Was しもうた ever a part of the standard japanese dialect? Do 昔話 borrow from the other current dialects to create their atmosphere or do they generally only use older Japanese forms?
It's hard to generalize about the entire genre, of course, but I think a combination of all of your theories is true. Modern "mukashibanashi" (as opposed to period novels 時代小説 and so on) are basically written for kids. The dilemma there is that you need to create the "mukashi" atmosphere, but you can't use actual premodern Japanese, because kids won't understand it (most adults wouldn't either). The solution is to use standard modern Japanese with a few well-known quirks thrown in.
For example, I think most kids learn that "なり = old-fashioned です" at a pretty young age, and so you can have your samurai saying "○○なり" and everyone will understand (a) what he means, and (b) that he is in the past.
The -うた ending is another sort of example, but it is a bit more complicated because, unlike なり, it is preserved in living dialects as you say. (In fact it is even preserved in standard Japanese in some cases: the commonly accepted past tense of 問う is 問うた, not 問った. Exactly the same phenomenon.)
Specifically, what we are looking at here is a way of forming the past tense (i.e. the -た ending; I will use the term "past tense" for reasons of simplicity) by adding -た to the dictionary form (that is, attributive + final form) of verbs that end in -ふ/-う (in the time period we are talking about, this is the same thing, so I'll skip the ふ thing in the explanation below). This is combined with the well-known sound change from /au/ to /o:/, so that even though they are written "...あうた" or, modernized, "...おうた", they are pronounced "おーた".
- 問う → 問うた
- 合う → 合うた
- 願う → 願うた
- しまう → しまうた = しもうた
This is, IIRC, still the standard way of forming the past tense for verbs of this type in Osaka dialect at least (not sure about Kansai dialect as a whole).
As I understand it, it is a direct descendant via sound change from the -ひた- form in classical Japanese, as seen in this example from the Tale of Genji:
- 弾きもの、琵琶、和琴ばかり、笛ども上手の限りして、折に 合ひたる 調子吹き立つるほど
But, the standard -った ending for -う verbs also descended from this -ひた-. So the same sound evolved in two different ways. Some speech communities, like the Osaka dialect, came to prefer the -うた version, while others ultimately came to prefer -った, and there was probably a lot of in-betweening and overlap (not to mention influence from the other sounds in the verb, interaction between speech communities, the varying de facto definitions of "standard Japanese", etc.) I don't know all those details.
But we can see the -ふた/-うた form in Tokyo/Edo Japanese as recently as, say, Higuchi Ichiyo's Nigorie:
... and, as I noted above, the past tense of 問う in standard Japanese is 問うた, not 問った. Same goes for 請う/請うて, etc. So clearly it is not solely a Kansai or even "dialect" phenomenon.
To summarize, from the point of view of a Tokyo-based Standard Japanese speaker, the -うた ending is seen in the following contexts:
- Dialects -- notably but not limited to Kansai dialects
- "The olden days"
... and to an extent, these overlap in people's minds. Everyone knows that people spoke differently in Edo than they do in Tokyo, but many people have a sort of unspoken assumption that the other dialects are more or less unchanged since the olden days.