Question about the CAP theorem

I've been reading about the CAP theorem, and I came across this article, with a snarky comment at the end (made by Francois), which got me thinking.

Is it possible to phrase the CAP theorem as follows: in the case of no network partition, everything is fine and you have both availability and consistency. In the presence of a network partition, you either:

1. bring down the entire system, losing availability but keeping it consistent in its offline state (obviously because there won't be any writes during downtime). Or perhaps keep the system online but in read-only mode if that's possible?

2. keep the system online (by which I mean it continues to accept write requests), but sacrifice consistency until the network partition is resolved. Once the partition is resolved you need to have some sort of heuristic to disseminate the information across partitions (if there were more than a single one online) such that you reach a steady state.