Didn't you just describe an example where the ambiguity of human language is a problem and a smart contract may actually have a better chance of performing as expected?
Why do you asssume smart contracts can't be ambiguous? A lot of people read the DAO contract, and yet it took a hacker a while to find a small error -- if I recall correctly it was a capitalization error -- that triggered an unexpected exit path.
At some point you have to ask yourself whether we really prefer that code be law, with the quality of code that's so common these days, or whether we actually like being able to specify something to a lawyer -- rather than a computer -- simply because the lawyer will return the contract and ask for clarification if something is ambiguous/unclear.
The human-to-lawyer interface is the best contract interface that exists, the only reason we use computers is because they're so cheap and fast. When we dream of AI we dream of having a computerized lawyer, who can ask clarifying questions and resolve ambiguity before it becomes a problem.
If tomorrow I sign a contract with my mobile phone carrier and it turns out that through some loophole in the contract they get entitled to the kidneys of my first born daughter then clearly I have a case to go to a tribunal and get it overturned as it's obviously not a reasonable clause.
I think people arguing that "code is law" (which is simply a modern form of "letter-ism") don't really know what they wish for. I guess is many are going to change their minds when those attacks get more and more common and they lose a ton of money because of an unforeseen and obviously unintended flaw in the contract code.
I feel like anything humans can use to communicate to each other could be considered a human language.
More relevantly, creating a synthetic language that's a subset of natural languages will mean that multiple natural concepts will map onto the same synthetic one.
