I believe they do stack, but that the duration is calculated twice, once for the pommel, then they take that number and use it for the calculation for the aura. Or something to that effect. Not just one base calculation of +22%, but two calculations, one of +12% and then one of +10% of that new duration number (as examples).
I'll test this fully tonight so I can be positive... there's a lot of conjecture I feel unsafe with on my part.
Edit: Tested it out, here's what I found
Code:
Enchantment: Vigorous Spirit
Base Duration: 30 seconds
Modifiers:
-(Level 0 Divine Favor) Blessed Aura +10%
-Staff Wrapping of enchanting +15%
-------------
Possibilities
-------------
Predicted total combined enchantment duration +25% -- 37.5 seconds
Predicted duration with +15% of +10% of the original duration -- 37.95 seconds
Predicted duration with +15% of +10% of the original duration (truncated) -- 37 seconds
Predicted duration with +10% of +15% of the original duration -- 37.95 seconds
Predicted duration with +10% of +15% of the original duration (truncated) -- 37 seconds
Predicted duration with +10% -- 33 seconds
Predicted duration with +15% -- 34.5 seconds
-----
Chart
-----
| +25% | +15% of +10% | Trunc. +15% of +10% | +10% of +15% | Trunc. +15% of +10% |
Predicted Duration |37.5s | 37.95s | 37s | 37.95s | 37s |
------
Trials
------
Trial | Actual Duration |
1 | 38.69s |
2 | 38.69s |
3 | 38.62s |
4 | 38.62s |
5 | 38.75s |
6 | 38.44s |
7 | 38.29s |
8 | 38.44s |
9 | 38.56s |
10 | 38.29s |
Assuming there were no other factors affecting the duration, and that I have about a ±.5 second error (using my amateur methods of finding error... didn't want to statistically find it), I think that the following, given the options accounted for in the chart, can be concluded:
-Enchantment enchancers DO stack
-Enchantment durations DO NOT truncate
-Stacked enchantment durations are calculated as a percentage of a percentage of the base, not as a sum percentage
-Priority doesn't matter, as 10% of 15% of a base is the same as 15% of 10% of that base
Of course, there was only a 1 second descrepancy max that I had to work with in the possible outcomes, so this may or may not be the exact calculation used in finding the durations. This isn't an exact science but is the closest thing I can come without having hard data from the devs themselves.
If there's anything I missed, please let me know~