lutecius said:
Your approach would only work if Int was the measure of a single ability. But it affects several thematically related things (in theory) like languages, knowledge skills and feats, and ability checks. Playing your high-Int simpleton wizard would require either to ignore these bonus or to find some really convoluted explanations.
Unless you're willing to give up your Str bonus for jump checks because "you play your character as if he has less lower body Str", your example only shows a lack of detail/complexity, not a contradiction like the high Int simpleton. If there were such things as Lower body Str and Upper body Str scores, I’d expect them to be consistent.
I already alluded to this in the latter part of my Strength example. The problem is that Abilities (game consturct) already cover an unreasonably wide range of abilites (abstract notion).
A high Con means you are resistant to pain, disease and poison, have excellent aerobic and anaerobic fitness etc...
High Int makes you harder to hit with swords and arrows, more likely to notice things in your environment, increases your book-learning and makes your magic more accurate and dangerous. Why should book-learning ever correlate to AC increases in the first place?
I could go on.
If we assume that an ability score is an accurate model of all the things it encompasses, vast swathes of perfectly reasonable concepts are impossible to play. What if I picture my character as a sprinter, with massively built legs and only average upper-body strength? Am I not allowed to describe him thusly? What if he works out his upper body, but has a beer gut? What if he's got a poor pain threshhold, but great fitness, or vice versa? Clearly, we are already disregarding the strict interpretation of ability scores with a large number of concepts. Things only get more muddied when we move onto more abstract notions such as charisma, wisdom and intelligence.
I'm not saying that my suggestion is the best way of treating abilities. However, I don't think it's any less realistic than the whole notion of Ability scores in the first place.
It is a very big step away from they way ability scores are traditionally treated, and I pointed out in my first post on the subject that I expected strong of opposition to the idea. If it doesn't suit your (generic you) preferred method of understanding Abilities, that's fine. Both systems are really quite poor abstractions of reality, when it comes down to it; the one advantage my system has is that it reduces the need for an ability to actually model more abstract notions in the first place. The major disadvantage, of course, is the disconnect between the common understanding of a term like "Strength" and what it actually ends up meaning under my use.