From a Keirseyian perspective, SJs aren't necessarily about tradition per se, it's more that they see time-tested methods as, you know, tested and reliable. Hence, they tend to trust them more and are suspicious of abandoning them for newfangledness.
NTs see everything as needing improvement, and have little or no respect for time as a crucible for "proving" an idea/method/whatever. They typically assume that if it's old, people probably haven't been thinking hard enough on how to improve it.
The idea that SJs blindly hold to convention and tradition is a bit of a myth. They'll throw it out if it's obviously inefficient like any reasonable person would. It's more that they have an attitude of trusting what's been established by long efforts and consistant, reliable use than just throwing that all out for something that should be theoretically more efficient. It's subtle, but different than blindly trusting tradition, for sure. Sort of the whole "If it ain't broke, don't fix it" saying. But if it is broke, any intelligent person will fix it. NTs think more that if it ain't broke, it doesn't have enough features yet (think of Windows Vista).
Oh, and CPs mean cognitive processes. Ni, Ti, Se -- all that.