The reason is that as in the above 8/16 table, the ordering of bytes, which
become synonymous with elements, become inverted.
+
+The "correct" fix to this problem is to view arithmetic LE/BE-reversing
+(reading/writing to/from the regfile on the way into and out of
+Arithmetic Operations) as **completely separate and distinct** from
+**Memory** LE/BE reordering, which a quick check on an IBM POWER8/9/10
+system using VSX arithmetic operations on 8/16 in BE order will confirm
+is already considered.
+
+The "correct" place to add such reversing would be the `REMAP`
+subsystem but the number of available bits is already under
+pressure so some care is needed. Even better would be the 24-bit
+RM but that is also precious and under pressure.