Phase I: the dog in Shang and Western Zhou burials

There has been much research across the twentieth and twenty-first centuries on the unique placement of dogs in Shang and Western Zhou burials. This topic has burgeoned to the point where state-of-the-field articles have sought to address current research questions and trajectories with regard to dogs in the archaeological and textual records of pre-imperial China.Footnote ² Archaeological evidence from the Shang period in particular largely presents fragmentary, slaughtered, or dismembered dog remains in three contexts: in waste deposits, in ritual pits, and in relation to human-occupied tombs.

The presence of fragmentary bones in waste deposits from Shang sites like Xiaomintun 小民屯, Anyang, is reflective of the consumption of dogs both as a food source and as a sacrificial victim.Footnote ³ As Li Zhipeng 李志鵬 and Roderick Campbell note, sacrificial practices cannot be fully understood through the study of interred animals in sacrificial pits alone as many sacrificial offerings or practices involved consumption, and these remains are likely to be mixed with regular food waste deposits.Footnote ⁴ Oracle bones certainly mention the sacrifice of dogs, as has been analysed by Zhang Bingquan 張秉權, Ling Chunsheng 凌純聲, and Yang Yang 楊楊.Footnote ⁵ These studies have shown that sacrificing dogs was largely part of ancestral worship, in relation to cardinal directions, or associated with winds.Footnote ⁶ Adam Schwartz's analysis of the Huayuanzhuang dongdi 花園莊東地 corpus of oracle bone inscriptions shows only five instances of dog sacrifice as part of royal ancestral rites carried out by a son of King Wu Ding 武丁 (r. circa 1250–1192 bc), either suggesting that dogs were a prestigious sacrificial offering, as argued by Yang, or a lowly offering, as maintained by Schwartz.Footnote ⁷ However, as Campbell notes with regard to finds at Anyang, ‘each context is different from the others and shows a different pattern of animal deposition, suggesting that a range of ritual practices occurred at Anyang, using alternative taxa and likely having a variety of objectives’.Footnote ⁸ Dogs may have been sacrificed in certain ceremonies more so than in others, but in totality were one of a range of domesticated animals raised for consumption and sacrifice in the Shang period.Footnote ⁹

Two ceremonial practices single the dog out above all else in Shang rituals: the consecration of buildings and the waist-pit (yaokeng 腰坑) of tombs.Footnote ¹⁰ The former aspect is evidenced by finds at Xiaotun 小屯, Anyang, with dog skeletons being found in ceremonial pits in Compound C and in seven important buildings in Compound B. According to Shi Zhangru's 石璋如 analysis, these burials were associated with the consecration of: a building's foundations; pillar bases; and the threshold when buried with a human victim.Footnote ¹¹

Unlike the ‘ritualised destruction’ seen in consecration victims for buildings or elaborate tombs, where the focus lies on the act of killing and an anonymised ‘stacking’ of victims, so-called death attendants present an individualised treatment and preservation of the body.Footnote ¹² Dogs and horses are by-and-large the only non-human death attendants found in Shang burials from Anyang and the Central Plains, suggesting that both animals were, for some reason, differentiated and selected for this purpose.Footnote ¹³ While the topic of horses and their interactions with humans is one worthy of further research, this article will predominantly focus on the dog because of the unique transformation it underwent in relation to the tomb in the post-Han period. Nevertheless, this will not be the last time we see the horse and dog singled out from other animals.

Turning to tombs, dogs are most often found in waist-pits in many types of tomb across the Shang period. As Campbell has noted with regard to finds at Anyang, humans and dogs could both be positioned in waist-pits, on tomb ledges, and in associated tomb fills.Footnote ¹⁴ With a noteworthy increase in the presence of dogs rather than humans in the waist-pits and tomb ledges of smaller burials, he asserts that dogs could have been substituted for humans in certain contexts.Footnote ¹⁵ For Shang tombs, approximately 50 per cent of dogs were placed in waist-pits, 30 per cent in related ceremonial pits or tomb fill, and only 20 per cent on ledges in smaller burials. Equally, half of the waist-pits at Anyang were empty, a feature that occurred in 60 per cent of tombs, suggesting that the function of the yaokeng could be accomplished merely by its existence.Footnote ¹⁶

Many Western Zhou tombs with yaokeng can be found in Shaanxi, Gansu, Shandong, Beijing, Shanxi, and the heartland of the previous Shang dynasty in Henan, with this practice being heavily associated with Yin 殷 adherents.Footnote ¹⁷ Clearly, the burial of dogs in waist-pits was regionally divergent and, particularly after the fall of the Shang, proliferated in certain areas but not others.Footnote ¹⁸

This practice dwindled, particularly during the Eastern Zhou period, and is interpreted by Han Wei as marking the dissemination of Zhou rituals in place of Shang practices.Footnote ¹⁹ The number of tombs with yaokeng containing dogs drops off rapidly and texts mention straw dogs, which may have been used in burials, though none seem to have survived.Footnote ²⁰ Clearly, dogs were still understood as being part of the ritual space of the tomb, but figurines could be substituted for their physical bodies. This coincides with the wider trend of using figurines in tombs, as reflected at a number of sites, which will be explored in the following section.

The preservation of the dog's body suggests that it was intended to serve the deceased in the next life as part of their retinue, and this again reflects a markedly distinct appraisal of the dog in relation to the ritual space of the tomb. When a human appears in the yaokeng with a dog, they are usually armed, which is suggestive of hunting or warfare.Footnote ²¹ Here, this has led many scholars to interpret the function of the dog as a tomb guardian needed to protect the deceased from threats either within the tomb or on the passage to the next life.Footnote ²² Arguments have also been made for a continued association of the dog with the four directions and the wind in these waist-pits, either in terms of geomancy or exorcising ‘evil winds’.Footnote ²³ Equally, the placement of dogs on the tomb ledge has been understood as marking the dog as a companion in death, that is, a beloved dog owned in life accompanying its owner in death.Footnote ²⁴

The argument for the dogs being companions in death is challenged by the findings of Li and Campbell, wherein many Shang dogs interred at Anyang were puppies, which is especially notable when compared with midden deposits in other sites that attest to the use of adult dogs as a food source. This suggests that the mortuary use of puppies in tombs was likely to have been economic in nature, as this involves less expenditure in raising a dog that would ultimately be killed. Of course, there may have been a ritual significance to the interment of puppies that, as yet, remains undiscovered.Footnote ²⁵ Equally, Song Yanbo's analysis of Western Zhou skeletal remains further attests to young dogs being buried in these waist-pits.Footnote ²⁶

If these dogs were not companions-in-death, then what were they? Clearly, Shang ritual culture associated the dog strongly with the tomb, especially given the fact that while other domestic animals could be used in sacrifice and consecration ceremonies, they rarely replaced the dog in the context of the yaokeng. Without explanatory texts from this time period it is difficult to be sure why humans from the Shang heartland singled out the dog for this ritual purpose nor what they felt such an interment signified. The Western Zhou period saw acute regional differences in the use of yaokeng, with new adaptations emerging. Western Zhou tombs at Puducun 普渡村, Chang'an 長安, all included items in the yaokeng alongside the dog skeleton, such as cowrie shells and jade pieces, which perhaps reflects alterations to the associated ritual practice or ceremony accompanying the burial of the dog under the tomb.Footnote ²⁷ Equally, other areas have little to no attested yaokeng. Even during the Shang, the Central Plains saw inter-regional differences around each metropolis ‘as if each megacenter was its own vortex and crucible of cultural transformation, pulling in regional traditions and creating new syntheses that lasted for a couple [of] centuries before becoming only part of the next metropolitan order’.Footnote ²⁸ The continued uneven distribution is thus reflective of the adoption, adaptation, and discarding of wider ritual practices, as well as the dog's role therein. Nevertheless, as we will see, the dog continued to be associated with the tomb to a greater degree than other animals.

While covering an exceedingly long time span and despite showing regional incongruencies, the Shang and Western Zhou sites discussed tie the dog to the tomb as a ritual accessory. Whether intended to act as a guardian, an appeaser of the winds, or a sacrifice consecrating the tomb, the dog was part of the ritual furniture of the tomb and its associated ceremonies. In Phase I, then, we can understand the burial of dogs as being closely tied to ritual practices surrounding human burials.

Phase II: the dog in Zhou and Han burials

As has been noted, the dog was no exception to the overarching introduction of substitute miniature figurines, or, in other words, a move towards ‘miniaturization, reduction and replacement’, in mortuary practices from the late Shang period onwards.Footnote ²⁹ Here, the lines between sacrifice and grave good become blurred—as physical dogs were replaced by figurines, did their ritual significance change? How was the dog, here as a miniaturised substitute, redefined in relation to the tomb? And what did the burial of physical dogs mean in this new dog-tomb dynamic?

The first issue is that such a transition is difficult to trace as the very materiality of these figurines means theoretically few survive.Footnote ³⁰ Equally, the increased use of figurines was not a change that occurred overnight nor did it necessarily involve the total replacement of physical dogs from the late Shang period onwards, as has been seen in the aforementioned Western Zhou sites.Footnote ³¹ This tendency to turn to figurines, however, is particularly salient when considering the sheer quantity of ceramic dog figurines unearthed in Han tombs, a period which saw a further transformation in the conceptualisation and range of grave goods included in the tomb.Footnote ³²

Despite evidence of grave good substitutes emerging in the late Shang period, according to Alain Thote it was during the Eastern Zhou period that tomb contents truly began to resemble ‘visual documents’ of life.Footnote ³³ Figurines within the tomb were used to represent property for the use of the deceased while they were in the tomb or when they went to the land of the dead. As Armin Selbitschka's analysis of miniature granaries as grave goods has shown, it would seem that grave goods were intended to furnish the deceased for a ‘perpetual afterworld that was imagined inside of tombs’.Footnote ³⁴

As his earlier analysis of grave goods more generally attests, there were two coinciding motivations for the use of grave goods: the first being related to nourishing the dead, that is, model granaries, wells, and stoves, and the second substituting, specifying, and enlarging the palette of people, animals, and items that could be included in the tomb. Figurines of people have been identified as being particular kinds of people not usually interred as sacrificial victims, including named family members. So too does the burgeoning range of architectural buildings appearing from the late Western Han 西漢 (206 bc–9 ce) onwards, including workshops and multi-chambered residences, illustrate a focus on social status. These trappings rendered the tomb a ‘microcosm’ of owned property in life.Footnote ³⁵ These grave goods provided both sustenance and income equivalent to a real-life estate, reflecting the expanding scale of private estates amassed by tomb occupants in this period.Footnote ³⁶

Building on Selbitschka's arguments for an increasing focus on the deceased's social persona and associated property throughout the Han period, I contend that ceramic dogs thus came to embody themes of property. First, these ceramic dogs were not placed inside yaokeng and so must have had an additional significance beyond substituting for the waist-pit dog that was once utilised for hunting, guarding of the tomb or the deceased, exorcism, or geomancy. Secondly, nearly all known examples date to the first century ce onwards—suggesting that these ceramics were part of the same transformation discussed by Selbitschka with regard to the increased range of grave goods incorporated into late-Western Han tombs.Footnote ³⁷

Dog figurines dating to the late Western Han and Eastern Han 東漢 (25–220 ce) periods from the Central Plains, particularly Henan and Shaanxi, are often depicted as harnessed, baring their teeth as if snarling, or else with open mouths as if barking (see Figures 1 and 2).Footnote ³⁸

Figure 1. Ceramic dark green-glazed dog, shown to be snarling and wearing a harness, dated to the Eastern Han period. H. 26.7 cm. 1991.253.1. Source: Metropolitan Museum of Art, New York. Open access image.

Figure 2. Monochrome image of a ceramic green-glazed dog, also depicted as barking and wearing a harness, dated to the first–second centuries ce. H. 21.4 cm. 1994.605.18. Source: Metropolitan Museum of Art. Open access image.

Many of the individual ceramic figures housed in museums are strikingly similar. While giving little indication of scale, the fact that many wear a harness attests to the size and strength of these dogs and to the fact that they were tethered.Footnote ³⁹ This is clearly a defensive image of a guard-dog which could therefore be misinterpreted as encompassing the supposed guardian role of the once-buried yaokeng dog. However, what was the yaokeng dog protecting? In Shang and Western Zhou tombs, it would seem that it was protecting the ritual space of the tomb or, given its placement underneath the body, perhaps the deceased personally. In Han tombs, however, this does not seem to be the case. Unfortunately, it is not clear exactly where many of the ceramic figurines were discovered in the tomb chamber. Nonetheless, earthenware domestic scenes featuring dogs add to our ability to interpret these figurines.

In the first instance, dogs can be seen next to ceramic stoves or in pens, presenting the dog as a source of meat and thus as a way of nourishing the deceased.Footnote ⁴⁰ This is in keeping with the first motivation for grave goods identified by Selbitschka, that is, to provide eternal nourishment to the deceased. Indeed, dog meat was consumed to some degree in certain regions, with Han rubbings in Shandong often showing a dog hung from a well-sweep for skinning and gutting.Footnote ⁴¹ Just how widespread the consumption of dog meat was and whether it was considered a luxury food reserved for the elite is unclear.Footnote ⁴² Nevertheless, ceramics of this nature underline the continued role of the dog in culinary history.

However, in line with the increased range of domestic buildings seen in late Western Han and Eastern Han graves, several architectural ceramics from this time period feature a guard-dog prone in front of an entrance, a very different placement to general livestock, which were positioned at the rear of such buildings.Footnote ⁴³ These dogs could be positioned lying down with their ears erect as if listening out for danger.Footnote ⁴⁴ There are also further examples of dogs sitting up with pricked ears.Footnote ⁴⁵

If these domestic scenes represent a miniaturised estate of the deceased, why are dogs so often added to the entrances of buildings, barns, and courtyards? As these scenes seek to recreate ‘a working world in miniature’, textual references to living dogs may make this placement clearer. Beyond hunting, a topic we will return to in due course, dogs were certainly used for guarding and ratting from at least the Zhou period onwards.Footnote ⁴⁶ While seeking to define the term for dogs—gou 狗—through homophony, it is their role as guard-dogs that is underlined in this Shuowen jiezi 說文解字 entry:

As Erica Fudge argues, the study of animals is ‘not the history of animals; such a thing is impossible. Rather, it is the history of human attitudes towards animals.’Footnote ⁴⁹ A dog's ability to guard is a canine behaviour that, when interpreted, written, and discussed in human-authored sources, becomes imbued with human consequences. This is manifested in texts like Fengsu tongyi 風俗通義 which reference dogs’ guarding abilities:

The mere presence of dogs is thus deeply associated with security, as is made clear in an analogy from Yantie lun 鹽鐵論 on border defences:

A dog is thus seen to be as essential to defining and defending the boundaries of a household as walls are. Dogs were not only used to physically patrol private borders but were also enmeshed in the very concept of security and peace within this context. It is clear, then, that the role of the dog in these domestic scenes is to protect the property of the deceased rather than the deceased directly.

As private estates in the late Western Han period expanded, the imagined afterworld of the tomb was increasingly equipped to provide eternal self-generating sustenance, wealth, company, and luxuries for the deceased. A dog, essential for guarding such property, was thus placed in these miniature scenes to watch for danger, and we can surmise that the larger individual ceramic figurines fulfilled a similar function. In Phase II, the dog was no longer there strictly to protect the ritual space of the tomb, and perhaps this function was entirely defunct. Instead, it was charged with guarding the property of the deceased. The dog was thus no longer a ritual accessory posed in the yaokeng but could be placed elsewhere in the tomb to fulfil this enlarged role.

How, then, should actual burials of dogs in human-occupied tombs from this second phase be interpreted? Several unusual finds have unearthed dogs buried with coffins or extravagant trappings. For instance, the excavation of Marquis Yi of Zeng's 曾侯乙 (circa 477 bc–433 bc) tomb shows a female dog buried in a wooden coffin accompanied by two stone discs and a bone object.Footnote ⁵³ In addition, the excavation of the tomb of the King of Zhongshan 中山 (circa 344–308 bc) has revealed two dogs buried in golden necklaces.Footnote ⁵⁴ These dogs were not positioned in a manner in keeping with yaokeng dogs, and the inclusion of objects alongside or on the bodies attests to their marked individualisation. It would seem that these dogs were specifically selected for inclusion in the tomb and were probably owned by the deceased in life. It is likely, given that these were elite tombs, that these were prized hunting hounds.

Hunting was undeniably linked with the royal clan in Shang and Zhou times, though even then the sport had its adherents among the wider populace.Footnote ⁵⁵ It would seem that a variety of dogs could be involved in the hunt; Han art certainly shows clear distinctions between different ‘breeds’ or roles of dogs within the hunt, whether sight-hounds or scent-hounds, pointers or sprinters. Hunting may have fostered a great interest in the collection of rare and prized hunting hounds, perhaps coming as tributes from vassal states. For instance, the last Shang ruler Zhou Xin 紂辛 (r. 1075–1046 bc) was depicted in later historiography as infamous for his extravagance and debauchery, as reflected by his varied collection of dogs (presumably for hunting), horses, and other rare objects.Footnote ⁵⁶ While his collection of dogs in the later records of the last Shang king underline his debauchery, which lent moral support to the Zhou conquest, this did not necessarily dissuade Zhou kings from similar practices. This is clearly articulated by the tribute of an ao 獒 dog by the Lü 旅 tribe to the first Zhou king Wu 武 (r. 1046–1043 bc).Footnote ⁵⁷

While it is difficult to prove that these particular dogs were rare hunting hounds perhaps imported from afar, it seems likely that they were at the very least prized hounds on two fronts. First, these dogs are placed incongruously from the ritually significant yaokeng dog seen in the Central Plains in the Shang and Western Zhou periods. Secondly, the use of coffins and precious objects effectively denotes these animals as art objects—de-animalising and redefining them as prized possessions. In keeping with Selbitschka's criteria for assessing prestige items in tombs, these burials clearly elevated the dog to a luxury item.Footnote ⁵⁸ Consequently, physical dogs found in human-occupied elite tombs, when placed so incongruously in the tomb chamber, were not ritually significant but socially significant. In this context, the dog embodies both political dominion, expressed by the receipt of tribute and the collection of exotics, alongside implied dominion over nature, as articulated by the very act of hunting itself.

In Phase II, then, we see two major shifts in the ways dogs were positioned in the tomb. Physical dog remains could be elevated into the tomb chamber, losing their ritual significance and incorporating themes of dominion, status, and luxury. Ceramic dogs from the first century ce not only exemplify the overarching trend towards miniaturisation, reduction, and replacement seen across the previous millennium, but also underline the growing relevance of social status to the conceptualisation of grave goods. Dogs were repositioned to guard the miniaturised property of the deceased, much as they did in life, rather than the ritual space of the tomb. The dog, in both cases, was not so much a ritual accessory as it was emblematic of changing themes of dominion, ownership, and property within this timeframe.

Phase III: a tomb of one's own: changing attitudes to the burial of dogs in their own tombs

What is clear in the previous two phases is that the death of the aforementioned dogs was unnatural, and timed with the burial of the deceased human. But what happened when a non-livestock dog died? Some were certainly consumed, given that the Qin statues discovered in Tomb 11 Shuihudi state that hunting dogs killed by park wardens should be eaten.Footnote ⁵⁹ The metaphor zougou peng 走狗烹 (the running dog is boiled) also references a non-livestock dog being cooked for consumption.Footnote ⁶⁰ According to Olivia Milburn, this metaphor uses the servile dog discarded on its death to comment on the fate of ministers, even though the consumption of working dogs after their death does not receive condemnation elsewhere.Footnote ⁶¹

There are two pre-imperial textual references to the burial of dogs in their own graves. In Yanzi chunqiu 晏子春秋, Yan Ying 晏嬰 (circa 578 bc–500 bc) is said to have criticised Duke Jing 景 (r. 547 bc–490 bc) of Qi's 齊 desire to have his zougou 走狗 (running dog) buried in a coffin with accompanying ceremonial sacrifices. Such criticism arose from the inequity Yan Ying, and indeed many other philosophers, saw in expending resources, particularly in death, on an animal in place of a human.Footnote ⁶² Second is a dog raised by Confucius himself, as recorded in Liji 禮記. Here, frugality is highlighted by Confucius, stating that worn cloth could be used to bury a horse or dog.Footnote ⁶³ The latter example from Liji appears to coincide with a later Han memorial, to be discussed, and so may be the result of heavy Han-era editing. Nevertheless, taking both texts together elucidates two points: the first being that once again the dog and horse are unique animals that operate on a different ‘level’ to other animals. Second, if a dog were to be buried, the burial should be highly economical. Confucius advocates using worn-out cloth or even a seating mat for this purpose, and it is the lavish trappings requested by Duke Jing that were most openly criticised by Yan Ying. The fact that in the end Duke Jing boils and serves up the dog is a marker of the parable nature of this anecdote, though this still suggests that the lavish post-mortem treatment of dogs, as we have seen in the two elite tombs above, was not inconceivable to people of the time.

Under certain circumstances, then, it would seem that non-livestock dogs could be deserving of a burial of their own as long as these burials were frugal and took place without excessive ceremonies. Gu Yong's 谷永 (d. circa 11~8 bc) memorial given in Hanshu 漢書 further explores the circumstances allowing for the burial of a dog or horse:

Evidently, it was the labour that dogs and horses performed on behalf of humans that allowed for their burial. But these burials might not always have been as frugal as thinkers advocated. The earliest known example of a dog-occupied tomb appears to be the Xin 新 (9–23 ce) burial of a small dog at Yueyangcheng 櫟陽城, Yanliang 閻良 (Shanxi); despite its simplicity, the placement of 31 small figurines in the tomb suggests a burial with some level of pomp and circumstance.Footnote ⁶⁶

The concept of ‘toil’ as an essential part of justifying burying dogs and horses is also forwarded by Zheng Xuan 鄭玄 (127–200) in his commentary on the aforementioned burial of a dog by Confucius:

In this commentary, the horse is specified as being owned by a jun 君 (ruler). This reinterprets the original Liji passage regarding burying a dog or horse as being restricted to the elite. Since Zheng Xuan commented on a later edition of the Liji compiled by Dai Sheng 戴聖 in the Western Han period, it may be that his interpretation was shaped by this edition or by prevailing attitudes of the time. Here, it would seem that both status and servitude were required to rightfully bury dogs and horses.

Returning to the Xin burial of a dog, the inclusion of ceramics may then be reflective of its master's status rather than affection for the dog. It would seem that across the first–second centuries there was a subtle shift regarding both how to bury dogs and horses and who should carry out such practices. Unfortunately, there is scant archaeological material attesting to further dog-occupied tombs for us to assess how these debates were actualised or the geographical spread, timeframe, or ubiquity of such burials. Nevertheless, the fact that such discussions even occurred heavily implies that people did bury their dogs, and so the burial of dogs in recompense for their labour in a manner befitting the master's social status around the first–second century ce will be termed as Phase III.a.

Though there is a continued dearth of archaeological materials to explore in the fluid post-Han landscape, it is clear that attitudes towards dog-occupied tombs underwent a stark reconceptualisation in this timeframe. This is manifested in anecdotes or zhiguai 志怪 (anomaly tales) from the post-Han period. A mixture of fiction, exaggeration, observation, and fact, these tales record and praise dogs performing heroic tasks on behalf of their owners.Footnote ⁶⁸ Rather than being transgressive shape-shifters poised on the threshold of life and death, canine-centred zhiguai in the post-Han period were unique in paralleling tales of filially loyal servants.

This striking phenomenon was first noted by both Fu Kaijing 付開鏡 and Keith N. Knapp in the context of filial animal tales more generally.Footnote ⁶⁹ Building on their research, I would like to first outline instances of dog burials in these tales before moving to contextualise the overarching transformation of the dog's relationship with the tomb.

One example of a dog being buried is the Shuyi ji 述異記 account of a hunting dog named Huang'er 黃耳 (Yellow Ears) delivering and returning with a letter for its homesick master Lu Ji 陸機, covering a great distance in a very short amount of time. The dog is explicitly said to have understood human speech and communicates with humans to agree to take the letter, to cross over lakes and rivers, and finally to obtain a letter in reply from Lu Ji's family. When the dog later died, it was buried in its owner's village graveyard.Footnote ⁷⁰ This account is given in a condensed form in Jinshu 晉書 with minor changes.Footnote ⁷¹

Furthermore, two similar Soushen ji 搜神記 accounts tell of a dog saving its drunken owner from a wildfire by shaking water onto the surrounding grass and, when its owner then falls into a well, barking to attract help from a passer-by. In one account this dog is owned by Yang Sheng 楊生 and, both surviving this ordeal, they remain inseparable.Footnote ⁷² In the second, the dog Heilong 黑龍 (Black Dragon) dies from exhaustion, having saved its owner.Footnote ⁷³ His owner Li Xinchun 李信純 is devastated on seeing Black Dragon dead and, from his damp fur, surmises what has happened. He tells the governor of this matter, who replies:

The dog is then buried in a burial mound. A further example of a dog burial is that of Hucang 鵠蒼 (Swan Grey) who was buried after its sudden transformation into a dragon when on the brink of death.Footnote ⁷⁵

Here, we see three examples of dogs being buried after performing monumental feats involving canine skills like running, barking, and shape-shifting. The burials are also much more extravagant—Huang'er is buried 200 steps from its owner's home, while Heilong is buried with an inner and outer coffin in a burial mound over ten zhang tall. While we lack the archaeological evidence needed to ascertain whether these post-Han burials were as extravagant in actuality, it is clear that the logic underpinning these burials has matured and developed in the post-Han period.

The tales of Huang'er and Heilong in particular reflect themes of filial loyalty seen in similar accounts of filial slaves. As Knapp notes, one tale of canine loyalty even directly parallels that of a dutiful son: a hunting dog named Diwei 的尾 (Real Tail) bites to death a snake which had constricted around its owner Hua Long 華隆. As Hua Long falls unconscious, Diwei barks to induce Hua Long's companion to follow it and carry Hua Long to safety. It refuses to eat for two days until Hua Long recovers.Footnote ⁷⁶ Diwei's refusal to eat parallels similar human cases, for example, the dutiful son Ru Yu 汝郁 who refuses food when his mother grows too ill to eat.Footnote ⁷⁷

The reference to reciprocity made by the Governor in relation to Heilong is key in understanding the changed conceptualisation of dogs as possessing human morals and an understanding of filial loyalty. Rather than simply labouring for humans and thus deserving recompense, or bao 報, from their master to deserve a burial, it is instead a dog's performance of bao that grants it a burial. As Knapp has shown, certain animals were credited with understanding bao in direct relationships across post-Han accounts. Nevertheless, the connection with themes of filial loyalty, as expected of slaves, is unique to dogs.Footnote ⁷⁸ The incorporation of the human master-servant relationship into the conceptualisation of the human-dog relationship speaks more to a humanisation of dogs than to a de-humanisation of servants. This is because the attribution of human morality, that is, a limited understanding of bao and filial loyalty, attests to dogs possessing human, even superhuman, moral potential.

In the third to fifth centuries, tales emerged of loyal dogs that would perform heroic acts to save their owners. Should they survive, they would be treated with great affection by their owner and, should they die, they would be buried in a manner that elevated the status of the dog beyond its master's standing. This then differentiates Phase III.b from Phase III.a in that the burial of dogs is no longer restricted by economic concerns but rather by emotional concerns centred on the theme of filial loyalty.