Writing in Neolithic China?

In this post we present selected parts of the very interesting and informative paper titled “The Origins of Chinese Writing: the Neolithic Evidence“, by Paola Demattè.

In China, a number of signs from some Late Neolithic contexts suggest that recording activities were well developed before Chinese writing became widespread during the Shang period. Archaeological and palaeographic evidence indicates that mature writing is likely to have evolved from these earlier signing systems as a result of the increasing social and political complexity of the societies of the Late Neolithic. This article analyses three Late Neolithic signing systems that may have led to the mature Chinese writing of the Shang oracle bone inscriptions, and argues that non-linguistic visual signing played a role in the emergence of writing systems.

The debate on the origins of Chinese writing The earliest undisputed texts from China are late Shang dynasty (c.1300–1200 BC) inscriptions on divination bones, mainly from the area of the last Shang capital, Yinxu, near Anyang (Henan). The Shang dynasty (c. 1600–1045 BC), a stratified state society of the Middle Bronze Age, was centred in the middle Yellow River valley, but its hegemonic power extended at times beyond the Central Plain. An important feature of the Shang state was its highly bureaucratic worship of royal ancestors, a system that operated ona ritual calendar and required divinatory inquiries and ceremonial offers. It was around this state-sponsored worship apparatus that various practices that define Shang culture developed. These include the sophisticated bronze industry that produced the food and wine vessels used for ritual offerings, the practice of human and animal sacrifice, and, eventually, an inscriptional system on bones to record oracular inquiries and associated events (Keightley 1978; 1999).

Scholars hold different opinions regarding the dynamics that led to the emergence of Shang writing. Some suggest that oracle-bone writing emerged suddenly with little evolution or at most a couple of hundred years of development.

Others believe thatthe process was much longer and gradual, and suggest that the earliest signs that can be connected with emergence of Chinese writing appeared as early as 4000 years before the oracle-bone inscriptions.

For many (but not all) Chinese scholars Early Bronze Age inscriptions and Neolithic signs are evidence of the gradual development of Chinese writing over an extended period of time and from a variety of earlier graphic systems (Chang Kuang-yuan 1991; Gao Ming 1990, end section 1–2; Li Xueqin et al. 2003).

Theories of writing The competing hypotheses (gradual vs sudden origination) raise questions about the dynamics and timing of the origin of Chinese writing, as well as on the nature of ‘writing’ as a social and cultural phenomenon. The question of the dynamics of the birth of Chinese writing is of foremost importance. Writing theory and empirical evidence indicate that though writing can appear rather suddenly, it frequently does so as a result of an outside stimulus. Sudden invention is more properly associated with the appearance of secondary writing systems, rather than primary inventions (various authors in Houston 2004a). Comparative data from areas of the world where primary writing emerged show how, in cases of independent development, writing undergoes an extended evolution before becoming an efficient language-recording tool.

If the theory of sudden origination is in contrast to available comparative evidence, its timing of the proposed emergence of Chinese writing is at odds with the archaeological record. With an origin around the fourteenth or thirteenth century BC, Chinese writing would have been absent in the early state societies of the Central Plain: the early Shang dynasty and the preceding Early Bronze Age Erlitou culture (c. 1900–1500 BC), a phase sometimes associated with China’s semi-legendary first dynasty, Xia (traditionally 2100–1600 BC) (Liu Li & Chen Xingzang 2003; Liu Li& Xu Hong 2007).

Even though not all state societies have graphic writing, a complex recording system —such as the Inca khipus — is generally necessary to handle the growing bureaucratic needs of a complex political organization, and it is unlikely that both the Shang and Erlitou lacked one (Urton 2003).

Studies also show that oracle-bone writing is linguistically complex and unlikely to have emerged as the result of a sudden invention. This script recorded efficiently (though sparingly) the contemporary language structure, had all categories of speech (nouns, verbs, adverbs, particles), included various sign types (pictographs, semantic and picto-phonetic compounds and phonetic borrowings), and included a grammar that was not dissimilar from that of archaic classical Chinese (Chen Mengjia 1956, 85–134; Norman 1988; Takashima 2004). This suggests that oracle-bone writing reached this level of complexity after undergoing a protracted evolution. In fact, although of great importance, the oracle bones were not the only inscriptions during the late Shang and are not the earliest writing from China. Late Shang ritual vessels were sometimes inscribed with names or brief formulae, and shorter inscriptions and/or single graphs are known from middle Shang (c. 1500 BC) and even pre-Shang bones, bronzes, pottery and jades (Chang Kuang-yuan 1991; Song Guoding 2003). In addition, signs that appear to share elements with early Chinese script have been documented on pottery, jade and bone excavated at various Neolithic sites (Cao Dingyun 2001; Cheung Kwong-yue 1983; Demattè 1996; 1999; Gao Ming1990, 1–2).

The above evidence questions the model of a sudden and circumscribed origin in the middle–lower Yellow River valley during the Middle Bronze Age and suggests that a gradualist explanation is more attuned to data and theory. If so, the documentation of Late Neolithic and Early Bronze Age signs is central to research on the origins of Chinese writing.

Some questionable signs In China as elsewhere, signs that may not be connected with writing include mnemonic marks for simple recording functions (pot-marks or tallies such as lines, crosses, combs etc.), decorative patterns and signs with no clear intentionality (doodles, scratches caused by deposition or fauna). Another group includes those signs whose authenticity or provenance is not clearly established. These types of signs are not uncommon in Chinese Neolithic contexts.

Signs that have been associated with the origins of Chinese writing are a variety of tallies and pot-marks common on Middle to Late Neolithic painted pottery from several sites, particularly in northwestern China. The best known date to the Middle Neolithic (5000–3000 BC) and have been unearthed in Yangshao contexts of the Wei River valley. Their highest concentrations are at Banpo and Jiangzhai, two early Yangshao villages in Shaanxi province (Fig. 1). Yangshao pot-marks are almost always carved on the black band running around the outer rim of select types of red pottery vessels, such as bo bowls. A few marks are carved at the bottom of the vessel. The inscribed containers appear to have been of higher value than other ceramic and were often used in child burials. The signs appear mainly as single entities rather than as compounds, they are stylistically similar, and re-occur on separate vessels and in different villages of the Yangshao horizon. Most were incised before firing with a sharp stone, wood or bamboo tool, though a few may have been carved after firing (Institute of Archaeology, CASS 1963; Wang Zhijun 1980; Xi’an Banpo Museum et al. 1988). These patterns of use suggest that the signs carried social meaning and that mark-making on valuable pottery was an established intra-village convention at Yangshao villages. Interestingly, similar marks are present also on ceramic vessels of other Middle Neolithic horizons of the northwest, such as Dadiwan and beyond, and may have formed an even larger regional tradition (Xie Duanju 2002). Though these facts make the signs valid candidates as potential sources of later Chinese writing, other elements — the overall simplicity of the signs, their early dates and the lack in later phases of development towards a more complex recording system — suggest that this may have been a local tradition with no direct relationship to historic Chinese writing. Perhaps future evidence will clarify their role.

Middle to Late Neolithic: Liangzhu, Dawenkou and Shijiiahe glyphs Quite different is the graphic production of three Middle to Late Neolithic cultures: Dawenkou in the lower Yellow River valley, and Shijiahe and Liangzhu in the middle and lower Yangzi river valley respectively. The signs retrieved from these contexts, which are approximately coeval dating to approximately the middle of the third millennium BC, respond to all the criteria listed above and though simple appear to form an interregional system. Dawenkou and Shijiahe pictographs are carved on large pottery urns (zun), while the evidence from Liangzhu includes both graphs carved on jades and inscriptions on pottery vessels or sherds. These sign systems are interesting because they show a degree of interrelatedness and a conceptual similarity with Shang script (Keightley1989).

Archaeological research has shown that Dawenkou and Liangzhu, which are located in relative proximity, were in close contact with each other, and that the cultural contacts continued also after the demise of Dawenkou and the onset of Shandong Longshan in the lower Yellow River valley. As for Shijiahe, archaeological evidence relating to the preceding Qujialing indicates that this horizon had significant exchanges with Dawenkou culture and that the impact of the central plain on Shijiahe was significant (Gao Guangren & Bian Fengshi 2004).

Dynamics and meanings Evidence suggests that the Dawenkou, Liangzhu and Shijiahe graphs were part of a Late Neolithic inter-regional signing system that emerged from the ritual recording needs of cultures interacting in the political-trading network uniting the Yellow and Yangzi River valleys in the middle of the third millennium BC. The growth of this inter-regional signing network, which is regulated by a relative internal coherence in sign form and usage, is to be understood in relation to the Late Neolithic spread of influence of a regional power beyond its local sphere. Since the Dawenkou evidence appears to be earlier, it is likely that these signs emerged in the Dawenkou heartland of Shandong, and from there spread south to the Liangzhu and Shijiahe territories, where they may have been adapted to pre-existing signing systems. In fact, the spread of these signs may have involved even more areas: comparable signs appear on ceramic vessels from Late Neolithic contexts to the west of Dawenkou. For instance, at Yancun (Linru, Henan) a large funerary jar serving as an adult burial was painted with a combination of three signs — a bird, a fish and an axe — which are in part the same as those of Liangzhu and Dawenkou (Linru County Cultural Center 1981). The southern and perhaps western expansion of this signing practice might be explained in a variety of ways: from the increasing commercial or military might of Dawenkou, to the physical move of Dawenkou people owing to environmental circumstances or climactic duress, or to the sheer power of rituals and associated symbols. Whatever the reason, it is possible that the Dawenkou culture expansion was also responsible for the demise of Liangzhu and Shijiahe and for the eventual development of the Longshan interaction sphere and associated social complexity (Liu Li 2004).

The meaning of these signs is difficult to divine, but based on their patterns of use, distribution and form, some conclusions can be drawn. Usage in all three Neolithic contexts indicates that the graphs were connected to ritual performance. The significance of the ceremonies is not discernible, but at least formally these activities bear some similarity to early dynastic practices of ancestral veneration in which writing played a role.

In the Neolithic, as in dynastic times, these signs appear to have been part of the ritual-political apparatus, though each culture used the inscribed objects in idiosyncratic ways integrating the symbols into pre-existing practices (Shaughnessy 1991; Fung 2000; Shao Wangping 1978).

The distribution of similar signs over a large territory suggests that these graphs probably did not record personal names. Clan names or names of ritual offices are more likely candidates. On the other hand, Dawenkou, Liangzhu and Shijiahe signs may have recorded different types of ceremonies to be performed, perhaps naming specific entities to be honoured. Alternatively, Dawenkou and Shijiahe marks may refer to the content of the jars or to place names indicating the origin of the jars or their content.

Finally, although it has been suggested that the Dawenkou, Liangzhu and Shijiahe signs may be ‘emblems’ or ‘blazons’ with no connection to writing, their forms and set-up recall Shang characters and indicate links with historic Chinese writing.

Are these graphs ‘writing’? The question of the nature of these Neolithic signs in relation to the origins of Chinese writing is bound by our definition of ‘writing’ and our understanding of the processes that led to its emergence. Due to the alphabetic nature of most contemporary writing systems, writing is seen today as inextricably connected with language and is often considered to be simply a tool for recording speech. The similarities between the two are, however, only superficial. No script can faithfully record the nuances of spoken language, and the equation ‘language = writing’ is less prominent in cultures that depend on logographic scripts, like Chinese. The differences between these two semiotic systems are highlighted by their distinct functions and origins. On a general level and leaving aside recent technological advances, we can say that language is associated with face-to-face communication, whereas writing is a visible recording system used to store information beyond the constraints of time and space. Recording systems do not need to rely on speech to effectively store information, in fact those that handle quantities and metrological relationships often transcend speech. For instance, music scores and mathematical equations use a quantitative code that can be understood by anyone who is informed about its logic regardless of linguistic abilities (Harris 1986).

Therefore the question to ask is: did writing originate specifically to record language or did it come into being to record something else (quantities, names, relationships), only to be subsequently transformed into a speech-dependent recording system? I favour the latter view and believe that writing developed out of early graphic recording systems and that language, though key to its eventual transformation, was incorporated into the system only later and gradually. This means that the origins of writing (in China as elsewhere) must be searched beyond its language recording capabilities, focusing on its original function. To do otherwise would be to adopt a teleological approach to the problem.

Evidence from areas of the world where primary writing systems emerged shows that early graphic signs that eventually developed into writing did not necessarily record speech.

This suggests that writing emerged as a systemization of pre-existing visible graphic systems (tallies, counting systems, pictographs, emblems, and narrative imagery), originally designed to handle quantities, kinds and names in pre-literate contexts. It was only at a later stage, when the bureaucratic necessities of political control took hold, that these once loosely organized systems became intertwined, structured and optimized. The systemization could take different forms (including non-linguistic ones such as the Inca
khipus), but apparently in many parts of the world language was chosen as a vehicle.

Bridging the evidence: from Late Neolithic to Early Bronze Age signs If the Dawenkou, Liangzhu, and Shijiahe signs of the Middle to Late Neolithic are likely the initial source of Chinese writing, the transition from these Neolithic pictographs to the mature writing of the Bronze Age is difficult to document. For various reasons, evidence of writing is scarce between 2200 and 1500 BC (the Late Neolithic Longshan era and the Early Bronze Age), though some has been found at Taosi (Shanxi), Chengziyai (Shandong), Shijia (Shandong) and Erlitou (Henan).

The earliest and most compelling is from Taosi, a large Late Neolithic site occupied between 2600–2000BC. A flat-back hu bottle inscribed with red-painted signs was found in an ash pit ascribable to the late occupation stratum (c. 2200–2000 BC). Though the pot is incomplete and some of the signs are difficult to understand as a text, one clearly resembles the archaic form of the character 文 (modern wen ‘writing’). This find is of great importance because Taosi with its rich cemetery, defensive walls, dwellings and early bronze use, fits the expectations of socio-political complexity associated with the use of writing.

In addition, eighteen different graphs inscribed on pottery cups or jars were recovered in the 1930s at Chengziyai, a Longshan era site. The simplest and most common are straight lines, crosses and multiple lines, but the more complex include pictographic forms similar to later forms for ‘leaf’ and ‘wing’. Two inscribed bones were excavated at Shijia (Huantai), a prominent Longshan, Yueshi and Shang walled site with remains of ritual platforms and sacrificial pits. The bones, dated to the Yueshi phase (c. 1700 BC), carry two characters structurally similar to Shang oracle bone writing (Zibo City Cultural Heritage Bureau et al. 1997). Finally, twenty-four different types of graphs carved inside
zun jars were found in the late stratum (1500–1400 BC) of the Early Bronze age Erlitou site. The signs range from single strokes to complex forms, including one resembling the Shang fish pictograph (Li Chi 1976; Institute of Archaeology CASS 1965). Though these inscribed artefacts show the transition from Neolithic graphs to mature Shang writing, more data are needed to complete the picture.

Conclusion Having analysed the graphic evidence from the Chinese Late Neolithic and the competing hypotheses for the origins of Chinese writing, in light of recent writing theory and available archaeological and palaeographic evidence, it is feasible to conclude that the Dawenkou, Liangzhu and Shijiahe graphs represent the beginning thread of the Chinese writing tradition.The forms and usage patterns of Dawenkou, Liangzhu, Shijiahe graphs suggest that, like other examples of early writing from other parts of the globe, these signs could perform simple recording tasks without having to handle sound or grammar. Ultimately, the question is not whether the Dawenkou, Liangzhu, Shijiahe signs are ‘writing’ (this depend on the inclusiveness of the definition), but whether or not they constitute the beginning of a thread that led to Chinese writing. Since they appear to be closely linked to the mature writing of the Shang period, I believe they do.

Research-Selection for NovoScriptorium: Isidoros Aggelos