I think that Gulfcapt 'nailed it' best. My company's policy, which closely follows Boeing, puts Check Lists into 3 categories -
(1) Normal Routine procedures (Example pre-flight preparation). These are accomplished from memory, and verified as having been actioned from a step by step Check List upon completion. The memory items, such as cockpit set-up are usually well aided by a good company provided flow chart / panel scans etc.
(2) Emergency Procedures (Example Engine Fire). These contain some memory items because they MUST be actioned immediately, and a number of non-urgent follow-up items. The memory items are accomplished from memory, and then the check list is used, item by item from the beginning, first to verify that the memory items have been accomplished, and then as a 'Read and Do' sequence to accomplish the follow-up non-urgent items.
(3) Non-Normal / Abnormal Procedures (Example Generator Failure). As non-urgent items, there is no requirement for the pilot to have memorised the particular items, and the Check List is used as a 'Read and Do' sequence to accomplish the entire procedure (Still a good idea for the pilot to have, in advance, a good general knowledge of what the Check List will achieve, and it's operational implications).
As a (sadly) historical record, Douglas separated Urgent from Non-Urgent procedures by using a 10 minute criteria. If no significant harm could befall the aircraft by inaction within 10 minutes, then it was a Non-Urgent procedure. Anything requiring action in 10 minutes or less was an Urgent procedure.
Regards,
Old Smokey