused perturbations in the pitch (vocal cord vibration, fundamental frequency, or F0) of the syllables, which first became salient, and then became phonologically contrastive (phonemic) when the consonants were lost during the normal process of syllable-final phonetic attrition. This process may have been completed by the late Han dynasty (c200 CE)50 or 400 CE51 and certainly had shifted from consonant status to tone status by the time of Late Middle Chinese.52

←14 | 15→

Middle Chinese had four tones: level (generally, present-day Mandarin first and second tones), rising (generally, Mandarin third tone), departing (generally, Mandarin fourth tone) and entering (lost in modern Mandarin, generally retained in the southern regionalects). Phonological reconstruction reveals that Old Chinese syllables ending in a vowel or a nasal gave rise to the Middle Chinese level tone, syllables ending in a glottal stop [-ʔ] came to have the rising tone,53 syllables ending in the voiceless fricative [-h] evolved to have the departing tone and syllables that ended in the stop consonants [-p], [-t] and [-k] would have the entering tone.

The four tones are thought to have developed from different consonants in the tonogenesis process described as follows. The voiceless fricative [h] coda caused a falling F0 pitch that strengthened to become the departing tone after the final [h] was dropped. The glottal stop [ʔ] final caused a rising F0 pitch that became phonemic after the [ʔ] was dropped. The F0 of the syllable-final vowel and nasal codas retained a simple level F0 pitch, contrasting with the falling and rising F0 pitch contours, while the length of the entering tone was naturally short due to the length-restricting properties of the [p], [t] and [k] syllable-final stops.

To give an example, consider the word 钉 ‘to nail’, pronounced ding4 (with a ‘falling’ or ‘fourth’ tone) in modern Mandarin. The Old Chinese pronunciation of 钉 is reconstructed54 as *tˤeŋ-s with an -s suffix (indicating denominalization) at the end of the syllable. Over time, the syllable-final -s became a voiceless fricative [h] , which caused a falling pitch (F0) at the end of the syllable. When that [h] then disappeared due to normal syllable-final weakening, the falling pitch became perceived as a distinctive sound, a falling tone. This change resulted in the current Mandarin pronunciation ding4 with a falling tone.

It is unlikely that the four tones existed at the time of Old Chinese (but see note 48), and if they were present at all, they likely would only have appeared in the form of the associated syllable-final glottal and fricative phonation features of what would eventually become the rising and departing tones respectively. Between the time of Old and Early Middle Chinese, the F0 changes caused by the syllable-final glottal and fricative phonation features would have evolved into phonemic tonal categories while the phonation features themselves gradually decayed and finally disappeared in Middle Chinese—a process that most likely would have extended across several centuries.55

Mandarin Reflexes of Derivation in Old Chinese

←15 | 16→

There is general agreement among scholars of Chinese linguistics that the Middle Chinese four tones (level, ping2平; rising, shang3上; departing, qu4去; and entering, ru4入) reflect the remnants of an ancient morphological derivation system that existed in Proto-Chinese and its parent language Proto-Sino-Tibetan.56 Specifically, the glottal stop coda [ʔ] that evolved into the rising tone (上声) is thought to be a reflex of an ancient grammatical suffix, the meaning of which is uncertain.57 Also, the voiceless fricative [h] coda that evolved into the departing tone—modern Mandarin fourth tone—is thought to have come from an earlier [-s] suffix58 that derived deverbal nouns and denominal verbs, among other possibilities.59

Those Mandarin fourth-tone words constitute evidence in modern Chinese of a productive morphological process in Old Chinese. The fourth-tone reflexes are clearly visible in modern Mandarin in the form of word pairs written with a single character, with one member of the pair having a departing tone with a clear derivational function that is still visible in the modern language.

A Chinese character will often have two readings (like the character 钉 having the two readings ding4 and ding1)—with one of the readings being a fourth-tone word—that has an obvious derivational meaning today and also represents the Old Chinese derivational process. As seen in Table 2.2, the derivational [-s] suffix in Old Chinese became a falling tone (fourth tone) in modern Mandarin. The meaning of the [-s] derivational suffix included both deverbal nominalization—creation of nouns from verbs (the top half of the examples in Table 2.2) and denominal verbalization—creation of verbs from nouns (the bottom half of the examples in Table 2.2).

As seen in Table 2.2, the base and derived words are written in most cases with the same Chinese character. The only difference between the base words without departing tone and the derived departing tone words in the rightmost column in Table 2.2 is that all of the fourth-tone words have been independently reconstructed as having the [-s] derivational suffix in Old Chinese.60

The modern Mandarin fourth-tone reflexes presented here constitute evidence of derivational morphology in Old Chinese. The relationship between the base and derived words outlined in Table 2.2 indicates a derivational process that is patently visible in the modern language upon examination.

The Linguistic Affiliation of Chinese

The model presented here assumes the existence of Proto-Sino-Tibetan, Proto-Tibeto-Burman and Proto-Chinese. It is believed that Proto-Sino-Tibetan was a monosyllabic61 proto-language that subsumed the Chinese and Tibeto-Burman languages and was spoken around 5,200 BCE62 along the upper and middle reaches of the Yellow River. This model presumes that Proto-Tibeto-Burman was spoken by the group that migrated south to Tibeto-Burman lands via the western-Sichuan/Yunnan Tibeto-Burman migration corridor, arriving in what would become Tibeto-Burman lands some time between 4,000 and 700 BCE.63 Finally, as discussed earlier in the chapter, it is proposed that Proto-Chinese—which

←16 | 17→

Table 2.2. Modern Mandarin reflexes of the ‘departing tone’ derivational morphology process in old Chinese. Source: Author.

a Old Chinese pronunciations from Baxter and Sagart (2014).

I have posited to be the direct ancestor language of the present-day Chinese regionalects—was spoken along the middle and lower reaches of the Yellow River around 4,000 BCE.

←17 | 18→

Figure 2.3. Proto-Sino-Tibetan language family tree. Source: Author.

The model described here follows the predominant assumption in East Asian linguistics that Sino-Tibetan is at the top of the language family tree (see Figure 2.3), and that the first branching of that tree consists of a Tibeto-Burman branch on one side and a Chinese branch on the other. Included in the Tibeto-Burman branch are languages like Tibetan, Burmese, Himalayan and Nepalese, while the Chinese branch would contain the Chinese regionalects Mandarin,64 Wu (e.g., Shanghainese), Yue (e.g., Cantonese), Min (e.g., Taiwanese), Hakka (Kejia), Jin, Gan and Xiang.

Chinese Is Not Related to Japanese

People often think that the Chinese and Japanese languages are closely related, an assumption that is incorrect. The language family relationships described above should leave no doubt that Chinese is related to, for example, Tibetan and Burmese, and not Japanese. The reason Chinese and Japanese are often thought to be related is because of the geographic proximity of the two countries, the physical resemblance of the Chinese and Japanese people and cultural similarities such as the copious loan words borrowed from Chinese into Japanese and the use of Chinese characters in Japanese writing. Those who make this assumption are usually surprised to hear that Japanese is more closely related to languages like Turkish and Mongolian than to Chinese. If we set aside words borrowed from Chinese into Japanese and the use of Chinese characters in Japanese script, then it is abundantly clear that the Japanese and Chinese languages have no familial relationship. In addition to the linguistic evidence, the absence of a close relationship between the Chinese and Japanese people is also supported by the genetic evidence, which invariably shows that Japanese populations pattern genetically with northern Asian groups rather than with Chinese groups.65

←18 | 19→

In the next chapter we refocus our attention on Chinese, and discuss how the Proto-Chinese speakers from the Yellow River valley around 4,000 BCE made their way southeast to become speakers of the Chinese regionalects as we know them today, and how the Proto-Chinese group that remained behind in the Yellow River valley became speakers of the language we now call Mandarin.


1Everett (2017).

2Out of Africa II refers to the migration of anatomically modern humans (Homo sapiens) out of Africa after their emergence at c. 300,000 to 200,000 years ago, in contrast to Out of Africa I, the migration of archaic humans from Africa to Eurasia between roughly 1.8 to 0.5 million years ago.

3Mitochondrial, or ‘maternal’, mt-DNA and nuclear or ‘paternal’, Y-DNA.

4These proposed migrations are presented as group movements because it is heuristically useful to do so. In reality, we do not know the extent to which the migrations represented discrete groups units moving together across specific places at certain periods of time. It could be that the posited migrations reflect tendencies of population movement or the movement of individuals rather than the migration of discrete groups per se.

5Soares et al. (2012).

6Soares et al. (2012).

7Gonzales et al. (2007).

8Crass (2009), p. 12.

9A single parent language can be posited under the Out of Africa II hypothesis because the theory asserts a largely single-group exodus from Africa. Under the multiregional hypothesis, the individual Homo erectus groups or their descendants such as floresiensis, denisova, neanderthalis, heidelbergensis and rudolfensis are more likely to have spoken different but possibly related languages—related because they would have diverged from a posited common ancestral language spoken by Homo erectus prior to leaving Africa.

10Ehret (2002), p. 35–36.

11Cavalli-Sforza and Feldman (2003).

12Di and Sanchez-Mazas (2011).

13Sanchez-Mazas et al. (2011).

14Disotell (1999), Su et al. (1999).

15Clarkson et al. (2017).

16Peng et al. (2011), Wang et al. (2014), Zhong et al. (2010), Shi et al. (2008), Jin and Su (2000).

←19 | 20→

17Wang et al. (2014); Shi et al. (2013).

18Zhong et al. (2011), Sanchez-Mazas et al. (2011).

19Wang et al. (2014).

20Norman (1988), p. 17.

21Sagart et al. (2019) give an estimated time of 5,400–5,200 BCE for proto-Sino-Tibetan, with a homeland in the Yellow River valley in the eastern half of the north Chinese loess plateau, during the final stages of Cishan culture (6500–5000 BCE) or the initial stages of Yangshao culture (5,000–3,000 BCE). W. S.-Y. Wang (2015, p. 26) and LaPolla (2017) place proto-Sino-Tibetan at c4,000 BCE, while Norman (1988, p. 183) estimates it to be c1,500 BCE.

22Li et al. (2017); Wang et al. (2014).

23Wang et al. (2018); Li et al. (2017); Wang et al. (2014); Zhao et al. (2009).

24Wang et al. (2014), p. 12.

25Packard (2019).

26Baxter (1992), p. 14–15. I have given precise dates for the boundaries of these time periods for convenience. For example, Baxter defines Early Middle Chinese as the language codified by the Qieyun 切韻, published in 601 CE, while I have set the beginning of that period as being around 500 CE, under the assumption that the Qieyun represented a language that was probably spoken both preceding as well as following its publication.

27Sagart (1999), p. 7.

28Baxter (1992), p. 14–15. I have taken the liberty of substituting precise dates for reference to the beginnings and endings of Chinese dynasties for convenience.

29Bodman (1980); Mei (2012); Baxter (1992); Sagart (1999); Pulleyblank (2000); Baxter and Sagart (1998), Sagart and Baxter (2012); Handel (2012).

30Although Sagart (1998) suggests that the level, rising and entering tones may already have existed in Old Chinese, and McCoy (1980) proposed that the language ancestral to Proto-Chinese had two tones, one high and one low.

31Baxter and Sagart (2014) reconstruct the earlier form *-s rather than *-h in Old Chinese.

32See, for example, Sagart (1998), Baxter (1992).

33Ting (1979), cited in Feng (1998), p. 224. Baxter (1992, pp. 177–183) gives the maximum Old Chinese syllable structure as CCGVCC.

34‘X’ represents a coda that may be either a high vowel ([i] or [u]), an [ɻ], an [n], or an [ŋ], and the ‘T’ is one of the four Mandarin tones.

35Cheng (1981).

36Feng (1998).

37See note 48.

38Coblin (2000).

←20 | 21→

39Coblin (2000), pp. 540–542. Coblin cites Morrison’s observations that at the beginning of the 19th century, Nanjing speakers of Mandarin distinguished apicals and velars before high front vowels, but had no palatals, while speakers in Beijing already had the palatal series, a trait he attributed to contact with the Manchu governing class. It is reasonable to presume that the Jin area would have been under greater Manchu influence, implying even earlier palatalization of velars for the Jin-area speakers than for speakers in Beijing.

40See, for example, Peyraube (2015).

41Bodman (1980) defines Proto-Chinese as the language form that existed between Proto-Sino-Tibetan and Old Chinese.

42As reconstructed by, for example, Baxter and Sagart (2014) or Zhengzhang (2013).

43Baxter and Sagart (2014).

44It is considered to be a form of the comparative method because it involves the comparison of languages with common descent from a shared ancestor to infer the properties of that ancestor. The Old Chinese reconstruction of Baxter and Sagart (2014) used the comparative method, because it included data from regionalects and from Chinese loan words in other languages.

45For a good summary of the Old Chinese morphological processes see LaPolla (2017) or Baxter and Sagart (1998). For a good summary of Proto-Chinese morphological processes see Sagart (1999) or Schüssler (1976).

46For example, Bodman (1980), Sagart (1999).

47Schüssler (1976), p. 129.

48Xu (1991) p. 269, cited in Feng (1998), fn. 19, estimates the date at between 200 BCE and 200 CE. Baxter (p.c.) estimates the date at around 300 or 400 CE.

49Zhu (2015). The term ‘tonogenesis’ was coined by James Matisoff.

50Xu (1991), p. 269.

51Baxter (p.c.).

52Baxter (1992, p. 185).

53Mei (1970).

54Baxter and Sagart (2014).

55Sagart (1998) suggests that the departing tone arose in the period between Old and Middle Chinese (between 500 BCE and 500 CE) as a result of the loss of final laryngeals in words that contained the other three tones, which already existed at the time of Old Chinese.

56LaPolla (2017), Sagart (2001), Sagart (1999), Baxter and Sagart (1998), Schüssler (1976).

57Sagart (1999), p. 133–134; Baxter and Sagart (1998), p. 60; Zhengzhang (1995), p. 280.

58Haudricourt (1954).

59Downer (1959).

60Reconstructed forms taken from Baxter and Sagart (2014).

61LaPolla (2017), p. 29.

62Sagart et al. (2019) place the origin of proto-Sino-Tibetan at c5,200 BCE. See note 21.

←21 | 22→

63van Driem (2005) posits Tibeto-Burman—not Sino-Tibetan—as the parent family, with Chinese belonging to a sub-family termed ‘Sino-Bodic’ (van Driem, 1997) located two levels below the parent Tibeto-Burman family. Bodman (1980) offered a similar proposal. van Driem proposes that the Tibeto-Burman homeland was in western Sichuan some time between 11,500–2000 BCE.

64Mandarin technically fits the definition of Chinese ‘regionalect’ because it is mutually unintelligible with the other seven Chinese regionalects. If we do group Mandarin with the other regionalects then it must be considered a ‘first among equals’ because it is China’s lingua franca, spoken over the entire country and understood by virtually the entire population. For a more extensive and nuanced discussion of the term ‘Mandarin’, see Weng (2018).

65For example, Di and Sanchez-Mazas (2011), Hammer et al. (2006).


Baxter, W.Η. III. 1992. A Handbook of Old Chinese Phonology. Berlin-New York: Mouton de Gruyter.

Baxter, W., and Sagart, L. 1998. “Word Formation in Old Chinese.” In New Approaches to Chinese Word Formation, edited by Packard, Jerome L., 35–76. Berlin: Mouton de Gruyter.

Baxter, W., and Sagart, L. 2014. Old Chinese: A New Reconstruction. Oxford: UK. Oxford University Press.

Bodman, N. C. 1980. “Proto-Chinese and Sino-Tibetan: Data towards Establishing the Nature of the Relationship.” In Contributions to Historical Linguistics: Issues and Materials, edited by van Coetsem, Frans and Waugh, Linda R., 34–199. Leiden: E. J. Brill.

Cavalli-Sforza, L., and Feldman, M. 2003. “The Application of Molecular Genetic Approaches to the Study of Human Evolution.” Nature Genetics Supplement 33: 266–275.

Cheng Xiangqing. 1981. “Research in Pre-Qin Bisyllabic Words.” In Research in Pre-Qin Chinese Language, edited by Cheng, Xiangqing, 44–112. Jinan, China: Shan-dong Educational Publishing Co. 程湘清(1981). 先秦双音词研究。程湘清(1981). 先秦汉语研究。济南:山东教育出版公司.

Clarkson, C., Jacobs, Z., Marwick, B., Fullagar, R., Wallis, L., Smith, M., Roberts, R. G. et al. 2017. “Human Occupation of Northern Australia by 65,000 Years Ago.” Nature 547: 306–310.

Coblin, W.S. 2000. “A Brief History of Mandarin.” Journal of the American Oriental Society 120, no. 4: 537–552.

Crass, J. 2009. “Afroasiatic Languages.” In Concise Encyclopedia of Languages of the World, edited by Brown, K. and Ogilvey, S., 12–15. Oxford, UK: Elsevier.

Di, D., and Sanchez-Mazas, A. 2011. “Challenging Views on the Peopling History of East Asia: The Story According to HLA Markers.” American Journal of Physical Anthropology 145: 81–96.

Disotell, T.R. 1999. “Human Evolution: The Southern Route to Asia.” Current Biology 9: R925– R928.

←22 | 23→

Downer, G. 1959. “Derivation by Tone-Change in Classical Chinese.” Bulletin of the School of African and Oriental Studies 22: 258–290.

Driem, G. van 1997. “Sino-bodic.” Bulletin of the School of Oriental and African Studies 60: 455–488.

Driem, G. van 2005. “Tibeto-Burman vs Indo-Chinese: Implication for Population Geneticists, Archeologists and Prehistorians.” In The Peopling of East Asia: Putting Together Archaeology, Linguistics and Genetics, edited by Laurent Sagart, Roger Blench, and Alicia Sanchez-Mazas, 81–106. New York: Routledge Curzon.

Ehret, C. 2002. “Language Family Expansions: Broadening Our Understandings of Cause from an African Perspective.” In Examining the Farming/Language Dispersal Hypothesis, edited by Bellwood, P. and Renfrew, C., 163–176. Cambridge: McDonald Institute for Archaeological Research.

Everett, D.L. 2017. How Language Began: The Story of Humanity’s Greatest Invention. New York: Liveright Publishing.

Feng, S. 1998. “Prosodic Structure and Compound Words in Classical Chinese.” In New Approaches to Chinese Word Formation, edited by Packard, Jerome L., 197–260. Berlin: Mouton de Gruyter.

González, A., Larruga, J., Abu-Amero, K., Shi, Y., Pestano, J., and Cabrera, V. 2007. “Mitochon-drial Lineage M1 Traces an Early Human Backflow to Africa.” BMC Genomics 8: 223.

Hammer, M., Karafet, T., Park, H., Omoto, K., Harihara, S., Stoneking, M., and Horai, S. 2006. “Dual Origins of the Japanese: Common Ground for Hunter-Gatherer and Farmer Y Chromosomes.” Journal of Human Genetics 51, no. 1: 47–58.

Handel, Z. 2012. “Valence-Changing Prefixes and Voicing Alternation in Old Chinese and Proto-Sino-Tibetan: Reconstructing *s-and *N-Prefixes.” Language and Linguistics 13, no. 1: 61–82.

Haudricourt, A.-G. 1954. “Comment reconstruire le chinois archaïque.” Word 10, no. 2/3: 351–364. (“How to Reconstruct Old Chinese”).

Jin, L., and Su, B. 2000. “Natives or Immigrants: Modern Human Origin in East Asia.” Nature Reviews Genetics 1: 126–133.

LaPolla, R. J. 2017. “Overview of Sino-Tibetan Morphosyntax.” In The Sino-Tibetan Languages, Second Edition, edited by Graham Thurgood and Randy J. LaPolla, 22–42. London & New York: Routledge.

Li, J., Zeng, W., Zhang, Y., Ko, A., Li, M., Zhu, H., Fu, Q., and Zhou, H. 2017. “Ancient DNA Reveals Genetic Connections between Early Di-Qiang and Han Chinese and Hui.” BMC Evolutionary Biology 17: 239.

McCoy, J. 1980. “The Reconstruction of Upper Register Nasals and Laterals in Proto-Cantonese.” In Contributions to Historical Linguistics: Issues and Materials, edited by van Coetsem, Frans and Waugh, Linda R., 200–213. Leiden: E. J. Brill.

Mei, T.-L. 1970. “Tones and Prosody in Middle Chinese and the Origin of the Rising Tone.” Harvard Journal of Asian Studies 30: 86–110.

Mei, T.-L. 2012. “The Causative *s-and Nominalizing *-s in Old Chinese and Related Matters in Proto-Sino-Tibetan.” Language and Linguistics 13, no. 1: 1–28.

Norman, J. 1988. Chinese. Cambridge: Cambridge University Press.

←23 | 24→

Packard, J. L. ed. 1998. New Approaches to Chinese Word Formation: Morphology, Phonology and the Lexicon in Modern and Ancient Chinese. Trends in Linguistics Studies and Monographs Series #105. Berlin/New York: Mouton de Gruyter.

Packard, J. L. 2019. “The Proto-Chinese Urheimat.” Paper presented to the The 52nd Annual Sino-Tibetan Meeting. Sydney, Australia.

Peng, M.S., Palanichamy, M.G., Yao, Y.-G., Mitra, B., Cheng, Y.-T., Zhao, M., Liu, J. et al. 2011. “Inland Post-Glacial Dispersal in East Asia Revealed by Mitochondrial Haplogroup M9a’b.” BMC Biology 9, no. 2.

Peyraube, A. 2015. “Periodization.” In Encyclopedia of Chinese Language and Linguistics, edited by Sybesma, Rint. Leiden: E. J. Brill.

Pulleyblank, E.G. 2000. “Morphology in Old Chinese.” Journal of Chinese Linguistics 28, no. 1: 26–51.

Sagart, L. 1998. “The Origin of Chinese Tones.” Proceedings of the Symposium/Cross-Linguistic Studies of Tonal Phenomena/Tonogenesis, Typology and Related Topics. Tokyo, Japan. Institute for the Study of Languages and Cultures of Asia and Africa, Tokyo University of Foreign Studies, pp. 91–104.

Sagart, L. 1999. The Roots of Old Chinese. Philadelphia: John Benjamins.

Sagart, L. 2001. “Vestiges of Archaic Chinese Derivational Affixes in Modern Chinese Dialects.” In Sinitic Grammar: Synchronic and Diachronic Perspectives, edited by Chappell, 123–142. Oxford, UK: Oxford University Press.

Sagart, L., and Baxter, W.H. 2012. “Reconstructing the *s-Prefix in Old Chinese.” Language and Linguistics 13, no. 1: 29–59.

Sagart, L., Jacques, G., Lai, Y., Ryder, R., Thouzeauc, V., Greenhill, S., and List, J. 2019. Dated Language Phylogenies Shed Light on the Ancestry of Sino-Tibetan.” Proceedings of the National Academy of Sciences of the United States of America 116, no. 21: 10317–10322. doi:10.1073/pnas.1817972116. PMC 6534992. PMID 31061123.

Sanchez-Mazas, A., Di, D., and Riccio, M.E. 2011. “A Genetic Focus on the Peopling History of East Asia: Critical Views.” Rice 4: 159–169.

Schüssler, A. 1976. Affixes in Proto-Chinese. Wiesbaden: Verlag.

Shi, H., Zhong, H., Peng, Y., Dong, Y., Qi, X., Zhang, F., Liu, L. et al. 2008. “Y Chromosome Evidence of Earliest Modern Human Settlement in East Asia and Multiple Origins of Tibetan and Japanese Populations.” BMC Biology 6: 45. doi:10.1186/1741-7007-6-45.

Shi, H., Qi, X., Zhong, H., Peng, Y., Zhang, X., Ma, R.Z., and Su, B. 2013. “Genetic Evidence of an East Asian Origin and Paleolithic Northward Migration of Y-Chromosome Haplogroup N.” PLoS ONE 8, no. 6: e66102. doi:10.1371/journal.pone.0066102.

Soares, P., Alshamali, F., Pereira, J., Fernandes, V., Silva, N., Afonso, C., Costa, M. et al. 2012. “The Expansion of mtDNA Haplogroup L3 within and Out of Africa.” Molecular Biology and Evolution 29, no. 3: 915–927.

Su, B., Xiao, J., Underhill, P., Deka, R., Zhang, W., Akey, J., Huang, W. et al. 1999. “Y-Chromosome Evidence for a Northward Migration of Modern Humans into Eastern Asia during the Last Ice Age.” American Journal of Human Genetics 65: 1718–1724.

Sybesma, R. ed. 2015. Encyclopedia of Chinese Language and Linguistics. Leiden: E. J. Brill.

←24 |

Ting, P.-h. 1979. “The Syllable Structure of Ancient Chinese.” Bulletin of the Institute of History and Philology, Academia Sinica 50, no. 4: 717–739.

Wang, C.-C, Wang, L.-X, Shrestha, R., Zhang, M., Huang, X.-Y, Hu, K., Jin, L., and Li, H. 2014. “Genetic Structure of Qiangic Populations Residing in the Western Sichuan Corridor.” PLoS ONE 9, no. 8: e103772. doi:10.1371/journal.pone.0103772

Wang, L., Lu, Y., Zhang, C., Wei, L., Yan, S., Huang, Y., Wang, C. et al. 2018. “Reconstruction of Y-Chromosome Phylogeny Reveals Two Neolithic Expansions of Tibeto-Burman Populations.” Molecular Genetics and Genomics 293: 1293–1300.

Wang, W.S.-Y. 2015. “The Peoples and Languages of China.” In Oxford Handbook of Chinese Linguistics, edited by Wang, W.S-Y. and Sun, C., 19–33. Oxford, UK: Oxford University Press.

Weng, J. 2018. “What Is Mandarin? The Social Project of Language Standardization in Early Republican China.” The Journal of Asian Studies 77, no. 3: 611–633.

Xu, T. 1991. Historical Linguistics. Beijing: Commercial Press. 徐通锵(1991). 历史语言学. 北京:商务印书馆.

Zhao, M., Kong, Q.P., Wang, H.W., Peng, M.S., Xie, X.D., Wang, W.Z., Jiayang et al. 2009. “Mitochondrial Genome Evidence Reveals Successful Late Paleolithic Settlement on the Tibetan Plateau.” Proceedings of the National Academy of Sciences of the United States of America 106: 21230–21235.

Zhengzhang, S. 1995. 郑张尚芳. 汉语与亲属语同源根词及附缀成分比较上的择对问题。 (The Problem of Correctly Choosing Cognate Roots and Affixes in Chinese and Related Languages). In The Ancestry of the Chinese Language. Journal of Chinese Linguistics Monograph Series #8, edited by Wang, W.S.-Y., 269–282. Berkeley, CA: Project on Linguistic Analysis.

Zhong, H., Shi, H., Qi, X.B., Xiao, C.J., Jin, L., Ma, R.Z., and Su, B. 2010. “Global Distribution of Ychromosome Haplogroup C Reveals the Prehistoric Migration Routes of African Exodus and Early Settlement in East Asia.” Journal of Human Genetics 55: 428–435.

Zhong, H., Shi, H., Qi, X.B., Duan, Z.Y., Tan, P.P., Jin, L., Su, B., and Ma, R.Z. 2011. “Extended Y Chromosome Investigation Suggests Postglacial Migrations of Modern Humans into East Asia via the Northern Route.” Molecular Biology and Evolution 28: 717–727.

Zhu, X. 2015. “Tonogenesis.” In Encyclopedia of Chinese Language and Linguistics, edited by Sybesma, Rint. Leiden: E. J. Brill.

←25 |


Chinese Regionalects

Now that we have seen that Proto-Chinese is likely to have been spoken along the Yellow River around 4,000 BCE, the next question is how the regionalects came to be spoken in their present locations, and why they have the linguistic form that they have—a form that is different from, yet clearly related to, the Proto-Chinese that is the ancestor language of all present-day Chinese regionalects.

Chinese is generally considered to have eight regionalects (Mandarin, Wu, Min, Yue, Gan, Xiang, Jin and Kejia), with all of them (except for Mandarin and Jin) located in southeastern China. These regionalects may be divided into three groups1: northern (Mandarin and Jin), southern (Hakka, Yue and Min, that is, those located on the southern bank of the Yangtze (Changjiang 长江) river and along its southern tributaries) and central (Wu, Gan and Xiang).

The central and southern regionalects came to be spoken in their present locations in southeast China following migrations from the north by groups who were originally speakers of Proto-Chinese. These groups are believed to have diverged from Old Chinese in the 1st millennium CE,2 migrating to the southeast along three primary routes: (a) an eastern route by sea and along the narrow coastal plain, (b) a central route along the Gan 赣 River watershed, and (c) a western route along the Xiang 湘 River watershed.3

←27 | 28→

There are three factors that explain why those central and southern regionalect groups in the southeast of China have such linguistic diversity. The first is due to the topography of southeast China, the second is the fact that the immigrants into southeast China came from different places along different migration routes and at different times, and the third is that different substrate languages were spoken by the original inhabitants of the eventual regionalect areas when the immigrants arrived.

Regarding the first factor, the topography of southeast China consists of separate linguistic areas that are isolated from each other by topological features such as mountains and rivers. When migrants move into those areas, even if they were to start out with exactly the same language, over the passage of time that same language would eventually turn into different languages because the groups that speak them are isolated from each other. This process of linguistic change is like Darwinian evolution, in which species that start out the same become different over time because they occupy different isolated ecological niches.

The second factor is that groups that migrated into the regionalect areas came from different places and at different times, and so the languages they brought with them were different as a result. When different incoming languages are imposed upon a receiving, substrate language, the resulting languages are also bound to be different, based on differences in the incoming languages.

As for the third factor, there were already people living in the separate isolated areas that would eventually become the southeast regionalect areas, and the languages that they spoke while living there were undoubtedly already different from one another. This means that each area probably had a different substrate language, so the languages that arose from the mixing of the incoming and substrate languages would naturally be different.

So, what different substrate languages were spoken by the people already living in the different regionalect areas? There are a number of possibilities. They may have been descendants of the languages spoken by the northern branch of the Out of Africa II Southern Coastal group4 in the vicinity of the Ancient Tea Horse Road, in which case the substrate languages and Proto-Chinese/Proto-Sino-Tibetan may have had a common ancestor language. The substrate languages could also have been languages related to Austro-Asiatic (such as Vietnamese, Khmer and Mon) or Austronesian (such as Indonesian, Tagalog and Paiwan), languages that would come to occupy neighboring areas of Southeast Asia. For all we know, the different substrate languages might even have been related to languages spoken by earlier homo populations that left Africa during Out of Africa I, that is, prior to the more recent 80,000–90,000 BCE L3 haplogroup emigration.

←28 | 29→

There has been no shortage of proposals regarding which immigrant and substrate languages were originally spoken in what would become the southeast regionalect areas. For example, it has been proposed that the Wu regionalect came about when immigrants imposed their language onto an Austroasiatic substrate.5 It has also been suggested that the Xiang regionalect resulted when immigrants moved into an area where Tai/Hmong-Mien were the substrate languages.6 Some have claimed that the Gan and Hakka regionalects came about when the Wu regionalect and a Xiang ancestor language split into the Gan and Hakka regionalects, when the Hakka speakers emigrated east, south and west.7 Finally, it has been proposed that the Yue were originally an Austroasiatic-speaking people,8 and that Hakka, Min and Yue share a common ancestor language which has an Austroasiatic substratum.9

It is likely that when Proto-Chinese speakers made their way to southeast China, the substrate languages that were already spoken there influenced the form of the regionalects that are spoken there now. So let us now take a closer look at those Chinese regionalects. Because Mandarin is considered China’s lingua franca, my tack will be to give Mandarin a relatively detailed description, and for the other regionalects I will indicate their most noteworthy characteristics especially where they differ from Mandarin.

The Regionalects

The presence of the regionalects in China is one of the most commonly-recognized characteristics of the Chinese language (for example, note the commonly used phrase ‘the Chinese dialects’). The fact that the Chinese regionalects resemble one another and share common linguistic features is one of the main reasons10 they are assumed to derive from the parent Proto-Chinese language. The relationship among the Chinese regionalects is roughly the same as the relation among the Romance languages such as Italian, French, Spanish and Portuguese. The Romance languages all derive from Latin as the parent language and are noticeably similar to each other, and yet they are still mutually unintelligible to the extent that, for example, a speaker of French will not understand what a Spanish speaker says, unless the French speaker also knows some Spanish. The situation is the same for the Chinese regionalects. A speaker of the Wu regionalect, for example, will have trouble understanding what a Yue (Cantonese) speaker says, unless that speaker also has knowledge of Yue. So the comparison to the Romance language situation is a good intuitive rule of thumb in thinking about the relationship among the

←29 | 30→

Table 3.1. Numbers of Chinese regionalect speakers. Source: Author. Data from Gu (2006).

a Shanghainese is a dialect of Wu.

b Teochew, Hokkien and Taiwanese are dialects of Min.

c Toisanese is a dialect of Cantonese, which is a dialect of Yue.

d Hakka is the Kejia pronunciation of ‘Kejia’.

Chinese regionalects. We will take a closer look at the mutual intelligibility of the Chinese regionalects later in this chapter.

The numbers of Chinese regionalect speakers may be seen in Table 3.1. As is clear from the table, the number of Mandarin speakers exceeds the next closest regionalect (Wu) by nearly a factor of ten. One reason for this is that Mandarin is the official language of China, functioning as China’s designated language for government and diplomatic functions, for the mass media, for the education system, and all other official purposes. Mandarin is understood in virtually all of China, even in the southeast regionalect areas, making it truly China’s lingua franca. Let us now take a closer look at Mandarin.

Mandarin (Putonghua, Guoyu, Huayu)

←30 | 31→

The term Mandarin has several different contemporary meanings11—it can mean Imperial Mandarin—the spoken language that imperial officials used to communicate with each other during the Ming (1368–1644) and Qing (1644–1911) dynasties; Standard Mandarin—the idealized national standard language of China; Geographical Mandarin—the family of mutually intelligible dialects that range from China’s northeast to southwest; and Local Mandarin—the individual instances of Geographical Mandarin spoken at the local level. In this section we use the term to refer mainly to Standard Mandarin, because it mostly refers to the idealized standard language, but at times it also touches on Geographical and Local Mandarin, for example when it makes reference to variations in Mandarin as spoken at the local level.

Mandarin (Standard) is translated into Chinese differently depending on national location: it is known as Putonghua (普通话 ‘common language’) in China, Guoyu (国语 ‘national language’) in Taiwan and Huayu (华语 ‘Chinese language’) in Singapore. Even though Mandarin is not limited to a narrow geographical region within China, it technically fits our definition of ‘regionalect’ rather than ‘dialect’ because it is not mutually intelligible with the other seven regionalects. But even if we consider Mandarin to be a ‘regionalect’ like the others, it must in a sense be considered a ‘first among equals’ because of its status as China’s official standard language and its lingua franca, and because it is used so widely throughout the entire country. The Mandarin regionalect itself has several dialects (see Table 3.5), which, although mutually intelligible, do vary considerably in phonology and vocabulary.

Sound System

As seen in Table 3.2, Standard Mandarin (based on the Beijing dialect of the Mandarin regionalect) has 21 initial consonants: a labial series [p] [pʰ] [m] [f ], an apico-dental series [t] [tʰ] [n] [l], an apico-alveolar series [t͡s] [t͡sʰ] [s], an apicoretroflex series [ʈ͡ʂ] [ʈ͡ʂʰ] [ʂ] [ɻ], a palatal series [t͡ɕ] [t͡ɕʰ] [ɕ] and a velar series [k] [kʰ] [x]. Mandarin also has a velar-nasal [ŋ] that does not occur as an initial consonant, but that—along with [n]—can occur in syllable-final position.

Mandarin is the only regionalect12 that has the retroflex consonant series [ʈ͡ʂ] [ʈ͡ʂʰ] [ʂ] and [ɻ] (pinyin <zh> <ch> <sh> and <r>13 respectively). The retroflex consonants are used in the Mandarin dialects in north and central China (see Table 3.5), roughly north of a line that would separate the watersheds of the

Table 3.2. Mandarin initial consonants, IPA and pinyin. Source: Author.

←31 | 32→

Yellow and Changjiang (Yangtze) rivers. South of that line, the retroflex consonants [ʈ͡ʂ] [ʈ͡ʂʰ] [ʂ] [ɻ] are produced instead as the apical consonants [t͡s] [t͡sʰ] [s] and [z] (pinyin <z> <c> <s> and ø14 respectively). As an example, the Mandarin word for ‘book’ is pronounced shu (IPA [ʂu], sounds like English ‘shoe’) in the Beijing dialect of Mandarin but is pronounced su (IPA [su], sounds like English ‘sue’) in the, for example, Chengdu dialect of Mandarin and in most other southwestern dialects of Mandarin.

Mandarin is also the only regionalect that has the ‘er-hua’ 儿化 (rhotaciza-tion) phenomenon, a phonetic process in which an r-sound (pinyin <er>, IPA [əɻ], consisting of the nuclear vowel [ə] plus the [ɻ] consonant as a coda) attaches to and usually partially supplants part of the syllable final. For example, the word ‘play’ in Mandarin is wan (玩; IPA [Uan]), but when spoken in the Beijing regionalect, following ‘er-hua’, is pronounced war (IPA [Uaəɻ]). ‘er-hua’ often functions as a diminutive suffix for nouns, but—especially in the Beijing dialect—the er often attaches merely as a stylistic phonetic element with no meaning change or grammatical function. The geographic distribution of ‘er-hua’ is generally the Mandarin dialects north of the Changjiang (Yangtze) river15 (see Table 3.5).

Mandarin is also distinguished by being the only regionalect that, in general, did not retain the historical ‘entering tone’, that is, the tone that includes pronunciation of the syllable-final consonants [p] [t] [k] and [ɂ] (glottal stop). With the exception of a handful of Mandarin dialects (see Table 3.5), the only consonants that Mandarin allows at the end of a syllable are [n], [ŋ] and in some cases [ɻ].

As seen in Table 3.3, Mandarin phonetically has ten nuclear vowels—the eight vowels [i] [y] [e] [a] [u] [o] [ə] and [ɤ], plus the retroflex apical vowel [ʅ] (that occurs only after the retroflex consonants [ʈ͡ʂ] [ʈ͡ʂʰ] [ʂ] and [ɻ]) and the alveolar apical vowel [ɿ] (that occurs only after the apical consonants [t͡s] [t͡sʰ] and [s]).

Chinese is a lexical tone language, which means that the meaning of a word is determined by the voice pitch used when the word is pronounced. Standard Mandarin has four tones, as do most of the Mandarin dialects, though some Mandarin dialects can have as few as three and some as many as seven tones (see Table 3.5). The four standard Mandarin tones are: first tone—high level, second

Table 3.3. Mandarin vowels. Source: Author

←32 | 33→

Table 3.4. The syllable ‘mi’ pronounced with each of the four tones. Source: Author.

tone—mid-to-high rising, third tone—mid-to-low-to-mid fall-rise, and fourth tone—high-to-low falling. Using numerical values to describe the tones—in which 5 represents a ‘high’ voice pitch and 1 represents a ‘low’ voice pitch—first tone is a 5-5, second tone is a 2-4, third tone is a 2-1-4, and fourth tone is 5-1.

To give examples using each of the four tones on the same syllable (see Table 3.4), the Mandarin syllable mi means ‘squint’ when pronounced with the high-level first tone mi1, ‘confused’ when pronounced with the rising second tone mi2, ‘rice’ when pronounced with the fall-rise third tone mi3 and ‘honey’ when pronounced with the falling fourth tone mi4. In more intuitive terms, first tone is produced as if it were a note that is sung and held at a high steady pitch like the note ‘mi’ on the do-re-mi scale. Second tone is produced like a single-word rising intonation question in English like who? or, me? Third tone is produced like an incredulous, exaggerated one-word question in English, like me..e..e? Fourth tone sounds like a one-word command in English, like stop! or, me! me! Table 3.5 summarizes the occurrence of retroflex consonants, the final er, the number of tones and the occurrence of ‘entering tone’ in a selection of Mandarin dialects.16


There are some minor differences among the regionalects, but by-and-large, they all have the same basic syllable structure. As seen in Table 3.6, the Mandarin syllable may contain four possible sounds plus a tone, CGVXT. The ‘C’ is an optional initial consonant (any of the consonants except [ŋ]), the ‘G’ is an optional medial glide vowel ([i] , [y] or [u]), the ‘V’ is an obligatory nuclear vowel (any of the vowels), and the ‘X’ is a coda that may be either a high vowel ([i] or [u]), an [ɻ], an [n], or an [ŋ], and the ‘T’ is one of the four Mandarin tones.

There are several additional restrictions that limit the phonetic shape of the Mandarin syllable, as seen in Table 3.7. Perhaps the most noteworthy is that except for [n] and [ɻ], no sound may occur twice in one syllable. Another restriction is that the palatal consonant series [t͡ɕ] [t͡ɕʰ] and [ɕ] may only be followed by the high front vowels [i] or [y], while the retroflex consonant series [ʈ͡ʂ] [ʈ͡ʂʰ] [ʂ] [ɻ] may never be followed by either of those two vowels.17 As a final example, the retroflex apical vowel [ʅ] and the alveolar apical vowel [ɿ] may only occur after the retroflex consonant series and the apical consonant series respectively, and they may not be followed by any other vowel or consonant.

←33 | 34→

←34 | 35→

Table 3.6. Mandarin syllable structure. Source: Author.

Chinese syllables are traditionally divided into initial and final segments, and the final is divided into a glide and a rhyme. In the CGVX template seen in Tables 3.6 and 3.7, the ‘C’ represents the initial, the ‘GVX’ represents the final and the VX represents the rhyme. The initial can be any of the 21 initial consonants. The ‘GVX’ final is comprised of the eight vowels in different arrangements, giving us 38 finals in total. Multiplying the 21 possible C initials by the 38 possible GVX finals yields a theoretically predicted number of 798 total syllables. The actual number of Mandarin syllables is 408. Multiplying the 408 syllables by the four tones yields 1632 predicted syllables, but the number of Mandarin syllables that actually occur is approximately 1200.

Table 3.7. Constraints on the Mandarin syllable. Source: Author.

←35 | 36→

Chinese Morphemes

In all the languages of the world, words are constructed out of morphemes, which are the smallest pieces of language that can be isolated as having a single, indivisible unit of meaning. Every language uses morphemes as the building blocks of words. To give an example in English, the word boy has one morpheme that means ‘young male’, and the word boys has two morphemes, one is boy ‘young male’ and one is – s, the plural morpheme. For a more complex English example, the word nonreusability (‘the property of not being able to be reused’) contains five morphemes: non-re-use -able -ity. The derivation of that word would be: use (verb) > reuse (verb) > reusable (adjective) > nonreusable (adjective) > nonreusability (noun).

The Chinese language has three properties that give it a rather special relationship with the concept morpheme. First, in Chinese virtually every morpheme consists of one syllable, and just about every syllable is a single morpheme.18 We can call this the ‘monosyllabic morpheme’ property. Contrast this property with the case of, for example, an English morpheme like ‘city’ which is a two-syllable morpheme in addition to being a two-syllable word. It is a two-syllable morpheme because there is no way to divide ‘city’ into smaller pieces of meaning, like one syllable that means one part of ‘city’ and one syllable that means another part of ‘city.’

Second, every Chinese morpheme contains only one unit of meaning and cannot ‘share’ two morphemes in one single form, as is common in other languages. Let us call this the ‘single morpheme’ property.19 A good example in English of two morphemes shared in one form is the word went, which contains two morphemes—one for the meaning go and one for the meaning past tense, with the single form went including both of those two meanings. There is no independent phonological identity for the part of went that encodes the meaning go and a part that encodes the meaning past tense. Likewise for the English pronoun him, which shares three morphemes in one form: gender (male), person (third) and for case (accusative, or direct object). For an example in Latin, in the phrase cave canem meaning ‘beware of the dog’, the word for ‘dog’ canem is in the accusative singular case, and so the word may be divided into three morphemes, the morpheme (root) can-meaning ‘dog’ and the ending –em which means both accusative case (direct object) and number (singular). Although the ending –em contains the two meanings ‘singular’ and ‘accusative’, there is no way to divide the –em ending into a part that contains ‘singular’ and a part that contains ‘accusative’.

←36 | 37→

Modern Chinese has no examples of this sort: there are no morphemes in Chinese that participate in such ‘double duty’ by encoding more than one meaning in a single phonetic form. As an example (virtually every Chinese morpheme serves as an example!), the Mandarin morpheme (word) gou 狗 meaning ‘dog’ has the same form irrespective of number, gender, person or case.

These two properties of Chinese morphemes—the ‘monosyllabic morpheme’ property and the ‘single morpheme’ property—are why Chinese is considered an isolating language par exellence in language typology. Chinese is a prototypical example of an isolating language because its morphemes are very easy to isolate, identify and associate with a single phonetic form.

The third remarkable property of Chinese morphemes involves the Chinese writing system, and that is—with few exceptions—that each Chinese character is equivalent to one single morpheme,20 which we may term the ‘one-character morpheme’ property. When you consider these three properties of Chinese morphemes—monosyllabic morpheme, single morpheme, and one-character morpheme—we are struck with the remarkable fact that in Chinese, the morpheme, the syllable and the written character are almost completely isomorphic, that is, the three concepts virtually completely overlap: every morpheme is a syllable, every syllable is a character, and every character is a morpheme.

Now that we have seen how readily Chinese lends itself to division into morphemes, let us consider two more properties of Chinese morphemes that are relevant to how Chinese forms words. The two properties are (1) whether the meaning of the morpheme expresses content or function information, and (2) whether the morpheme is free or bound.

A morpheme is a content morpheme if it represents objects or actions. For our purposes, that would be if the morpheme has a lexical meaning involving a noun, verb, adjective or adverb. In contrast, a morpheme is a function morpheme if it does not have a specific lexical meaning, but rather expresses a grammatical relationship between other parts of an utterance. In linguistics, this content/function property may be represented on a continuum using the concept of the cline,21 which signifies how ‘empty’ of meaning a morpheme is. The content/function cline represents the fact that the content/function property is not a discrete, all or nothing value, but rather varies along a continuum according to whether the meaning of a morpheme is more, or less, empty.

On the ‘content’ (‘left’) side of the cline continuum we have nouns, verbs, adjectives and adverbs, with adverbs being to the right of nouns and verbs on the continuum because adverbs do not carry their own lexical reference to action, but rather ‘borrow’ that ‘action’ information from a verb elsewhere in the discourse.

←37 | 38→

Table 3.8. Four Morpheme types in Chinese. Source: Author.

On the ‘function’ side of the cline continuum, we have copulas, classifiers, articles and pronouns. Copulas are function words because their function is simply to set up an equivalence between two nominal elements, to which they contribute no lexical meaning.

Function words are traditionally considered closed class items, because there is an upper limit to the number of function words there can be in a language, while for content words there is, in principle, no upper boundary on the number of content words that a language can have. Pronouns would fall a little to the left of copulas on the continuum, because while pronouns are nouns, nonetheless they carry no lexical meaning of their own but rather receive their lexical content from nouns elsewhere in the discourse. It is worth pointing out that the distinction between content and function has a long history in Chinese language and literature, embodied in the distinction between words (ci 词) that are either ‘empty’ (xuci 虚词 empty-word ‘function word’) or ‘real’ (shici 实词 real-word ‘content word’).

Following the function/content property, the second property of Chinese morphemes that will be relevant in our discussion of word formation is whether the morpheme is free or bound. A morpheme is free if it can independently occupy a syntactic form class (i.e., ‘part of speech’) slot, and it is bound if it can only occupy a syntactic slot if it is augmented with additional morphemic information. It is important to point out that when a morpheme is free it is always considered a word.

To give an example from English demonstrating the free/bound morpheme distinction, the morpheme micro-from the word microchip is bound, because micro-cannot occupy a noun slot by itself—it needs to be augmented with other morphemic information like chip or computer before it can occupy a noun slot. A good example of a bound morpheme in Chinese is fang-房 meaning ‘room’ or ‘house’. The morpheme fang-is not a word because it cannot stand by itself as a

←38 | 39→

Table 3.9. Chinese text sample—no word parsing information. Source: Author.

a Modified from the ‘dúbái’ (独白 alone-white ‘monologue’) entry in the Baidu Online

Encyclopedia. Translation (JLP): “People are the sum of their social networks, with both their words and deeds constrained by their web of social relations. Under certain social conditions it is difficult for people to face each other with sincerity. Many behaviors are carried out only as a last resort and many statements are made that violate speaker intent. But when a person has only himself to face, the situation is quite different—he can reveal the complete truth and make a clean breast of everything.”

noun—it must be augmented with additional morphemic information. So fangzi 房子room-NOM is fang-followed by the nominalizing suffix – zi 子to form the word meaning ‘house’, chufang 厨房 kitchen-room is the word for ‘kitchen’ and fangjian 房间room-room means ‘room.’

Having identified the two properties function/content and free/bound, it turns out that every morpheme in Chinese can be identified by its combination of function/content and free/bound values as seen in Table 3.8: if a morpheme is free and content it is a (content) word; if a morpheme is free and function it is a (function) word; if a morpheme is bound and content it is a bound root and if a morpheme is bound and function it is an affix.

There is a subset of affixes called grammatical affixes, consisting of the suffixes – le (了 aspect), – guo (过 aspect), – zhe (着 aspect), – men (们 plural), the resultative affix – de– /– bu– (得/不 aspect) and the adverbial suffix – de (得 aspect) that introduces a complement of extent (e.g., haode shei dou xianmu ta 好得谁都羡慕她 good-asp who all admire-admire her ‘so good that everyone admires her’).22


XXIV, 206
ISBN (Book)
Publication date
2021 (May)
New York, Bern, Berlin, Bruxelles, Oxford, Wien, 2021. XXIV, 206 pp., 25 b/w ill., 3 color ill., 71 tables.

Biographical notes

Jerome L. Packard (Author)

Jerome L. Packard is Professor Emeritus of Chinese at the University of Illinois. Professor Packard received his PhD in Chinese linguistics from Cornell University and is the author of five books and over 30 articles on the Chinese language. Professor Packard has served as a Fulbright scholar in China and has taught Chinese at the University of Massachusetts, Cornell University, the University of Pennsylvania and the University of Illinois.


Title: A Social View on the Chinese Language