Cinnamon, Corvée and the Translation Trap: Three Sources on Kandyan Labour

Principle: reliability is not a property of a source but of a source-question pair, so the ranking must change with the claim being tested.

• VOC Colombo resolution. Generated by the party that set the quota, for internal discipline and audit, addressed upward to the Governor-General and the Company's directors. It is strong evidence that quotas existed, that they were seasonal, and that shortfalls were a live administrative problem. It is weak evidence for how peelers understood the work: desertion is recorded as a fact, its cause is not.
• Temple palm-leaf record. Its purpose is to fix the temple's rights and its tenants' obligations, and such records are typically copied and recopied. It is strong evidence that cinnamon peeling was one item in a recognised schedule of services attached to land. It tells us about categories and legitimacy, not about quantities or pay.
• 1968 oral narrative. A 170-year gap and a recording context shaped by the land politics of the late 1960s make it good evidence for family memory and for how coercion was remembered, and poor evidence for the 1760s.
• The translation is itself evidence. Rajakariya denotes an obligation attached to land tenure; rendering it 'forced labour' imports the very conclusion she is testing. The Dutch 'corvée' does the same work in the opposite direction.

The contradiction is largely an artefact of standpoint, and reconciling it is more productive than adjudicating it.

• Same institution, two positions. From the temple's side, service is lawful because it is attached to land held under a tenure that carries obligations; it is bounded, known and inheritable. From the Company's side, it is a deliverable with a deadline, a price and a penalty.
• Coercion can increase without the category changing. Quotas, cash commutation, punishments for desertion and pressure on landholders would all raise the burden while rajakariya remains, in local terms, a duty rather than a seizure. Both sources can therefore be accurate.
• What would actually settle the question. Whether service fell on tenants or on the landholder; whether it was commuted for money and at what rate; whether deserters were punished and how; how long the peeling season ran; whether the same families appear across seasons.
• Read each source's silence. The temple record is silent on wages and on punishment; the VOC record is silent on consent and on the legitimacy of the obligation. Those silences are findings, not gaps to be filled by assumption.

Each proposed source must be tied to a specific claim, or the list becomes window-dressing.

• Dutch thombo (land) registers would show which holdings were assessed for cinnamon and in whose name, establishing whether the obligation ran with land or with persons.
• Records of the Landraad and of the local courts could show prosecutions of deserters or disputes over service, establishing whether coercion was routine and judicially enforced.
• Cinnamon warehouse and depot accounts would give quantities delivered per season and any payments made, testing whether the trade was monetised.
• Kandyan land grants and temple documents of an earlier period would show whether the service predates the Dutch, which decides whether the Company created the obligation or inherited and intensified it.
• A fresh translation of rajakariya passages by a Sinhala-reading historian, checked against the Dutch, would remove the glossary problem at its root.

A negative result is still a result: if no court records of desertion survive while the Council complains of mass desertion, the enforcement gap becomes the argument.