Dispelling the Myth of the “Tang-Song Transition Theory”


Administrative statutes in the Tang clearly recognized that the fields of commoners could be held through private ownership. Field ownership structures in the recently restored Tang Statutes, while seeming to support ideas of land nationalization, did not actually change the private landowning practices that had been in place since the Qin and the Han dynasties. Numerous tenancy contracts unearthed in Dunhuang and Turfan dating back to the Tang and Five Dynasties show ample evidence that, prior to the establishment of the double-tax system in 780, a highly developed system of contract tenancy was already in place. Tenancy was clearly the leading form of agricultural production outside subsistence farming. This proves that the labor force during the Sui and Tang dynasties consisted not of “slaves and tenant farmers” or “agricultural dependents and serfs” but of commoners who were legally free. The Sui, Tang, and Five Dynasties, as described by Japanese historian Naitō Konan, bear no resemblance to the historical reality of this period. In many instances, Naitō’s arguments have distorted the history of these dynasties in an effort to make China’s history fit neatly into the framework of medieval European history. Consequently, his premises, arguments, and his central conclusion are all wrong. It is crucial that we dispel the myth of Naitō’s “Tang-Song transition theory” and return to historical reality.

1 Originally published in Naitō Konan 內藤湖南, "Gaikatsuteki Tō Sō jidai kan 概括的唐宋 時代觀 [A General View of the Tang and Song Dynasties]," Rekishi to chiri 歷史と地理 9, no. 5 (1922). This section is based on a translation in Liu Junwen 劉俊文, ed., Riben xuezhe yanjiu Zhongguoshi lunzhu xuanyi 日本學者研究中國史論著選譯 [Translation of Selected Works by Japanese Scholars on Chinese History] (Beijing: Zhonghua shuju, 1992), 1.10-18. 2 All direct quotations from Naitō's work in English have been taken from Joshua Fogel's translation of the original Japanese text. See Naitō Konan, "A Comprehensive Look at the T'ang-Sung Period," trans. Joshua Fogel, Chinese Studies in History 17, no. 1 (1983). Transliteration of Chinese names and dynasties have been converted from Wade-Giles to pinyin for consistency. Here, "aristocratic government" is Fogel's translation of what Naitō termed kizoku seiji 貴族政治. As this article is referring to a Chinese translation of Natiō's work, the term used here should be understood as guizu zhengzhi 貴族政治, not as kizoku seiji. Given that it is argued later in this article that the term guizu 貴族 refers to the nobility, not to the aristocracy, of medieval China, and the term guizu zhengzhi refers to a system of government in the pre-Qin centered on the nobility, the words "nobility-based" are added to this translation in order to maintain Fogel's original translation and ensure consistency with other English-language scholarship, which almost invariably uses some form of "aristocratic"' in the translation of this term. In all other instances of the term "aristocracy" or "aristocratic" in a Naitō quotation translated by Fogel, the term has not been altered or added to because the term kizoku as used by Naitō (translated by Fogel as "aristocracy") and the term shizu 士族 used in the original Chinese article (translated here as "aristocracy") both refer to the same thing. -Trans. 3 Naitō Konan 內藤湖南, "Gaikuo de Tang-Song shidai guan 概括的唐宋時代觀 [A Comprehensive Look at the Tang-Song Period]," in Riben xuezhe yanjiu Zhongguoshi lunzhu xuanyi,1.10,13,14. subscribe to the term "Tang-Song transition."4 However, for them, the Tang dynasty was still an ancient period centered on a slave society whereas the Song dynasty was a medieval era based on feudalism and serfdom. This school of thought came to be known as the Tokyo school.5 Naitō's Tang-Song transition theory has had a lasting influence on the study of Chinese history in Japan and internationally6 and has also found many proponents within China. Although some have called the theory into question,7 their reservations have not garnered significant attention.
I understand Naitō's theory to be based on two major arguments. The first concerns the organizational structure of government, whereas the second 4 One of the earlier articles to summarize the specific term "Tang-Song transition" was Ikeda Makoto 池田誠, "Tō Sō no henkaku ni tsuite no saikentō 唐宋の変革についての再検討 [A Reexamination of the Tang and Song Transition]," Nihonshi kenkyū 日本史研究 24 (1954). 5 For the Chinese translation of the article, see Maeda Naonori 前田直典, "Gudai dongya de zhongjie 古代東亞的終結 [The End of the Ancient Period in East Asia]," in Riben xuezhe yanjiu Zhongguoshi lunzhu xuanyi, 1.135-52. 6 Many Tang and Song scholars in the United States and England have accepted the validity of the Tang-Song transition theory. However, although some aspects of the original theory have been retained, others have been modified and revised. For example, Denis Twitchett argues that "Naitō's theory was stated in very general terms. He was not originally an academic historian … but the general outline which Naitō perceived-largely by intuitive understanding-has stood up remarkably well to the progress of modern research" (The Cambridge History of China: Sui and T'ang, 589-906, Part 1, ed. Denis Twitchett and John Fairbank [Cambridge: Cambridge University Press, 1979], 3.9-10). The American scholar Peter Bol has called Naitō's Tang-Song transition theory the "Naitō hypothesis" (" Tang  pertains to socioeconomic issues, and both have a decisive effect on changes in the nature or developmental stage of society. The fact that the Kyoto and Tokyo schools focused their discussions on questions concerning these two areas shows that they were able to grasp the essential issues at stake. It is somewhat of a pity, then, to find that the Sui [581-618], Tang, and Five Dynasties [907-960] as described by Naitō bear no resemblance to the historical reality of this period. In fact, Naitō's "A Comprehensive Look at the Tang-Song Period" is little more than an abstract framework developed according to a Eurocentric paradigm. In an effort to make China's history fit neatly into medieval European history, Naitō's arguments have altered the history of the Sui, Tang, and Five Dynasties in a way that seriously deviates from the historical reality of those periods. As for the changes in the organizational structure of government and socioeconomic structure that Naitō described as leading to this transition, these changes either did not take place during the period in which he describes them or were totally at odds with the historical reality of the time.

1
Propositions by Naitō and His Students Concerning the Organizational Structure of Government during Tang and Song Were Divorced from Historical Reality

Naitō and His Students Defined the "Aristocracy"
Inaccurately as "Nobility" When Naitō speaks of the nobility [guizu 貴族], he is actually referring to the aristocracy [shizu 士族]. By describing the "aristocracy" of the time as "nobility," Naitō defined nobility in a way that was inconsistent with both popular understanding of the term throughout history and the basic characteristics of the aristocracy.
The term guizu appears in written texts throughout Chinese history to refer to the relatives of the imperial family or leaders of ethnic minorities. This meaning is consistent with the understanding of the nobility in medieval Western Europe and the nobility as found in some countries today. Given that, from the past to the present, the term "nobility" has had its own specific meaning, Naitō's unique use of the term to define the "aristocracy" from the Wei [220-265], Jin , and Southern and Northern Dynasties [420-589] until the Sui, Tang, and Five Dynasties is certainly inaccurate.
Throughout history, the nobility-both inside and outside China-has held an entrenched legal status that was hereditary and could be passed on to successive generations. In contrast, what we understand as the aristocracy of the Wei, Jin, Southern and Northern, Sui, and Tang dynasties was not a group with an entrenched legal status that could be inherited.
The legal status of people throughout Chinese history has been determined by the rules set out in administrative statutes and criminal codes as well as in records in household registers. However, whereas statutes and codes in the Wei, Jin, Sui, and Tang periods record distinctions between officials and citizens [guanmin 官民] and subordinated people [liangjian 良賤], they made no distinction between the "upper class" and the masses, thus no distinction between the "aristocracy" and "commoners." Household registers and other related documents, such as household declarations [shoushi 手實]8 and financial registers [ jizhang 計賬],9 also distinguish only between free and subordinated people and between officials and the masses, with no records indicating differences between commoners and the aristocratic elite. This evidence demonstrates the absence of a clearly defined legal concept of the aristocracy at the time.10 Prior to the edict of Emperor Xiaowen 孝文帝 [r. 471-499] of the Northern Wei  in 495, which assigned certain lineages an "aristocratic" status based on their previous bureaucratic positions, the designation of individuals as either aristocratic elite or commoner was assigned through public opinion. This absence of an officially recognized classification scheme made it impossible to have a uniform standard that determined who made up the aristocracy, inevitably leading to difficult situations in which court officials, scholars, and genealogists had different theories about which families should be included.11 8 Household declarations were records of declarations made by civilian families to the village head recording information such as the age of the people in their household and the amount of land they had. These documents would form the basis of the Household Register.-Trans. 9 Financial registers were records of changes in the status or situation of residents of a particular area and were generally made by the village head.-Trans. Although the aristocracy existed as a social group over a long period, its membership was constantly changing. This was due, in part, to the fact that the bureaucratic positions of aristocratic officers were not hereditary. All these points demonstrate the marked differences between the aristocracy and the nobility.
Although some might argue that the Jin dynasty household levy [hudiao 戶 調] system and landowning system [zhantian ketian 占田課田] provided the aristocracy with certain economic privileges, such as exemptions from taxes and corvée labor, this was not actually the case. As the treatise on financial administration in the History of the Jin Dynasty [ Jinshu shihuozhi 晉書·食貨 誌] clearly states, land held by ranked officials was exempt from tax and labor requirements whereas that held by aristocratic landlords was not.12 It would be inaccurate to use the concept of the aristocracy as a substitute for ranked officials, as this would be applying to an entire social level what was true for only some.
Aristocratic landlords and commoner landlords made up two separate groups within the landlord class, with the two groups constituting the foundation of landlord political power. The Wei, Jin, and Southern and Northern Dynasties had no substantial difference in economic position between aristocratic landlords and commoner landlords, and it was customary for the two groups to work together with government officials in order to maintain feudal power. An example of this is the influential Sui ministers Gao Jiong 高熲 [541-607] and Su Wei 蘇威 [542-623]; Gao Jiong was originally from a commoner family, and Su Wei came from an eminent aristocratic family. The two ministers were on close terms with each other, and historical documents record that "during the time that Gao Jiong and Su Wei came together to assist each other, the administrative statutes and criminal codes, regardless of how important or otherwise, were all carefully planned out. Accordingly, it came to pass that, after several years, the fortune of the dynasty was recast and all under heaven was in put in order."13 During the early years of the Tang  Not only was Naitō misguided in his use of "aristocracy" and "nobility" but all his assertions regarding the state of the aristocracy at the time were also incorrect. For example, he notes that from the Six Dynasties until sometime before the mid-Tang, "government in this period was completely possessed by the overall aristocracy, and no one outside it could rise to a high official post."14 On the contrary, many of the top-ranking officials during the Wei, Jin, and Southern and Northern Dynasties came from humble beginnings. According to calculations by Wang Zhenglu 汪征魯 in Research on the System of Official Selection during the Wei, Jin, and Southern and Northern Dynasties, during the Western Jin [265-317], ninety-one of the subjects of biographies in the History of the Jin Dynasty who served as officials came from upper-class aristocratic families, eighty-four came from common aristocratic families, and fifty-nine came from lower classes. This means that a total of 25.2 percent of officials originated in the common people. This group contained thirteen high-level officials at a rank of three or above, namely: Yue Guang 樂廣, Chu Tao 褚陶, Yan Ding 閻鼎, Tao Kan 陶侃, He Fan 何樊, Gou Xi 茍晞, Fan Gui 範晷, Xiong Yuan 熊遠, Wang Xun 王遜, Yu Yu 虞預, Gao Song 高崧, Wei Gai 魏該, and Li Ju 李矩.15 Naitō also noted that when the Tang emperor Taizong 唐太宗 [r. 626-649] ascended the throne, he gave instructions for an inquiry into aristocratic genealogies to be carried out and that "even imperial authority could not alter pedigree ranking."16 Naitō is fundamentally wrong here. In fact, precisely the opposite is true, for it was none other than Emperor Taizong who used the power of his imperial authority to arbitrarily change the rankings of aristocratic families across the empire, raising his own family ranking to the first rank and demoting the eminent Cui 崔 family to the third rank.
Naitō further states that "in the Southern Dynasties as well, the Wang 王 and Xie 謝 families were far more important than the family pedigrees of the emperors."17 In making this point, Naitō was probably drawing on the popular adage regarding the influence of the Wang family from Langya  Naitō, "A Comprehensive Look at the T'ang-Sung Period," 89. 17 Naitō, "A Comprehensive Look at the T'ang-Sung Period," 89.

Yang
Journal of chinese humanities 6 (2020) 129-152 Simas rule the empire together."18 However, because of his scant evidence and limited understanding of the history of the Eastern Jin and Southern Dynasties, Naitō once again got many of his facts wrong. The reality is that this adage referred to a time before the founding of the Eastern Jin, more than a hundred years before the beginning of the Southern Dynasties. At the time, Sima Rui 司馬睿 [r. 317-323], a distant relative of the ruling Sima family, who would later go on to found the Eastern Jin, was garrisoned at Jiankang 建康 (modern-day Nanjing 南京) as merely a prince of the royal family. He was under the command of Sima Yue 司馬越 [d. 311], the prince of Donghai [Donghai wang 東海王]. As such, it was only natural that the two brothers from the eminent Wang family, Wang Dao 王導  and Wang Dun 王敦 [266-324], enjoyed much greater prestige than Sima Rui at this time. When Naitō says that "in the Southern Dynasties as well, the Wang and Xie families were far more important than the family pedigrees of the emperors," he is wrong in terms of both the people involved and the time that it took place.

China Was a Monarchical Autocracy from the Qin and Han Dynasties All the Way through the Ming and Qing
It is widely known that the political history of ancient and imperial China clearly went through two distinct stages of nobility-based aristocratic government and monarchical autocracy. However, nobility-based aristocratic government did not hold sway during the Sui, Tang Although a host of historical evidence points to the contrary, Naitō willfully chose to ignore it, as it did not match the European model in which throughout medieval Europe monarchical power was declining. Quite a few of the events and situations that Naitō considered as taking place in the Wei, Jin, Southern and Northern Dynasties, Sui, Tang, and Five Dynasties never even happened. They were simply a case of Naitō making up history out of thin air. For example, in the putative nobility-based aristocratic government from the Wei to the Five Dynasties, "the monarch was the common property of the aristocratic class; government was effected by recognizing the special powers of the aristocracy; and alone, he could not hold absolute power."21 The situation that Naitō is describing here fits in well with that of medieval Europe. However, the reality is that from the Wei to the Tang and Five Dynasties, no individual enjoyed a higher standing than the emperor. Emperors simply did not need the approval of the aristocracy or nobility to govern their empire. 19 Sima Qian, Records of the Grand Historian, trans. Burton Watson (New York: Columbia University Press, 1993), 58. 20 Naitō, "A Comprehensive Look at the T'ang-Sung Period," 89-92. 21 Naitō, "A Comprehensive Look at the T'ang-Sung Period," 90.

Yang
Journal of chinese humanities 6 (2020) 129-152 During the decline of monarchical power in medieval Europe, it was common practice for the nobility to openly unite and take a stand against the monarch. However, during the period of monarchical autocracy from the Wei to the Five Dynasties, aristocrats, commoners, and ministers alike did not dare to usurp the authority of the emperor unless they had sufficient power to do so.
Naitō noted that the important organs of state in the Tang were the Department of State Affairs, the Imperial Secretariat, and the Imperial Chancellery…. The Imperial Secretariat represented the emperor, while the Imperial Chancellery represented bureaucratic, namely aristocratic, public opinion, but of course all high officials in all three organs of state came from the aristocracy. Thus the aristocracy did not pay absolute obeisance to imperial orders. For that reason, the language used in imperial responses to memorials from officials was extremely friendly and never commanding in tone…. All the [grand chancellors] of the Tang came from the aristocratic class, and it was the custom that upon attaining this post not even the emperor could freely shake their power.22 Naitō's statements here are clearly at odds with the historical reality in the Sui and Tang dynasties. Many of the grand chancellors in the Sui and the Tang came from the common people. Naitō's comment that "the Imperial Chancellery represented bureaucratic, namely aristocratic, public opinion" is simply a flight of fancy. If we look at the facts, all the authority of the three major state organs came directly from the emperor. These institutions were the tools through which the emperor governed his empire, and they existed to serve the emperor. They were absolutely under his command. The separation of responsibilities between the three organs was only a separation of the functions of the empire's governing apparatus. At no stage was there any question as to which organ represented the emperor and which represented government officials. Although the Imperial Chancellery had the authority to deliberate on and reject proposals put forward by the Imperial Secretariat, whether its opinions were followed was determined by the emperor. Not only could the emperor decide whether to listen to his ministers but he could also immediately remove any official who disagreed with him from his post or even have him executed. In short, regardless of their rank or position, every government official was nothing more than a subject of the emperor and, when all was said and done, had no choice but to unconditionally submit to his will.
As for comments such as "the language used in imperial responses to memorials from officials was extremely friendly and never commanding in tone," again this is Naitō imagining Tang China through the prism of declining monarchical authority in medieval Western Europe. There is no evidence of this in China whatsoever.
Naitō also notes that "it was the custom in the Tang that local officials often held the same powers in their localities that the sovereign held at the center."24 Here, he is clearly conflating local officials in Tang China with the local magnates who lived during the period of declining monarchical authority in medieval Western Europe. Naitō does not seem to have understood that the medieval European context was fundamentally different from that of China beginning in the Qin and Han dynasty. From the time that Qin Shihuang unified the empire, China had a centralized government in which authority in key military, political and fiscal matters was all held by a centralized state. All appointments or dismissals of local officials (e.g., commandery governors, commandants, and prefects) were made by the central authorities, with the power of local officials heavily restricted. In an imperial China run by a centralized state, the crimes that local officials feared most were those concerning Naitō, "A Comprehensive Look at the T'ang-Sung Period," 92.

Yang
Journal of chinese humanities 6 (2020) 129-152 individual power grabs, the overstepping of one's authority and the formation of cliques colluding to usurp power.

The Assertions Made by Naitō and His Students Regarding the Socioeconomic Structure of the Tang and Song Dynasties Are Inconsistent with the Historical Reality at the Time
Naitō stressed that during "the era of aristocratic government … the people were looked upon as slaves of the body of aristocrats. In the Sui-Tang years, the people were released from the hands of the aristocracy and came under the direct control of the state."25 Here, Naitō's emphasis on the people coming "under the direct control of the state" is informed by changes that took place in medieval Western Europe, where the practice of feudalism meant that although the monarch was the nominal overlord, in practice, they could only govern the subjects residing within their personal fiefdom. A popular saying at the time described this as "my vassal's vassal is not my vassal." It was not until the later stages of the Middle Ages, marked by the growing power of monarchs coupled with the declining power of the nobility, that monarchs began to wield personal power over the majority of households within their realm and commence compiling statewide household registries. The situation in China was completely different. Even before Qin Shihuang unified China, Zhou [1046-256 BCE] vassal states had already established household registers that put the control of civilian households directly into their own hands.
The amount of material related to state control of households is too numerous to list in detail. An example of this evidence is household data recorded in the histories of the Qin, Han, Wei, Jin, and Southern and Northern Dynasties, which confirms that civilian households were under the direct control of the state in each of these periods. Living at a time when he did not have access to as many pre-Tang household register documents as we do today, Naitō can perhaps be forgiven for making the wild claim that "in the Sui-Tang years, the people were released from the hands of the aristocracy and came under the direct control of the state." However, this just goes to show that many of Naitō's assertions regarding the socioeconomic structure of the Wei, Jin, Southern and Northern Dynasties, Tang, and Song periods do not stand up to historical scrutiny.
Turning now to Naitō's comments regarding taxation reforms in the mid-Tang, we find Naitō stating that the Tang system of zu-yong-diao 租庸調 required that people pay a land rent, serve corvée labor, and contribute goods to the government. From the middle years of the Tang, this system gradually dissolved and was replaced by the double-tax system. This development enabled people to choose their residence freely. Since the land tax could be paid in money, the position of the people began to open up widely from their early slavetenant farmer status in which they had been tied to the land.26 This statement is riddled with obvious errors. First, it should be pointed out that, under the double-taxation system in which land and household taxes were collected separately each year, the land tax was assessed in bushels of grain whereas the household tax was assessed in cash. In practice, however, a certain proportion of household taxes was often collected in cloth; furthermore, there is no evidence that cash was used instead of grain to pay the land tax. Naitō's statements regarding the double-taxation system mixed up which taxes could be paid in cash. Additionally, with regard to workers being tied to the land, as far as I can see, both tax systems were much the same. One of the principles in determining taxes under the double-taxation system was the elimination of the distinction between long-standing and newly settled households, with all households required to register in the location of current residence.27 This did not make it easier for the common people to leave the land. On the contrary, it made it more difficult.
Moving now to the issue of the distribution of land in the Tang, we find Naitō saying that "the position of the people and their private power in wealth underwent a great change from the era of aristocratic government … the system of land distribution was closely related to this … in the Song dynasty, the sense of popular land ownership was gradually secured through the new laws of Wang Anshi."28 In fact, the private nature of land ownership by common people in the Tang dynasty was already clearly acknowledged in the administrative statutes and criminal codes at the time. Examples can be found in the household and marriage section of the Tang

Yang
Journal of chinese humanities 6 (2020)  to "all cases of illegal cultivation of private and public land," "all cases of wrongfully laying claim to or illegally trading or selling public or private land," and "all cases of officials who encroach upon private land."29 Here, the references to "public land" specifically referred to government-owned land, and instances of "private land" refer to land held by the common people. An article that refers to "all cases of public or private wasteland" in the section on land regulations in the recently restored Tang  on land around the capital that had previously been used to provide officials with salaries but was now abandoned, Li Yuanhong explicitly stated that these fields which have been returned by officials are spread across various districts and cannot be grouped together. The common people have worked hard to cultivate their private land and this land cannot be taken from them. Should agricultural colonies be established, their establishment would require an exchange of private land for public land.31 Such a clear conception of land ownership rights coming from the highest echelons of state power was rare in imperial China. Li Yuanhong's views appear in stark contrast to the "public fields" [gongtian 公田] land nationalization policy of the Song grand chancellor Jia Sidao 賈似道 , which forced landowners to sell one-third of their holdings to the state, often for much less than the market value.
Whereas Naitō's key arguments revolved around an analysis of bureaucratic systems, he rarely touched on economics and left out any mention of issues regarding the tenancy system. They were later taken up by his student Miyazaki Ichisada, and two of his articles, "East Asia's Early Modern Age" and "From Agricultural Dependents to Tenant Farmers,"32 focused on the issues of manorialism and tenancy. Technically speaking, ever since Qin Shihuang abolished the patrimonial system of appointing subordinate rulers and created a unified state through centralized power, manorialism and its closely associated seigniorage ceased to exist in China. Written documents from the Qin, Han, Wei, and Southern and Northern Dynasties do not contain any mention of the term "manor" [zhuangyuan 莊園], and no record of the term has ever been found in any official administrative documents, such as household registers, household declarations, or tax registers in any dynasty. Although transmitted texts from the Tang dynasty and manuscripts unearthed at Dunhuang 敦煌 mention the term "manor," it is clear that, as used in the Tang dynasty, it was merely a synonym for "cultivated fields." Thus, we have a situation in which not only were owner farmers in the practice of calling a small section of their fields a "manor" but so were many of the landholding military officers and soldiers who lived throughout the large administrative circuits across the empire. Indeed, even the legendary Dong Yong 董永, the personification of filial piety who sold himself into slavery to pay for his father's funeral, was also in the habit of styling a portion of his fields as a "manor." Needless to say, the term did not hold any specific connotations. The "manors" in the Tang had little in common with manors in Europe or shōens in Japan, and even less with any associated systems of slavery, feudalism, and serfdom.33 Miyazaki's statement that, between the Tang and the Song, the primary labor force in agricultural production shifted from agricultural dependents [buqu 部 曲]  These sayings reveal the important position that male and female slaves occupied in the terms of social production. From the Sui onward, what had once been a "common saying" regarding the consultation of slaves now became a historical aphorism: "the ancients said, for plowing consult a [male] slave, for weaving consult a [female] slave."36 This transformation indicates that consulting slaves in matters of plowing and weaving had become a historical memory for those living in the Sui and the Tang, which in turn illustrates that the role of slaves and agricultural dependents in agricultural production at that time was no longer as great as it had previously been. Miyazaki notes that, the manors of medieval aristocrats … were rented out on land contracts to the tenant who agreed to pay the highest rent. In this way, agricultural dependents in the middle ages were released and replaced by tenant households…. not long after the Song dynasty had been established, probably sometime around the reign of Emperor Zhenzong 宋真宗, the system set up around agricultural dependents completely dissolved.37 This assertion mixes up the timing of events and is another case of Miyazaki seriously departing from the historical reality at the time. At no stage did governments in the Sui and the Tang carry out measures to free large numbers of agricultural dependents. Whether private landlords allowed freed slaves to leave their households was a matter of personal choice, and there is no way that any unified measures could have been taken. The real reason for the decline of agricultural dependents was not their release but, rather, a range of measures in the Tang to limit slavery, which dried up the supply of both slaves and agricultural dependents. Furthermore, tenant households during the Tang did not emerge from agricultural dependents. The bulk of tenant households were made up of owner farmers, part-owner farmers, and people who had been displaced, all of whom were legally free. With regard to the timing of events in Miyazaki's thesis, whereas the agricultural dependent system ended during the Tang, the system of tenancy had been in existence long beforehand. By no later than the early Tang, tenancy arrangements had become the most common farming practice outside subsistence farming.
To date, more than ten tenancy contracts have been unearthed in Dunhuang, all of which are from the late Tang and Five Dynasties, when the area was governed by the Returning to Righteousness Army [Guiyi jun 歸義軍]. In addition, over 120 tenancy contracts (including a few fragments) have been unearthed at Turfan, most of which date back to the era of the Gaochang 高昌 state  and the early Tang dynasty. Together, they total around 140 contracts (one of which records an instance of a tenant claiming ownership and subletting his field, leading to a situation in which one field had two owners),38 which is more than three times the number of tenancy contracts from the Song, Liao , Jin , Yuan [1206[ -1368[ ], and Ming [1368[ -1644 currently known to exist.39 In general, these tenancy contracts had quite a short duration, and more than 85 percent of them for a period of one to two years. This indicates that the relationship between landlords and tenants at the time was temporary and that tenants were relatively free. Although some instances occurred in the Song in which landlords prevented tenants from leaving their land, no such examples of this practice have been found as yet in the Tang. Indeed, whereas Song regulations mandated that the relationship between landlords and tenants was that of master and servant, this was not the case in the Tang and Five Dynasties, and no special regulations existed in the criminal codes and administrative statutes of the time regarding tenants. As noted in the General Principles section of the Tang Code, for "all cases of crimes committed by official bondsmen, agricultural dependents, or government or private slaves, where the specific article has no formal text concerning these inferior classes, treat them as free 38 Yang Jiping, "Lun Tang, Wudai suojian de 'yi tian er zhu' yu yongdianquan 論唐、五代 所見的'一田二主'與永佃權 [On Perpetual Tenancy and the 'One Field, Two Owners' Phenomenon Found in the Tang and Five Dynasties]," Zhongguo jingjishi yanjiu 中國經 濟史研究, no. 3 (2018). 39 According to incomplete data, tenancy contracts dating from the Song, Yuan and early Ming (before the Hongzhi 弘治 period [1488][1489][1490][1491][1492][1493][1494][1495][1496][1497][1498][1499][1500][1501][1502][1503][1504][1505]) currently include no more than ten contacts written in Chinese characters, three tenancy contracts in Old Tibetan 吐蕃, ten tenancy contracts in Tangut 西夏, and around thirteen contracts in Old Uyghur 回鹘. This gives a total of a little over thirty contracts.

Yang
Journal of chinese humanities 6 (2020) 129-152 persons."40 This indicates without a doubt that tenants enjoyed a legal status as free men and that no legal basis existed on which to distinguish between landlords and tenants in terms of higher and lower status or being master and servant. Facts speak louder than words. Naitō and Miyazaki's theory about the Tang's being a medieval slave society and the Song's being a time when "the position of the people began to open up widely from their early slave-tenant farmer status" can no longer stand up to the scrutiny of new material evidence that has emerged in the past twenty to thirty years. The same is true of Maeda Naonori's theory developed with others at the Tokyo school, which regarded the Tang as an ancient slave society and the Song as a medieval and feudal serf society. Prior to the mid-Tang, except for owner farmers, tenancy systems were the dominant structure of agricultural production. This historical reality, confirmed by the tenancy contract system, means that no evidence remains to support the various assertions made by the Kyoto and Tokyo schools regarding whether the use of agricultural dependents in the Tang constituted a serf society or a slave society.

Dispelling the Myth of the "Tang-Song Transition Theory"
The fact that many major changes take place in each relatively long period of history is not surprising in the least. For example, if we examine the Spring and Autumn and Warring States periods, we naturally find that a whole host of changes took place. Politically, these changes included the decline in the power of the monarchy, coupled with the rise of vassal lords. This led to increasing power struggles between expanding states, which in turn laid the foundation for Qin Shihuang to unify China. Major cultural changes also took place during this period and are manifested in the decay of traditional Zhou state rites and music practices as well as debates between competing schools of thought. In science and technology, advances in smelting practices and improved iron tools led to substantial increases in social production.
In the Qin and Han periods, extensive changes took place in government organization and fiscal management. After China's unification, Qin Shihuang fundamentally changed the structure of state power by replacing the previous system of patrimonialism with one of bureaucratic centralism, allowing him to create a unified multiethnic state that was both autocratic and centralized. At the same time, major changes occurred in the makeup of the bureaucracy. In line with the movement toward state centralization, a system of three senior statesmen and nine ministers was established in the capital, whereas in local regions, commanderies and districts were set up and managed by governors and magistrates, respectively. In financial affairs and taxation, new measures for land taxes, poll taxes, and corvée labor were introduced. After the upheaval created by the division of the empire and several partial reunifications that followed, a period of large-scale cultural and ethnic integration occurred in the Wei, Jin, and Southern and Northern Dynasties. During this period, changes in bureaucratic systems shifted real power away from the three senior ministers to the Imperial Secretariat. Changes in the process of selecting officials included the introduction of a new system in which "Recommending Legates" [zhongzheng 中正] were used to promote and rank candidates for posts in the imperial bureaucracy based on nine official ranks [ jiupin 九品]. This system and its subsequent ossification led to the gradual rise of an aristocracy that increasing held political power. In terms of financial and taxation measures, political disorder and the ensuing lack of detailed census data resulted in a shift from proportional taxation to a fixed quota based on land area whereas the poll tax was replaced by a household levy.
Reunification during the Sui, Tang, and Five Dynasties enabled the return of a unified multiethnic state that was both autocratic and centralized. Changes in the government structure facilitated the rise of the three departments and six ministries, and the nine-rank system was gradually replaced by a selection system centered on imperial examinations. With the implementation of new measures limiting the use of slaves, contract tenancy became the leading mode of production, other than subsistence farming. After fiscal changes and reforms in the taxation system in 780, the double-taxation system replaced the previous zu-yong-diao system, in which fixed taxes had been paid in grain, cloth, and labor by each registered taxpayer.
During the Song, partial reunification of the empire once again was followed by the reestablishment of the centralized state. Although socioeconomic advancements proceeded at an unprecedented rate, and cultural practices and academic scholarship flourished, the state was increasingly hampered by its weak fiscal conditions and the lack of military force.
All these changes are just as important as those that took place between the Tang and the Song, if not more so. Yet until now, scholars have rarely encompassed these changes in theories about a transition from one particular period to another, and even when they do, these theories never even get close to being considered as describing "settled historical fact" in the way that popular support has bestowed on the Tang-Song transition theory. No other "transition theory" can lay claim to a such a wide influence over historians and academics, to the point that it becomes the basis for future research.
Only Naitō and his students, who, in comprehensively describing their imagined "major transition" between the Tang and the Song across a variety of topics-politics, economics, culture, intellectual discourse-were able to come to the conclusion that the Tang was the culmination of the medieval period, whereas the Song marked the beginning of the modern era. Accompanying this feat was the packaging of their entire suite of arguments into a single "Tang-Song transition theory." Although this theory, as summarized by Naitō and his students, is seriously divorced from the historical reality of the Tang and Song periods, it nevertheless fit nicely with Naitō's original motives and worked well with the meaning of the term "transition" [biange 變革]. By referring to more than simply a gradual change, the term bian'ge denotes expelling the old and bringing in the new and describes a fundamental change in the intrinsic nature of that which is changing. The connotations of this fundamental change can be seen in Naitō and his student's description of the Tang as an aristocratic government in which the people were slaves and tenant farmers, in contrast to his description of the Song as a monarchical autocracy that further validated the significance of land ownership for commoners. It can also be seen in the contrast between the Tang as the culmination of the medieval period and the Song as the start of the modern era.
With regard to the major political and socioeconomic transitions imagined by Naitō and his students to have taken place during the transition between the Tang and Song, they either never actually took place or were already taking place in the early stages of the Tang long before the period of transition between the two dynasties. For example, as mentioned above, China did not wait until the Song to become a monarchical autocracy-it had been one from the moment that Qin Shihuang unified the empire and claimed the title of emperor. Similarly, changes in the nature of the aristocracy were not exclusive to the transitional years between the Tang and the Song. The decline of aristocratic families was already well underway by the beginning of the Sui with the abolition of the nine-rank system and the establishment of imperial exams. Nonetheless, traces of the aristocracy continued to exist throughout the Tang and Song and could even be found after the fall of the Song.41 One could say that the dissolution of the aristocracy during the transition between these two dynasties was a fait accompli. Furthermore, records in the Tang Code and Tang Statutes show that the right of civilian households to own private land was clearly recognized in the law in the early stages of the Tang. There was no need to wait until after the mid-Tang for such recognition, let alone until the commencement of the Song. The same is also true of changes in agricultural production, with transmitted texts and unearthed manuscripts revealing that slaves and agricultural dependents were no longer the main force in production by the Sui and early Tang. These texts also show the high prevalence of tenancy and contract tenancy, with tenancy clearly being the leading form of agricultural production at the time, other than subsistence farming. These changes occurred before the mid-Tang and long before the transition period between the Tang and Song.
In conclusion, the Sui, Tang, and Five Dynasties as described by Naitō bear no resemblance to the historical reality at those times. Naitō's arguments, in many instances, distorted the history of these dynasties in an effort to make China's history fit neatly within the framework of medieval European history. Consequently, his central conclusion, main arguments, and so-called premises are all wrong. The approach taken by Naitō is not a scientific one in which he undertook research based on historical material; rather, he started with a predetermined conclusion and then selected and interpreted historical materials to support his argument. It is crucial for us to dispel the myth of the "Tang-Song transition theory" and return to historical reality.
The "Tang-Song transition theory" as set out by Naitō and his students is a myth. Another theory, highly regarded in some academic circles, that Zheng Qiao's 鄭樵  statements regarding changes in official selection and marriage practices from the Tang to Song are evidence that he was the first proponent of the "Tang-Song transition theory" is also just a myth. It is crucial for us to shatter the myth of the "Tang-Song transition" as advanced by Naitō and his students and, in line with the tenets of dialectical materialism and historical materialism, make proper use of transmitted texts and unearthed manuscripts from the Tang and Song dynasties (especially unearthed manuscripts published over the past forty years). In this way, we can study the changes that occurred during the Tang and Song dynasties as well as during the intervening period and perform a practical assessment of the status of these changes within the context of Chinese history. (Beijing: Zhonghua shuju, 1977), 46.908. These examples show that, as during the Tang, vestiges of the aristocracy could be found throughout the Song.