. The main topic of balance is the balance of power among all Civs in a given scenario. Having minor civs that can colonize, or that can't colonize but are as powerful as a full Civ can drastically alter the rate and manner at which full Civs expand. It will also mean greater use from alliances with minor civs. Imagine allying with every minor civ on a map? Making a United Federation of Planets all your own, gaining access to possibly unique technologies and such.
. A secondary topic is the concept that a race can be equal but weaker. To take your example of the Arceans, while it might be a fun challenge to play as them and try to win, what's the point of having them on a map? If the "turtle" strategy of their AI makes their ultimate defeat a forgone conclusion, what's the point of having them?
. With minor civs, this is fine...they're minor. Full Civs though should always have a roughly equal chance of winning, all things being equal. The Arceans shouldn't just turtle, they should be like an inexorable wave of conquest taking one world at a time like an unstoppable glacial wall. It might take a long time for them to conquer the whole map, but good luck standing in their way. Obviously they don't currently work that way, but I hope you see my point.
. Finally, with all the other ways of adjusting the game's difficulty, creating imbalanced Civs that exist solely as irritating cannon fodder meant to hamper your progress and nothing more...well, it's just excessively complex and not really very interesting. Again, the minor civs are already intended to fulfill this role.
. Now, it might be reasonable to have full Civs that are more complicated to manage and difficult to play correctly with more unconventional concepts behind their racial strategies, such that they only shine as AI Civs when the game's difficulty and such are turned up. But aside from that, having weaker full Civs is really both unnecessary and doesn't add enough to the game compared to what it takes away.
. It could be possible to mod new full Civs into the game that are as handicapped as a minor civ purely for the challenge of it, but that sounds to me like a failing of the original game to provide sufficient challenge on its own. After all, mathematical exploitations, and actual tactical and strategic thinking have almost nothing in common. Most tactical and strategy games employ little if any actual tactics or strategy in the first place because games simply aren't yet capable of processing that many variables of such broad ranges. Whether that inability is a failure on the part of programmers and designers, or of contemporary technology is a matter of debate at the moment.
. In closing, I think every game should have major and minor elements, and that each should stick to its role in a game. Real life doesn't care about balance or fairness, and as such, simulations shouldn't either. But true games should care about balance and fairness. The idea behind a game is that you learn the rules of the game and exploit that knowledge to complete the victory conditions of the game. The concept being, all things being equal, you and your friend will be evenly matched in every scenario and thus comes down to who actually plays better.
. For instance, Hopscotch isn't the kind of game whose rules you can really exploit mathematically. It ultimately becomes a matter of how good your hand-eye coordination is (for throwing the token into a particular square) and how good your balance is (for the hopping up and down the track). The rules are perfectly consistent at all times for all players, but some players just suck at it. And the same should be true for all games. As games have gotten more complex and varied, this has been lost to the point that most don't even remember the fact. Hopefully it makes a comeback soon because games nowadays are just getting worse and worse.
. Here's hoping.