Towards An International Court of Smart Contract Arbitration

Firms involved in international commerce routinely contract that disputes are to be resolved by private courts of arbitration such as the International Court of Arbitration, the London Court of International Arbitration or the Singapore International Arbitration Center. These courts of arbitration compete for clients and thus have an incentive to resolve disputes fairly, quickly and inexpensively. Courts compete, for example, to provide arbiters who are experts not simply in the law but in the relevant area of commerce. The New York Convention of 1958 says that private arbitration decisions will be enforced by the national courts of any of the 159 signatories; thus private arbitration leverages national enforcement but is otherwise not tethered to national law (e.g. in US see, Mitsubishi v. Soler Chrysler, National Oil v. Libyan Sun). Over time private courts of international arbitration have developed a system of law that transcends nations, an anational law–this is the new lex mercatoria.

I propose that courts analogous to the courts of arbitration that govern international commerce be created to govern smart contracts in virtual space. Arbitration of smart contracts will develop a new private law that will evolve to meet the needs of virtual commerce, a true lex cryptographia. At first, it might seem contradictory to advocate for courts of smart contracts and the development of lex cryptographia. Isn’t the whole point of smart contracts that no courts or lawyers are needed? Similarly, lex cryptographia is usually understood to refer to the smart contracts themselves–code is law–rather than to law governing such contracts. In fact, it is neither desirable nor possible to divorce smart contracts from law.

Smart contracts execute automatically but only simple contracts such as those involving escrow are really self-enforcing. Most contracts, smart or dumb, involve touchstones with the real world. Canonical examples such as the smart contract that lets you use an automobile so long as the rent has been paid illustrate the potential for disputes. Bugs in the code? Disputes over the quality of the car? What happens when a data feed is disputed or internet service is disrupted? Smart contracts applied to the real world are a kind of digital rights management with all of DRMs problems and annoyances.

Some of these problems can be dealt with online using decentralized mechanisms. But we don’t yet know which decentralized mechanisms are robust or cost-effective. Moreover, when marveling at the wisdom of crowds we should not forget the wisdom of experts. Nick Szabo once remarked that if contract law was suddenly forgotten it would take hundreds of years to recover the embedded wisdom. Contract law, for example, is filled with concepts like mistake, misrepresentation, duress, negligence and intention that are not easily formalized in code. Contract law is a human enterprise. And the humans who write contracts want law with terms like negligence precisely because these terms fill in for gaps which cannot be filled in and formalized in contracts let alone in code.

I am enthusiastic about smart contracts on blockchains. Smart contracts will significantly reduce transaction costs and thus let people create valuable, new private orderings. But it will be more profitable to integrate law and code than to try to replace law with code. Integration will require new ways of thinking. The natural language version of a contract–what the parties intend to agree to–may not map precisely to the coded version. Arbiters will be called in to adjudicate and thus will have to be experts in code as well as in law. Smart contracts can be made by anonymous parties who may want a dispute resolved not just privately but anonymously. Smart contracts can be designed with escrow and multisignatory authority so arbiters will also become decision enforcers. All of these issues and many more will have to be understood and new procedures and understandings developed. The competitive market process will discover novel uses for smart contracts and the competitive market process among arbiters will discover novel law. Law will adjust to business practice and business practice to law.

In short, the best way to create a vital new lex cryptographia is through competitive, private arbitration built on the model that already governs international commerce.


What Tabarrok calls a "smart contract" is actually a "dumb contract". What? A smart contract is one drawn for a specific transaction between specific parties; a dumb contract is one drawn for an unspecific transaction between unspecific parties. As I mentioned in my comment to Cowen's blog post this morning, in civil law countries the codes fill in the gaps; thus, contracts in civil law countries are simpler and shorter than in common law countries. But that produces results that many would find unacceptable: almost every instance is unique, and generalized principles may not fit very well. To make the generalized principles (i.e., the codes) acceptable would require them to be vast in length, covering every possible instance, and written by experts with a vast amount of experience and knowledge. Is that possible? Is that even desirable? Dumb contracts are dumb, indeed.

A smart contract is one drawn for a specific transaction between specific parties. I work on complex transactions, and write long and detailed contracts for them. Many of the provisions I describe as "litigation avoidance" provisions: they produce the result that is best for my client if the unexpected occurs. The more experience I have, the longer my contracts become because the more unexpected events I have seen. Years ago I wrote a long contract for the purchase and sell of some business or asset. I met with the lawyer for the other party to review and negotiate the contract. As he sat across the table from me flipping through the pages of the contract, he asked why the contract had to be so long. I answered that many of the provisions are there for litigation avoidance. After flipping through more pages, he looked up at me and asked: "Why would we do that?" A lawyer, he preferred a dumb contract that is more likely to produce a conflict that required litigation to resolve.

Agreed. If you want to learn about this, go to WestLegalEdCenter and take a course on legal aspects of blockchain technology and smart contracting.

Alex should learn more about the subject before he writes. The Chicago Bar Association had a presentation you can see on WestLegalEd Center last week entitle: Breaking Down Blockchain: The Possibilities and Pitfalls. Took it online for CLE requirements.

If you think of a smart contract as a letter of credit, which pays on the presentment of documents, you will quickly realize that contingencies and discrepancies are part of many commercial transactions, which necessitate modifications or exceptions and even interpretations.

Some aspects of contracts can be smart--what is the CPI as of a certain date, etc.--and executable.

The other problem I have with Alex's proposal is that we have a large body of law, including precedent, for arbitral proceeding and interpret ion of contracts which need not be replicated with another set of rules.

What is interesting is that contracting parties often try to maintain a "reputation" so that you can take them at their word, or recognize that if there are gaps in interpretation, one or both parties, eager to maintain the relationship, will accommodate or interpret an ambiguous clause to mirror the parties intentions. In other words, reputation makes contracting easier. But, if contracting parties are unknown or secret, then there is less room for flexibility because there is no benefit.

This idea needs to wait until smart contracts can actually accomplish something useful that might need to be arbitrated.

I don't think that a judge in the Hague is going to want to hear that a Solidity bug made his cryptokitty the wrong color.

It is now and has been litigation in crypto contracts. The litigations we are seeing look a lot like the instruction sheet on how to define a crypto security rather than a crypto coin. In other words, geeks learning to count past two, and they are figuring it out.

In the short term, I think smart contracts primarily serve as an enforcement mechanism. You can think of all the other things that go into a lawful contract today--negotiation, mutual assent, consideration--as happening off the blockchain, in meatspace, and the smart contract really just steps in to provide an automatic transfer of property or rights when a given on-chain condition is triggered.

From that idea, maybe a neat medium-term solution could use automatic execution coupled with 1. a code-based dispute mechanism and 2. reputation ramifications. So if the smart contract executes in a way that the parties did not intend, the property or rights are still transferred automatically but the aggrieved party gets to log (on the blockchain) a complaint. An arbitration tribunal adjudicates the claim like in today's off-chain world, and maybe cryptographically endorses or doesn't endorse the smart contract's execution. On the blockchain you'd be able to keep score of everyone's behavior in these circumstances, like a vendor score on Amazon, so that in entering new deals you would see whether the person you're dealing with abuses the dispute mechanism or ignores the arbitration outcome, comparing their stats with the stats of other similarly situated players in the industry.

That strikes me as a setup that manages incentives pretty well and preserves the bonus of automatic execution.

Poe's Law? I've read this several times and I still can't work out whether it's intended to be genuine or a Swiftian "Modest Proposal".

Blockchains are brilliant! Smart contracts are brilliant! And the only things we need to make them work are hundreds of years of embedded contract laws and a system of international courts filled with arbiters who are experts in coding as well as law. Simple!

This is exactly what Mattereum and the Internet of Agreements is up to - have a look.

"A paper outlining how Mattereum implements Ricardian contracts—​smart contracts with the legal force of natural language contracts—​in a way that integrates smoothly with existing arbitral and legal frameworks."

That's just BS.

That's a really great point about DRM. Smart contracts and DRM both require giving up control over the machines we own, and instead giving our machines control over us. Paging Cory Doctorow...

I'll add that converting fuzzy human desires into code is the essence of programming. The only reason this might seem easy is that bugs are often fixable by pushing out another patch. Although we try to avoid mistakes, this is software that remains under the control of humans, for good reason.

In domains where patching it later is unacceptably expensive (as in aviation or chip design), writing software requires much more care than ordinary programming. This is very slow and expensive.

Everything in the world is speeding up, except government. That is a fatal flaw. The only move is to free more and more things from government. This appears t be an example of that. Go for it.

I would argue that the very lack of unsettled substantive law militates against the use of a novel forum. Indeed, the idea that a private forum -- without appeals or stare decisis -- could unilaterally determine what the law should be, as opposed to what it is, seems an invitation to a suboptimal result. Arbitration works best in fields where the substantive law is highly settled -- e.g., securities, commercial paper, letters of credit, diamonds -- and where specialized knowledge or ability to inquire into *facts* can help to expedite the determination of disputes using settled law.

Sorry, should have written "lack of *settled* substantive law."

This is along the lines of the digital jurisdiction and court system we're working on at Aragon. We just released a new version of our whitepaper describing how it works, would love your feedback if you get a chance to check it out:

Comments for this post are closed