Not sure why I'm fascinated by this study . Maybe it's the 98% slot-filling rate. Maybe it's the notion that out-of-grammar responses are 7 times more likely for types of pain than for body parts (yeah, think about that...) Maybe because it's a well-written
