In Japan, enormous vessels were used for everything from cooking to burial — scientists have uncovered their history
In Edo-period Japan, a large ceramic vessel could accompany a person through almost every stage of life – and even after death. Such vessels were used to store water and soya sauce, to transport fertiliser, to collect human waste, and some were even used as burial urns.
Scientists have now analysed the shape of 243 such vessels and identified two clear patterns. From the 17th to the 18th century, they became increasingly uniform, whilst examples from burials differed markedly from those used for domestic purposes. To identify these changes, the researcher combined mathematical analysis of the shapes with machine learning.
A huge pot could be used for just about anything
We are talking about ogame – large stoneware vessels produced in the Hizen region of Kyushu during the Edo period, which lasted from 1603 to 1868.
It would be difficult to describe their use as highly specialised.
One such vessel could be used to store drinking water or soya sauce. Others were used for agricultural fertilisers and for collecting human excrement, which was used as fertiliser in Japan.
But the uses of ogame did not end there. Some large vessels were transformed into kan—funerary vessels in which human remains were placed. Thus, a single type of pottery served both the most ordinary domestic needs and the rites of death.
It was precisely this diversity that caught the interest of archaeologist James Francis Loftus from the Institute of Science Tokyo.
The scholar digitised the shapes of 243 vessels
Loftus collected data on 243 ogame from 11 archaeological sites in the modern-day prefectures of Saga and Fukuoka.
Two sites were associated with pottery kilns, and nine with cemeteries. Instead of the usual visual comparison, the researcher converted the contours of each vessel into a set of numerical data.
To do this, he used geometric morphometrics—a method of quantitative shape analysis. In particular, the contours of the vessels were described using elliptic Fourier analysis, which allows small differences in the curvature of the walls, shoulders, neck and base to be mathematically recorded.
The data were then processed using the Random Forest algorithm – a machine learning method.
Here, too, the AI performed a fairly specific task: it helped to classify poorly dated vessels into presumed chronological groups based on similarities in their shape. This is not a case of ‘AI unravelling the mystery of the past on its own’—the research remained grounded in archaeological data and quantitative analysis of form.
In the 17th century, the vessels were diverse; subsequently, they began to become uniform
The analysis revealed a distinct trend.
At the beginning of the 17th century, the shape of the ogame was quite varied. Over time, the variation decreased, and by the 18th century, the vessels had become increasingly standardised.
Statistical analysis confirmed a decrease in shape variability throughout the Edo period. Individual elements of the vessels—their shoulders, necks and rims—became particularly standardised.
This change coincided with major transformations in the Japanese economy and craft production.
During this period, the Tokugawa system of government was consolidated, domestic production networks expanded, and mass production of ceramics and porcelain grew. Trade was also influenced by events outside Japan — in particular, the Qing Empire’s maritime restrictions in the mid-17th century.
The authors suggest that tighter control over production, the specialisation of workshops and the development of domestic markets may have gradually led to the standardisation of everyday pottery.
However, an important caveat is in order: the study demonstrates a correlation between morphological changes and historical processes, rather than proving that, for example, a specific political decree directly compelled potters to alter the shape of their vessels.
Even the official Japanese press release formulates its conclusion cautiously: historical events may have left their mark even on the shape of everyday objects.
Vessels for the dead looked different
The second pattern proved no less interesting.
The ogame found near the pottery kilns typically had broad shoulders, a flared mouth and a relatively compact base. This design is well-suited to the storage and transport of liquids or fertilisers.
Vessels from burials, however, looked different.
They were taller and narrower, with a more closed neck. Quantitative analysis revealed a distinct morphological distinction between the two groups.
This is important for understanding how ogame came to be used as funerary vessels.
A simple explanation might go like this: after a person’s death, their relatives would take an ordinary household vessel and use it as a coffin.
New data suggest that the situation was more complex. Some vessels may have been specifically manufactured for burial or selected in advance for their suitable shape, rather than simply being transferred from the kitchen to the cemetery by chance.
However, it is impossible to determine which of the two scenarios prevailed — specialised production or the selection of suitable household vessels — based on shape alone.
The algorithm had to work with just six reliably dated samples
The study has a significant limitation.
For many ogame, there is no precise date of manufacture. This is precisely why the author turned to machine learning — to use the shape of the vessels as an additional clue for their chronological classification.
However, there turned out to be only six reliably dated reference specimens. These date from the period 1629–1800.
It was these six vessels that were used as reference points, after which the Random Forest algorithm categorised the other finds into six chronological stages based on similarities in shape.
Therefore, the result should not be interpreted as meaning that artificial intelligence has ‘accurately determined the age’ of hundreds of archaeological finds.
This is a model-based classification based on a relatively small chronological database. The Institute of Science Tokyo explicitly states that an increase in the number of reliably dated vessels in the future should improve the accuracy of the method.
The author himself describes the work as exploratory research—a first quantitative step towards a more systematic analysis of this ceramic tradition.
Historical sources from the Edo period provide detailed accounts of the shoguns, trade, political reforms and major economic events. Yet the everyday lives of ordinary people feature far less frequently in written history.
Ceramics can help to fill in this picture.
If production gradually shifted from a wide variety of forms to more standardised vessels, this may reflect the specialisation of workshops, changes in the organisation of labour, the expansion of markets and new consumer demands.
And the differences between domestic and funerary ogame show that standardisation did not at all mean the production of a single, universal pot for all occasions.
On the contrary: vessels were becoming both more uniform within their own categories and more specialised depending on their intended use.
In future, Loftus hopes to combine form analysis with other methods — such as portable X-ray fluorescence spectrometry and the study of the chemical composition of the pottery. This will make it possible to examine not only the shape of the vessels, but also the origin of the raw materials and the organisation of production.
Source
Author: James Frances Loftus, Institute of Science Tokyo.
Journal: Open Archaeology, 2026. The article was published on 19 August 2026.