I am totally confused by this thread...
First of all a question here:
Where should the classes end and the themes begin?
I think you mostly mean themes as "kits" here, because 5e themes so far are merely collections of feats, while you are trying to see if a class can be turned into a theme, am I right?
If you are instead really considering to use 5e themes to represent some of the classes, then the big deal is which class-defining features can you turn into feats? Nothing with a level-based progression can easily be turned into feats, which by their nature aren't scalable (or are only minimally scalable). Thus, you cannot turn a Sorcerer into a theme unless you change its mechanics a lot and just try to represent the concept with feats.
And here comes my main question. Are you trying to "tier-ize" mechanics or character concepts?
If you're trying to tier-ize mechanics, this might make sense for the designers, it might simplify their work, but doesn't have to have any consequence for the players. There is no need why a mechanically derived class must imply that it is also more rare in the world.
If you're trying to tier-ize character concepts, I am in general against it. Which concepts are more common or not is way too much dependent on the setting and the DM's preferences, to standardize such thing brings no benefit to the game at all, and only brings unnecessarily assumptions from some players, while forcing the DM to make it explicit if she's not following the standard.
Also keep in mind that something like this would significantly affect the amount of material published for different classes, because a "common" class will get more supplementary stuff than a "rare" class.
In general, I see no reason why since "Fighter" is a more general concept than "Barbarian", you have to assume that there are more fighters than barbarians in the world, just like there aren't necessarily more BSc than MSc (which might be true in the US, but is absolutely not the case in several EU countries).
Especially with spellcaster classes, this kind of assumptions are more grounded in D&D legacy rather than on the consideration of the character concept in an average setting. Everyone assumes "Wizard" should be the most common class, but that's mostly because
we have seen a lot of wizards in our games and much fewer sorcerers or warlocks for example, since "Sorcerer" and "Warlock" weren't even a class before 3ed. OTOH if you think about the
concept, then being a wizard could be a tremendously more difficult job than being e.g. a witch. Note that despite the main concept behind 3ed sorcerers is "natural born spellcaster", the practical use of it in some settings has been more similar to witchcraft (that's IIRC more like the meaning of "sorcery").
There's quite a lot of room there for rarity of character concepts, that I would like the
DM to be totally free of choosing, without being told what should be the standard. Some might want a high-magic world full of academies of wizards where a sorcerer is a rare individual born with magic, some other wants a low-magic world where commoners learning some magic rituals are sorcerers and being a wizard is rare because it requires superhuman intellect.