I won't go through it in detail as I am posting from my phone, but scenarios where fort legal status is relevant include:
Prison colonies
Bandit camps
Separatist groups
Civ sponsored strategic military outpost
Frontier colonies with only loose ties to parent civ
Claiming a site with special magical properties
The possibilities are very broad.
If you don't want to be subject to restrictions, you presumably could play a standard fort.
There is no such thing as an independent colony (three of the above with I have bolded do not therefore exist as genuine starting scenarios). That means that realistically legal status of the fort is pretty much always going to be the same, which is dependent upon whatever created them in the first place. The historical trajectory is always or virtually always that the colony starts off highly dependent and with little autonomy, then it obtains greater amounts of autonomy as it becomes more developed, either peacefully (Canada) or violently (USA). There is a reason why that is, any colony is in competition with it's own homeland for people and the accumulated capital of centuries or millennia cannot be replicated by a handful of people starting from nothing.
The present DF then is built on a lie. The situation works because the surplus value produced by dwarves with minimal or no capital investment is so staggering (surplus value is the difference between the value of what dwarves produce and the value of what they consume) that the fantasy of the independant colony is a possibility. Realistically the amount of the surplus value produced by the colony is too small to independently allow anything but an extremely primitive lifestyle and production for centuries, which means the embark goods are absolutely critical to the situation rather than being optional as they presently are. While situation works fine at the moment, it is not compatible with the Economy actually existing, unless what we have in mind is post-scarcity.
The restrictions should be based upon this issue, in order to create a viable colony with any degree of long-term prospect somebody has to foot the bill. That somebody has to have a reason to do so, the starting scenario is the reason why the colonists were given the goods needed to set up a site. The starting scenario represents a situation where what the colonizers had in mind was something very specific in the sense of a return on their investment, they only accidentally ended up creating an actual settlement with independent viability; the initial restrictions derive from the purpose the accidental colonizers had in mind.
A standard fort on the other hand represents a situation where the civilization rulers actually decided to flat-out build a new fortress. In that situation the conditions are that you remain part of their civilization and ultimately contribute in whatever way an ordinary settlement would contribute in your situation. Independence should still be an option, but it should not be a starting scenario but something declared by a settlement advanced enough that it's declaration is taken seriously, it is an end-game scenario in effect not a starting scenario.
As for different legal castes - yes, it is unnecessary. But you could say that of any feature. it would add flavour to the game and would create a useful framework for modders etc. And, as always, if you don't want serfs or whatever you could turn them off in world gen.
You don't know what a can of worms you just opened there.
You see the amount of working hours in Toady One's life so happens to be a finite resource. Adding serfs in for no reason or purpose is an example of wasting that finite resource, since nothing is really added to the game by making it more oppressive, that we can opt out of the system does not change the fact that Toady One hours were wasted adding it in to the game, hours that could actually be used developing worthwhile elements. But different legal castes *are* needed, because we need to be able to distinguish in prison colonies (and forts more genuinely) between prisoners and free people; my point was solely that we did not need starting scenarios for that.
Provided the framework is flexible enough, it can be extended to a wide range of other options beyond prisoners. Really what we need for prisoners is occupational restrictions (prisoners cannot be hunters), low-priority for goods distribution (aka poverty), mobility restrictions (obviously) and lack of certain legal rights and a name that identifies them as prisoners of a certain type. As long as the above things are defined loosely enough, it becomes possible to add your serfs into the game simply by reusing the tokens that are used for prisoners in a different context.