The earliest historical linguistic evidence of the spoken Chinese language dates back approximately 4500 years, while examples of the writing system that would become written Chinese are attested in inscriptions made on bronze vessels and oracle bones during the Late Shang period (1050 BCE), with the very oldest dated to .
Proto-Sino-Tibetan and Proto-Sinitic (before 1300 BCE)
The oldest attested written Chinese—comprising the oracle bone inscriptions made during the 13th century BCE by the Shang dynasty royal house in modern Anyang, Henan—is also the earliest direct evidence of the Sinitic languages. Most experts agree that Sinitic languages share a common ancestor with the Tibeto-Burman languages, forming the primary Sino-Tibetan family. However, the precise placement of Sinitic within Sino-Tibetan is a matter of debate. Reconstructing the common ancestor of all Sino-Tibetan languages, known as Proto-Sino-Tibetan—a language last spoken thousands of years before the historical record—has been challenging. Even though Chinese characters are logographs that do not encode the sounds of speech directly, written Chinese allows for the partial reconstruction of how ancient forms of the language sounded. No such evidence exists for earlier stages, including where Sinitic would have split from other branches of Sino-Tibetan. There are several reasons that the comparative method conventionally used in historical linguistics has had limited success in reconstructing the history of Sino-Tibetan compared to the results achieved with other major language families like Indo-European: Relatively few Sino-Tibetan languages, of which Chinese is one, have written traditions old enough to provide useful comparative evidence. In areas where Sino-Tibetan languages are spoken, there have historically been high rates of population transfer. This has resulted in many areal features shared between languages, making it difficult to determine which relationships—not only cognates in vocabulary, but also grammatical features—are phylogenetic. Languages that make heavier use of morphology generally provide more data useful in historical comparisons, i.e. the conjugated and derived forms of words. Compared to Indo-European languages like Latin, Greek, and Sanskrit, Sino-Tibetan languages are highly isolating and exhibit very little morphology, making it difficult to identify cognates.
Pre-Classical period (1300–500 BCE) Old Chinese, sometimes known as "Archaic Chinese", is ancestral to all current Chinese languages. The first known use of the Chinese writing system is divinatory inscriptions into tortoise shells and oracle bones during the Shang dynasty (1766–1122 BCE). During the first half of the Zhou dynasty (1122–256 BCE), writing descended from the Shang is found in texts including inscriptions on bronze artefacts, the Classic of Poetry, the history of the Book of Documents, and portions of the I Ching. Phonetic elements found in the majority of Chinese characters also provide hints to their Old Chinese pronunciations, as do the pronunciations of borrowed Chinese characters in Japanese, Korean, and Vietnamese. Old Chinese was not wholly uninflected. It possessed a rich sound system in which aspiration or rough breathing differentiated the consonants, but was probably still without tones. Work on reconstructing Old Chinese started with Qing dynasty philologists.
Classical period (500 BCE – 1 CE)
Words in Old Chinese were generally monosyllabic; as such, each character denoted an independent word. Affixes could be added to form a new word, which was often written with the same single character. In many cases, the pronunciations then diverged due to the systematic sound changes caused by the affixes. For example, many modern readings reflect the departing tone present in Middle Chinese, which many scholars now believe is a reflex of a derivational suffix that served a range of semantic functions in Old Chinese. This is possibly the only example of inflectional morphology extant in what was otherwise an analytic language: + Character Gloss → → 'to transmit' → → 'a record' → → 'to grind' → → 'grindstone' → → 'to stay overnight' → → 'celestial mansion' → → 'speak' → → 'exhort' Another common sound change occurred between voiced and voiceless initials, though the phonemic voicing distinction has disappeared in most modern varieties. This is believed to reflect an Old Chinese de-transitivising prefix, but scholars disagree on whether the voiced or voiceless form reflects the original root. Each pair of examples below reflects two words of opposite transitivity. + Character Gloss → → 'to see' → → 'to appear' → → 'to defeat' → → 'to be defeated' → → 'to bend' → → 'to be broken by bending'
A divergent dialect split from Old Chinese in this period: Ba–Shu Chinese was first described during the Western Han dynasty and represented one of the earliest splits from Old Chinese.
Eastern Han period (1–300 CE) Eastern Han Chinese is the period which is considered to be intermediate between Old and Middle Chinese. Modern Min Chinese dialects are believed to have descended from southeastern dialects of Eastern Han Chinese which split off during this period. Proto-Min may have been a daughter language to an early form of Sinitic in the Lower Yangtze region known as Ancient Wu, though this is not fully accepted. Pointing to features of Min varieties that are also found in Hakka and Yue varieties, Jerry Norman suggests that all three groups are, to varying degrees, derived from a variety spoken in the lower Yangtze region during the Han period, which he calls Old Southern Chinese. He argues that this dialect belonged to the group of dialects known as Wu or Jiangdong () in the Western Jin period, when the writer Guo Pu (early 4th century AD) described them as quite distinct from other Chinese varieties.
The consonant clusters postulated for Old Chinese had generally disappeared by the Eastern Han period.
The customary writing style of the period was strongly modelled on the classics, and thus provides only occasional glimpses of contemporary grammar. Some works contain passages in a more colloquial style thought to reflect contemporary speech. Many such examples are found in translated Buddhist literature, particularly direct speech, or paraphrases and commentaries of the classics written for the benefit of novice students. In comparison with Warring States texts, colloquial Eastern Han texts display a massive increase in compound content words. Monosyllabic words were largely replaced by disyllabic compounds.
Old Chinese had a range of personal pronouns, including case distinctions. In the Eastern Han, these were reduced to first person and second person . The demonstrative also came to be used as a copular verb in sentences of the form A B (as in modern Chinese), replacing the typical classical pattern A B (). They also make much less use of function words in favour of periphrasis.
Middle Chinese (300–1200)
Middle Chinese was a form of Chinese used during the Sui, Tang, and Song dynasties between the 4th and 10th centuries. It can be divided into two periods: Early Middle Chinese is documented in the Qieyun (601), the first rime dictionary, and a later revision in the Guangyun (1008). Late Middle Chinese is reflected by 12th century rime tables such as the Yunjing. The evidence for the pronunciation of Middle Chinese comes from several sources: modern dialect variations, rime dictionaries, foreign transliterations, rime tables constructed by ancient Chinese philologists to summarize the phonetic system, and Chinese phonetic translations of foreign words.
Middle Chinese tones developed from Old Chinese syllables with final consonants which disappeared, with "departing" tone developing from syllables that ended with , "rising" tone developing from syllables that ended in , the "entering" tone being syllables who retained a final stop consonant, and "level" tone from syllables that ended in a vowel. As these suffixes were part of the derivational morphology of Old Chinese, tonal variation often appeared either in a single word or in semantically related words.
The Tang dynasty form of Middle Chinese was highly influential, forming much of the reading pronunciations in Korea, Japan, and Vietnam, and displacing pre-Tang regional varieties across all of its territories except for individual words and colloquial Min Chinese. However, Norman and Coblin (1995) questioned the assumption that all non-Min Chinese varieties descended from Middle Chinese as indicated in the rime dictionaries, demonstrating that spoken language in the Middle Chinese period was much more diverse and less phonologically conservative than the rime dictionaries indicate. They questioned whether the non-Min varieties had split only during the period of Middle Chinese (7th century or later).
Wu Chinese can be traced back to the Eastern Jin dynasty, when commonfolk typically spoke the earlier local form of Sinitic, Ancient Wu, or their native Shandong or northern Jiangsu Chinese, and the aristocracy, typically spoke a variety not dissimilar to that of early medieval Luoyang. This linguistic overlap eventually led to the formation of modern Wu, with many early coincidental strata that are hard to differentiate today. Shandong and northern Jiangsu dialects likely played an important role in the formation of Wu Chinese, as more than half of the migrants came from this region. Norman also argues that today's Hakka and Yue resulted from overlays of "Old Southern Chinese" by successive waves of influence from northern China. According to Sagart, Gan Chinese has a thick Late Middle Chinese layer brought by Tang dynasty northern migrants, combined with an archaic layer derived from the same source as Min, Hakka and Yue, although less prominent due to its greater exposure to northern influence.Sagart, Laurent. 2002. Gan, Hakka and the Formation of Chinese Dialects. Dialect Variations in Chinese, 129–153. Papers from the Third International Conference on Sinology, Linguistics Section. On top of the Tang-era influences, Gan Chinese absorbed continued northern influences due to southward flows of refugees in the Southern Song.
Old Mandarin (1100–1400)
After the Middle Chinese period, non-Han rule in Northern China weakened the conservative tradition of the Qieyun. Local speech patterns in the North developed a new form of spoken language which scholars now refer to as Old Mandarin.
Shao Yong's adaptation of the rime tables, without reference to the Qieyun tradition, to describe the phonology of either 11th-century Kaifeng, or his hometown Fanyang, show that the speech of this time already displayed some of the phonological characteristics of Northern Mandarin, such as the diphthongization of vowels in syllables with -k finals. Zhou Zumo (1966) concluded from Shao Yong's tables that his voiced stops and affricates had also devoiced and aspirated in level tone and devoiced without aspiration in other tones, a change which was not visible in contemporary Western Xia records. Similar records of Chinese proper nouns in the Khitan small script also show that diphthongization had occurred, as well as other Mandarin sound changes such as the loss of stop consonant finals, and devoicing according to the aforementioned tonal rules. Zhongyuan Yinyun, created by Zhōu Déqīng () in 1324, show the even tone ( ) divided in upper and lower tones called and , respectively. Apical vowels emerged as a new rime class. The entering tone disappeared, and syllables in the entering tone are distributed between the other tones, but placed after the other syllables with labels such as (, 'entering tone makes departing tone').