Rather than consulting game rules, many times the GM is required to make a judgement call in response to a player action, based on what he thinks is realistic (whether it is real-world realistic or game-world realistic is an important choice for the GM). Because this is a human element adjudicating player actions, the presence of rules-artefacts will arise less frequently.
I applaud you thinking hard on this topic, but I think you need to take it one step further.
While its true that human judgment gives a richer experience than any possible rules set, if what you propose is true, then the most 'realistic' game system would be one with no rules at all. The GM simply responds to any proposition with what he thinks should happen.
But I think you'll quickly see that if this was the case, there would be an enormous number of 'judgment artifacts' that would arise in the course of play where the GM is forced to make a snap judgment and ends up making a bad call. How the GM decided to handle it was unrealistic in some fashion or full consequences were poorly thought out.
What I want you to see is that there is no real difference between a 'rules artifact' and a 'judgment artifact'. In any rules-light system where the Game Master was striving for realism, very quickly he'd find himself making rulings on the same sort of things again and again. In order to reduce his work load and bring some consistancy to the game, the GM would handle the same ambiguity in the same fashion again and again. In effect, the GM will have created an unwritten house rule. Once the GM does that, the rule becomes part of the game system, just as if it was part of the official rules as written.
Suppose that aging really did play a big role in our hypothetical game, but that the rules were unspecific on how aging was to be handled. Our novice GM would probably invent some house rule which reduced physical attributes, and possibly increased certain mental ones. Perhaps then he'd realize that the existing rule was inadequate, because after a certain point mental attributes should begin to decline as well. So a new house rule would be created as the GM's system increased in complexity. Then the GM might eventually find even this system inadequate to deal with reality in a way that felt real, so the GM might invent a system by which the aging characters would acquire various conditions that hindered them in various specific ways (incontinance, poor balance, lame, fragile, hard of hearing, whatever). At last, the system might now feel sufficiently real, but alas in an effort to be both fair and realistic the real system in use would no longer be the simple elegant one at the beginning. As someone else has pointed out, D&D's aging rules don't fail because they are too rules heavy - they fail because they are too rules light.
All game systems have a tendency to become more and more codified the longer you play them as in any system, however complete or incomplete, the GM institutes rulings to cover ambigious situations and reasonable player propositions that aren't covered under the rules. Each ruling becomes part of the common law of the table, and is referenced probably both by players and the GM when determining future courses of action and how to handle new ambiguities. The problem is of course, that not all GMs - even good GMs - are also good rulesmiths. Even the best GMs make bad rulings, and some of the best GMs make alot of really awkward rules that never achieve what they intended them to achieve.
So strange artifacts are going to develop whether you use a rules light system or a rules heavy system. The real question isn't how many rules do you have, but how comprehensive they are. Do the rules in fact cover all possible player propositions? Typically they don't. Rules light systems tend to leave alot up to the DM to decide on the fly, and eventually evolve into rules heavy systems over time. Rules heavy systems tend to have the illusion of completeness, but only by limiting player imagination to the propositions which are covered by the rules. Whether rules light or rules heavy, there is a tendency for players to look at a rules heavy system and simply discount as impossible any option not expressedly permited.
Take d20. Can you study the stars to see signs and omens that hint at the future course of events? According to the rules, only if you cast a 'divination' spell, at which point the stars under the rules become unnecessary. The result is that D20 players typically don't become astrologers, no matter how much a part of fantasy source material that may be. What happens to a D&D player who isn't a cleric who offers up or fails to offer up a sacrifice to a diety? According to the rules, nothing, so d20 players tend not to make religious sacrifices. If you are wanting to model your fantasy after the Illiad and the Oddyssey, then here come the house rules in some form. Very likely these will quickly evolve beyond, "The DM randomly smites or blesses you.", in order to make the game feel more fair and less arbitrary.
And so on and so forth.
The problem that any GM running any system has is that 'reality' is complicated. It's not simple. An almost endless variaty of things can take place in 'reality'. If these are to have any effect other than color, rulings will have to be made to cover them. So any rules light system that strives to be realistic inevitably becomes a rules heavy system, and often a very awkward and clumsy one if the original system isn't comprehensive enough.