Egyptians for example would get full culture from temples and cathedrals but less from libraries and universities because they aren't scientific. I could see people in a religous non scientific not appreciating libraries as much as they do temples.
I've been mulling on this for the past few days, and have had a couple of related notions that seemed worth throwing out for comment.
One, which I do not think is doable in Civ 3 as is, would be to make buildings whose cultural or indeed other output varied with government type. So that if one had a theocratic government, religious buildings would get a culture boost - and possibly a higher one again given a fundamentalist government in which the impact of science-boosting buildings, on culture or otherwise, was reduced. Whereas if, connecting on to my earlier point about wanting government types to improve over time, you had a government that represented the post-French Revolution First Republic [ I'm stuck for a snappy name for it; technically it wants to be Enlightenment Republic, to distinguish it from Roman-type republics, but that's just bound to get misread by people reading "enlightenment" as moral approval ] or indeed Communism. religious buildings would be enough against the prevailing ethos to have drastically reduced or negative culture.
The second, which is a somewhat more radical notion, is to reverse the causality suggested in Marsden's original post. Rather than picking a scientific and industrious culture and therefore getting cultural bonuses for scientific and industrious buildings, why not have the traits of your culture be something you define by your early build strategy ?
I'm thinking something roughly like each trait being assessed by
- Industrious - how many shields you get compared to the maximum you could get with those cities in those locations
- Agricultural - how many food you get compared to the maximum you could get with those cities in those locations
- Commercial - how much commerce likewise, or what proportion of your income goes into cash
- Religious - how many religious buildings you build
- Scientific - how many scientific buildings you build, or what proportion of your income goes into science
- Militaristic - how many barracks you build, how strong your military is, how aggressive you are
- Expansionist - how many squares you explore, how many cities you build
And then at a certain point, your civilisation is assessed on how you are doing on each of those counts, and you get assigned two traits, and the appropriate bonuses, both the existing Civ 3 ones and Marsden's suggested cultural ones.
It is, admittedly, hard to see exactly how best to compare all of these traits evenly. And while one way would be to compare your own Civ's scores for all the traits and assign you the two highest, another would be to compare you with the other players in the game and assign you the traits at which you were doing best relative to the average of everyone else, which might not necessarily be the same as which you were doing best at on an absolute scale.
The real appeal of this, for me, would be in a context of having more Ages. If the game had, for example, Ancient, Classical, Medieval, Renaissance, Industrial, Modern, and Space Age, and every time you went into a new Age your traits were assessed and might change, I think that would raise interesting possibilities. I don't think that really works as a mechanism with only four Ages though, particularly with so many of the most efficient ways to win involving not going into the Modern or even Industrial Age at all. I think this would add the same sort of benefits to gameplay as having advanced later governments and game balance that encourages switching into them.