This actually turned out to be quite a fun bit of logical deduction. You need some good source material though. I started with transcriptions of the Morgan County, KY censuses from 1830 through 1880, the Morgan Co. marriage license book and the marriage register. Once I had exhausted all of those possibilities I branched out to neighboring counties' marriage records and censuses where I had them available. Of course I had a small beginning in that Harley Little had provided some from his own research and I had some more from my own research, but this method should work even if you're starting out cold.
Since it was the 1850 Morgan County census I was concentrating on here is how I proceeded. I'd look at a household listing and the age of the man and wife (assuming both were listed) and the age of the oldest child listed. That gives you a rough estimate of when the marriage might have taken place (unless the age gap between the parents and the oldest child listed is such that there were likely some older children who have already married and left the nest). The marriage license book is alphabetized by the first letter of the groom's surname only, with the listing within each letter in chronological order. Say the oldest child listed in 1850 was 7, you would look in and prior to 1843, working your way backward to where you either found a marriage or else you found a point at which the couple would be absolutely too young to be considered. (And, in the hills, that age is mighty young!) If you get an exact name match you can figure you've found the marriage. If you don't get an exact match but you get what appear to be standard nicknames for the given names in question you can pretty safely assume that that is the one you're after. If you can't assure yourself that you've found it with those steps you then look at subsequent censuses to see if the names were given in a different form. You can also look in the marriage register and try to find the corresponding entry you've located in the marriage license book. In any event, if you can find enough sources that give a consistent name, then it is likely safe to assume that that is the correct name.
If you exhaust the resources of the county you're researching you can look in neighboring counties to see if the marriage might've taken place in another county. It is a quite fun procedure and, after you get a feel for it, quite a productive exercise as well. I figure that I quadrupled or quintupled the number of maiden names that I started with.