I might be in the vast minority here, but I also think that it's worth asking if, historically, large empires lost some amount of effeciency in innovation and wealth generation. The Greek City-States and Venetian Republic jump to mind as examples of cultures that seemed to punch above their weight, population wise (though I'm certainly no historian, so take those specific examples and the larger point with a grain of salt, if you'd like). Call me crazy, but I think that the real world provides tons of fun game mechanics if abstracted properly. Anyway, it's food for thought.
Paradox games: Europa Universalis, Victoria, Crusader King, HoI, etc. In them, being bigger is better very often too, but it's not the sure death sentence like in most civ-alikes.
One of the difficulties is that diplomacy doesn't matter as much in most games as it does in the real world. I mean in reality the US could conquer Canada to get at that oil, but why would they? Despite huge millitary superiority it would cost a fortune, it would be hard to hold territory, would do damage to the resources they want to capture AND the infrastructure required to exploit them, it would lead to worldwide condemnation, etc... and isn't worth the headache when Free Trade accomplishes similar results (access to resources) but instead creates a diplomatic ally.
There is a very neat mechanic for that in Europa Universalis. You have a stability number (from -3 to +3, giving significant bonus and malus to all aspects of ones Empire plus it heavily influences the revolt risk).
Declaring war against a friendly neighbour then gives you -1 stability for no casus belli, -2 for attacking a friendly state. Such a hit to your stability then can easily mean empire wide revolts, large money losses (iirc +3 gives 30% bonus production, -3 gives -30% production penalty), lower fighting efficiency of ones troops, etc. If you repeat that then and land on -3, it can easily mean, that you'll spend the next ingame years trying to recover from that, or even lose parts of your Empire to revolutions. Increasing one's stability takes time and lots of money.
In other words, stability describes the internal stability of an Empire, which is a very good mechanism to simulate populations. For example, making harsh changes in one's politics can mean a hit to stability. Declaring wars against friends harms it. Losing a war? Big hit to stability. Espionage can lead to a stability hit as can diplomatic blunders. On the other hand, wise decisions, good economic sucess, etc. can lead to an increase. Ie. you do unpopular stuff, you lose, you do popular stuff, you gain.
Hmmm... What would really help in Elemental to even out research in small factions versus large faction is to make the defining limit on studies be the cost not the build time. In the current system it's build time, so the more cities you have the more studies you can pump out, and more research you get. If it was cost then small factions could keep up with large ones if they planned well.
Another idea from another Paradox game, Victoria. In that game, you have a literacy rate. Depending on that percentage, you got a big bonus to research or a lower one or none (it's granular). As a small state, you could easily increase your literacy rate to the high nineties. As a big, expanding empire though, conquering illiterate indigenous people your literacy always stayed low and increasing it to high numbers would have cost twice or thrice your yearly income per year. Thus, as a small state, you could focus on literacy and research almost as fast as some bigger, but illiterate, empires.
I always liked those two solutions to easy war and easy research, since they give an easy metric to gauge why and how one can improve, instead of diminishing returns as seen in the older civ titles with corruption or the prestige penalty in WoM.
You see, those are just eight examples off the top of my head. Someone actually getting paid to do this should be able to come up with upwards of 30 of these kinds of technological intersections. That's what I mean by 'nuanced tech function.' And you're right, the other two aspects of my first post would be just refraining from naming the techs as explicit descriptions of the bonus they provide you or mechanic they unlock. This doesn't do any favors for immersion. And flavor text is certainly important. Ambience is possibly the most important thing that separates a good game from a classic. And, luckily, they are low cost additions to the game in terms of man-hours.
Oh how I'd love research with imaginitive and intricate details, descriptions and flavour. You research large smelters or use fire magic for steel production. Illusion not only gives spells but helps with diplomacy. Water mills? We have wind mills and wind magic! Who builds walls when one can grow one through earth magic. My Empire's sorcerers have researched five old languages (bonus to magic research) but also invented a new way to cast magic through song, which coincidentally also boosted singing to a high art form, pleasing the bored nobility. Or whatever, I'm not good at it, but LightofAbraxas is.
Kaleb, hire that guy to develop the individual techs!