I made this list by combing through both services’ API documentation and extracting noun-like words and verb-like words from the names of resources or methods. Facebook count: 24 verbs, 43 nouns. Twitter count: 15 verbs, 24 nouns.
Here’s what I think these numbers mean.
Facebook manages many different kinds of content, and allows you to perform many different kinds of actions on that content, though the actions that you can perform on one kind of content are inconsistent with the actions you can perform on another. Twitter has fewer types of content, and a more consistent set of HTTP-like actions to perform on that content.
The Facebook API tends toward the baroque and insular, while the Twitter API tries its best to be a part of the web. In general, the Twitter API is much more straightforward.
I think this simplicity—this paucity of nouns and verbs—has been an important factor in Twitter’s widespread growth among both users and developers.
Developers can be confident that—even if Twitter’s API changes—they’ll still be doing mostly the same actions (getting, posting, updating) on mostly the same things (statuses, friendships, direct messages). Users know exactly how all the moving parts of Twitter work together, and are therefore better able to understand how a given application might augment that.
On Facebook, the opposite is true. User statuses, notifications, event invitations, feed stories, photos—they all have different interfaces and behave in different ways. Even long-time developers can’t keep track of how everything works together.
Interestingly, Facebook’s platform roadmap says that one of Facebook’s goals is to “focus [...] communication on the stream and Inbox,” which will make communication on Facebook feel much more like Twitter. I think this is a smart move—anything that simplifies the user’s model of how communication on Facebook works is a good idea, both for users and developers.