There are a number of good exhibits available on the NTSB website for this accident.
This link included excerpts from the AA A300-600 Operations Manual, both before and after the accident, indicating how to use the rudder:
AA Operations Manual
This link covers the FDR and simulator sessions showing exactly what happened in the final seconds of flight AA587. Page 19 particularly shows what happened to the rudder and the ailerons, when the pilot attempted to coordinate them.
NTSB FDR and sim analysis with graphs
The AA Operations manual discusses using the rudder in coordination with the ailerons to assist with rolls during certain upset recovery maneuvers. Wino mentioned that this was in the training course. The accident data shows that the ailerons were sometimes at their limits as the pilot tried to roll the aircraft level during the second wake encounter.
However its seems to me that the mere mention of using the rudder to coordinate with ailerons to effect roll maneuvers in both the AA manuals and the training course, assumes that the pilot MUST have the ability to control the amount of rudder (and side force) being applied. The rudder limiter of the A300-600 (and its all-or-nothing design flaws) basically prevents the pilot from controlling the amount of side force generated at high airspeeds, therefore the ability to affect good coordination between the ailerons and the rudder is compromised. To me, this is clearly shown in the FDR and sim analysis in the above link. I don't think that this accident was caused by a training problem, because the AA training program assumed that the rudder system operated correctly. Clearly after the accident, the training program had to be modified to accommodate the limitations of this particular rudder system on this particular aircraft.
I personally think that a really good lawyer should be able to tag Airbus for this one, without too much trouble.
(edited to fix a link)