We are moving to a more Lean based approach, Kanban the lot. We started breaking down our user stories into very small ones, however a lot of questions are being raised as to what stories should have acceptance tests and what shouldn't.
For example we have split one story into 2. The first creates an entity and some basic fields. The other expands on those fields and adds more functionality. Normally we would write acceptance tests at the end of the second story. However should we be writing a small acceptance test now we have a smaller story and then expand on the acceptance test(s) when we do the second story?
The argument for it is, that all stories should be testable, the counter argument is its anti-lean because we are splitting the stories to get them to the testers quicker, but by writing acceptance tests (on tiny logic and tests that are going to change) slows things down to them anyway. What is the consensus with Lean on how to approach acceptance testing (or BDD testing) with Lean user stories?