Thanks for playing!
That’s a good point about speech queuing up when using a controller. I mentioned this specifically in a discussion on the Discord server about the difference between aria-live="polite" and aria-live="assertive". This is an example where I made the wrong choice and will be fixing it in the first post-jam update. (I used polite but should use assertive.)
The reason it queues up like this with a controller is because, when using a screen reader, pressing keys will typically interrupt the speech, but the gamepad doesn’t have a similar mechanism.