Hardship, end to end

A question in plain words becomes an address, the address becomes a rule, the rule names the facts it needs — and two of those facts turn out not to be the model's to supply.

We signed a five-year supply contract in 2023, priced in a currency that has since lost about two thirds of its value. Performing now costs us far more than we could have priced in. The contract is governed by the UNIDROIT Principles. Can we require the buyer to renegotiate?

The Principles do have a hardship regime (Chapter 6, Section 2), but I am not going to recite it from memory. Let me put your question to the model of the instrument and see which provisions it addresses.

Article 6.2.2 is the address. The candidate carries a ready-made next call, but before I supply any facts I want the rule itself — it will say exactly which facts it needs, so I do not have to guess.

Six premises, one of them negative. From what you told me I can supply four of them as facts of the case: the performance became more onerous, the currency collapse was beyond your control, it happened after conclusion, and you did not assume that risk. The remaining two — that the equilibrium was fundamentally altered, and that the events could not reasonably have been taken into account — I will deliberately leave out and see what the engine does with them.

There it is. The answer is not "no" — it is NEITHER, and the engine says why: the rule would derive hardship on these premises, four of them tick, and the two that are missing are marked as questions for the tribunal. That is not a gap in the model. Articles 6.2.2(b) and (c) are written as judgements, and the model refuses to make them on the tribunal's behalf — and on mine.

You told me the arbitral tribunal has already made both findings in its partial award. That changes the input, not the law: I can now supply them as facts of the case.

TRUE_ONLY — hardship is established, by article 6.2.2, on those six premises, against this exact version of the model (programHash). Whether you may then request renegotiation is article 6.2.3, and it is a separate question with its own premises — I would put it the same way rather than assume it follows.