Heijō Palace: Governing One Transaction at a Time

Two wooden labels recovered from Heijō Palace record shipments of processed bonito from Izu, a province well to the east of the capital. One, dated to the ninth month of 733, identifies a district, a village, a named contributor, and a quantity of fish delivered as tax. The other, from the tenth month of 769, describes its bonito as goods obtained through exchange or procurement. The commodity was much the same. The reason it belonged to the government was different.^[1][2]

That difference could not be read from the fish. Someone had to write it down and keep the writing associated with the consignment. Without the label, a receipt of food would lose part of its administrative meaning: where it came from, what quantity was declared, and which kind of transaction it represented.

The two pieces are separated by decades and by their findspots within the palace. They do not demonstrate that purchasing replaced taxation. They do reveal a court using more than one way to acquire supplies—and using small written distinctions to keep those ways intelligible. This is a useful entrance to Heijō Palace, whose grandest buildings explain where authority appeared, while its discarded records help explain how authority worked.

At Nara, where the capital was established in 710, the palace formed a government compound containing the ruler's household as well as the places where officials conducted business. Its ceremonial architecture presented that institution at its most concentrated. The First Great Audience Hall, or Daigokuden, stood within a large enclosed court used for major state ceremonies during the earlier part of the Nara period. The building now standing there is a reconstruction; it gives volume to a setting whose ancient form has had to be recovered through research.^[3]

The hall makes it tempting to picture government as a single gathering around the sovereign. The excavated offices require a different scale of attention. In one northern sector, archaeologists traced successive arrangements of buildings associated with food administration. Their reconstruction includes offices, storage areas, and increasingly differentiated groups of buildings. Many working structures had posts set directly into the ground; their roofs need not have resembled the imposing tiled roof of the audience hall.^[4]

This variety was functional. An audience brought participants together for an occasion; a storehouse held things until they were required. An office needed to distinguish one consignment or request from another. Putting all three under the word palace can obscure the different problems they solved. The compound had to support actions that happened at different speeds: a ceremony with a fixed occasion, a delivery arriving from elsewhere, and a written instruction awaiting execution.

The wooden records, known as mokkan, belong to those smaller intervals. Their value is not that they expose a secret government behind a ceremonial façade. Ceremony and routine were different activities within the same institution. Food, personnel, buildings, and documents had to remain available between the moments when sovereignty was publicly displayed.

The 733 bonito label compresses a chain of identification into a narrow strip. It gives Izu Province, Naka District, more local divisions, and a contributor's name before stating the commodity and quantity. Its declared weight uses contemporary units, kin and ryō. The object was found in a ditch in the government-office area east of the later audience-hall and residential precincts.^[1]

Such labels had a place in the legal arrangements for taxation. The research institute's account of the tax regulations explains that identifying information could be written directly on the edge of a textile payment. For goods such as salt or marine products, the prescribed solution was a wooden label attached to the package. The information and its physical carrier were adapted to what was being delivered.^[5]

This small design choice made a distant obligation easier to recognize at its destination. A parcel could carry not just food but the information needed to place it within a larger set of obligations. Repeating the same kinds of fields—origin, person, commodity, quantity—made unlike shipments more readily comparable. An official did not need to know the contributor personally to understand what the label claimed.

There are limits to what the surviving claim proves. The name on a tax label does not identify every person who caught, processed, packed, or transported the fish. Nor does the declared quantity guarantee that it arrived intact. The label preserves the government's way of identifying a payment, rather than a complete history of the work behind it.

That distinction also guards against mistaking administrative precision for administrative success. An elaborate address could assist checking; it could not, by itself, prevent a shortage or settle a dispute. What the object establishes is the existence of a practical means of assigning responsibility across distance. At the palace, the written parcel could be treated as a particular obligation, not merely as another quantity of fish.

The 769 label introduces another category: kōeki, exchange or procurement. Its wording records Izu bonito and a quantity, but not a purchase price or an individual bargaining encounter. It is therefore evidence for a different route into government supply, not a window onto a fully recoverable market transaction. Its eventual disposal in the palace's southeastern ditch places the record within the administrative compound without revealing every stage of its journey.^[2]

Taxation and procurement created different explanations for a receipt. Once food had arrived, further decisions remained: where to hold it, how to account for it, and when to issue it. None of those decisions follows automatically from ownership. The gap between acquisition and use is where much of the palace's routine work becomes visible.

Food archaeology offers a particularly direct reason for that work. A research-institute account of Nara-period dining reports textual evidence that officials received meals at the government's expense twice a day. Large quantities of standardized vessels are interpreted in connection with organized food service. These findings concern evidence from the capital more broadly, so they should not be converted into a single imagined palace dining room or a menu shared by every official.^[6]

Even with that boundary, the institutional problem is clear. Supporting people while they worked required recurring allocations, rather than occasional displays of abundance. Standard vessels could serve repeated distributions; a consignment label identified the incoming supply. Between those two ends lay decisions about entitlement and issue.

The excavated food-administration sector supplies the spatial counterpart. Separate office and storage areas, as interpreted by its excavators, allowed things and information to be handled differently. A building containing supplies was not itself an instruction to release them. For an unusually specific glimpse of the authorization that could connect the two, another wooden tablet is more revealing than either bonito label.

Catalogue tablet number 1 from Heijō Palace is a request from a temple. It lists red beans, fermented seasonings, and vinegar, and connects the request with Tsukuba no Myōbu, a court woman. Its date is the sixth day of the third month. No year survives in the inscription. It was recovered from a disposal pit in the sector associated with the palace's food-service administration.^[7]

The missing year and unnamed temple have to be supplied, if at all, by other evidence. The modern catalogue identifies the temple as probably Hokke-ji and places the request in the circumstances of the early 760s. The retired empress Kōken was then living at that temple while Emperor Junnin occupied the palace. This is a reasoned identification based on historical and archaeological context, not a place-name plainly written on the wood.^[7]

Tsukuba no Myōbu can be placed more securely within the network of court service. The research institute's account links her with service to Kōken and with responsibility for food provision. Her appearance here gives the transaction a human connection: a request associated with an identifiable woman could connect a royal residence outside the palace with supplies administered within it.^[8]

The tablet does not preserve an ensuing conversation, a signature acknowledging receipt, or the delivery itself. What survives is one stage of the process. Yet that stage explains something important about the palace's reach. The boundary of the compound did not coincide neatly with the boundary of the household relationships it supported. A member of the imperial family living elsewhere could remain connected to its administrative resources.

The document also gives a woman a more specific historical role than that of an anonymous recipient of royal generosity. Tsukuba appears within the authority attached to a request. That is evidence of a particular position and relationship, not proof that court women in general had equal access to offices or resources. The strength of the example lies in its specificity: a named participant, a set of provisions, a written request, and a probable institutional destination.

Hokke-ji has its own history. Here its relevance is the connection recorded in palace rubbish. The request shows why a government household needed ways to recognize authorized demands coming from beyond its immediate walls.

Food was not the only thing that needed classification. Officials themselves generated records. The research institute's study of palace personnel tablets distinguishes material concerning a ministry's own staff from an earlier group associated with the central handling of assessments submitted by multiple agencies. This was administration at two levels: an office had to manage its people while also participating in a wider system for evaluating service.^[9]

A broken tablet found south of the central state-hall compounds preserves only a fragment of that system. Its catalogue interprets the opening characters as referring to entry into service in the previous year. Another surviving expression concerns the descendant of a family enjoying rank-related privilege. A few characters thus belong to a world in which service history and family status could both be relevant; they do not describe recruitment on an exclusively modern, merit-based model.^[10]

The problem of assessing an official differed from weighing a consignment. Service extended over time and could take place in an agency other than the one processing the assessment. To make it administratively usable, information had to be put into a form another office could recognize. The palace records reveal part of that conversion from a person's activity or standing into something that could be handled as a case.

The surviving archive is much too uneven to measure how consistently this worked. A fragment cannot tell us whether an assessment was just, and a central record cannot make every official's experience visible. It can show the effort to preserve distinctions that mattered to decisions: whose record was being considered, which period it concerned, and what category of service or status was involved.

These records bring a different meaning to the space between the audience hall and the office compounds. Officials were not only participants arranged for ceremonial occasions. They were also subjects of the government's own information-gathering. Keeping a court in operation required decisions about the people who performed its work, as well as about the materials they consumed.

Wooden documents can give the misleading impression that this was a government waiting for paper. Both materials were in use. The early publication of the palace tablets describes wood serving requests, receipts, labels, and other short communications. It also records large numbers of shavings left when written surfaces were cut back. A tablet could be altered or prepared for another use; the removed writing then entered the rubbish.^[5]

Paper has survived by a less obvious route. Two lacquer-permeated documents were recovered from the palace's southeastern ditch. Waste paper could be reused as a cover for a lacquer container, and contact with the lacquer could preserve writing that would otherwise disappear. These finds bring paper-based documentation into the same archaeological landscape as the wooden personnel records.^[11]

A writing surface therefore had more than one possible career. It might carry information, lose its immediate administrative value, and acquire another material use. For the historian, those careers can run in opposite directions: reuse that ended a document's official life might be the very event that saved it.

This does not establish a neat division in which every important text went on paper and every minor matter on wood. The surviving examples are evidence for particular uses, not a complete office manual. They do establish that the mass of excavated wood is an uneven sample of a mixed documentary system.

That unevenness is essential to reading Heijō. A large number of surviving labels does not mean labelling occupied most officials' time. Fragile sheets may be absent because of what happened after disposal, while a discarded shaving may preserve words never intended to last. What archaeology recovers depends partly on the government's choices and partly on the later lives of its waste.

The records also complicate the reconstruction of the palace itself. Research on the southeastern sector links different bodies of evidence to the Ministry of Personnel and, in a later phase, to the Council of Religious Affairs. Precisely when the change occurred has been debated. A modern plan that assigns one institution to a rectangle can therefore represent a phase or an interpretation, rather than an identity fixed throughout the palace's history.^[9]

Even a good match between texts and buildings needs care. The study of lacquer-permeated documents points out that waste paper might move before being discarded. A document's last location need not be the place where it was composed. Identifying an office requires the combined evidence of inscriptions, deposits, building phases, and associated objects, rather than a single institution's name lifted from a ditch.^[11]

Sometimes a very modest object can provide a firmer limit than an impressive building outline. A shipping label dated to the third month of 710 was recovered beneath the southern corridor of the First Great Audience Hall compound. The report uses the dated object to constrain the construction sequence: the work sealing it beneath the corridor cannot be earlier than the label. It does not prove that every part of the palace was unfinished when the capital moved.^[12]

This is a second kind of administrative afterlife. A date originally useful for identifying a shipment becomes evidence for the order in which a government complex was built. The historical value comes from the relationship between writing and buried position. Remove either, and the conclusion becomes weaker.

The palace emerging from these relationships is adjustable and layered. Its departments could change location; its structures could be rebuilt; its discarded documents could end up in deposits younger than their inscriptions. The apparently empty spaces around the reconstructions contain the evidence needed to keep those different moments apart.

The audience hall offers an architectural image of concentrated authority. The smaller finds show that exercising authority also involved distributing decisions. Someone had to identify a payment, recognize a request, record service, or retain information long enough for another person to act on it. No single one of those tasks amounted to governing the realm. Together they made parts of government repeatable beyond a single audience with the sovereign.

The objects discussed here do not form one continuous file. The bonito labels came from different decades; the food request preserves no surviving receipt; the personnel fragments concern other kinds of business. Joining them into an imagined journey of one parcel through the palace would erase precisely the distinctions that make them useful.

Their common contribution is more concrete. They show what the court needed to make identifiable: places and obligations, authorized demands, people and periods of service. The surviving categories are evidence of administrative capacity, but not a certificate of fair treatment or perfect control. They tell us how certain claims could be made legible inside the institution.

Heijō Palace can consequently be understood at two scales without sacrificing either. Its monumental buildings provided settings in which the court assembled and represented sovereignty. Its offices handled transactions that could proceed when the assembly had dispersed. Among the most informative remains of that second palace is a strip of wood small enough to accompany a quantity of fish—and precise enough to explain why the government believed it should receive it.