Especially in 1st ed AD&D I think these can be tricky, as your list brings out.I think I need help with the vocabulary and differentiating between defining characteristic, niche, trope, and stereotype.
Agreed.The ranger with tracking - in 1e we picture the character having it, they have mechanical support for it, and they're the only ones
Except that the ranger gives you the surprise bonus. I think it's a mechanical weakness of 1st ed AD&D that it doesn't integrate its surprise rules with its stealth rules in any very consistent way.Being sneaky - we picture the 1e thief having it and they have mechanical support for it basically only at the higher levels
Being sneaky in the woods - the 1e flavor text makes it sound like the ranger should be, but there is no mechanical support for it and the thief is the one you'd actually want
In his DMG (p 86), Gygax gives as the characteristics of a poorly-played magic-user "seek[ing] to engage in meee or ignor[ing] magic items they could employ in crucial situations", and of a poorly-played fighter "hang[ing] back from combat or attempt[ing] to steal, or fail[ing] to boldly lead".Being smart - we picture the mage being the smartest, and there is flavor text to support, but no mechanics in 1e, and in 3e the mechanics actually support other classes being more knowledgable.
<snip>
The mage rocking out the knowledge checks doesn't seem to be at the same level as the thief being the best at opening doors and sneaking around or the mage being the one with a huge variety of interesting spells.
I think that the smarts of an MU, and the contrast with other PCs in that respect, were expected to emerge, at least in part, via the way the PC was played.
I don't know 3E well enough to know what other knowledgeable classes you have in mind (the bard? but won't INT be weaker for a bard than for a wizard, so that the wizard will somewhat make up the gap in skill points and have a better stat bonus to Knowledge?).
In 4e, the wizard is I think the most knowledgeable class, having access to all 5 knowledge skills (Arcana, Dungeoneering, History, Nature and Religion), having good INT and (for some builds, at least) secondary WIS, and having good access to rituals. An invoker is also a good scholar, with good WIS (but no class access to Dungeoneering or Nature), and (for some builds, at least) secondary INT. Like wizards, psions favour INT, but lack Nature and Religion as class skills; swordmages favour INT, but have lack Dungeoneering, Nature and Religion.
Warlocks can also have secondary INT, but are likely to have lower WIS, and lack Nature and Dungeoneering as class skills, and so are second-tier intellectuals. Bards have access to the full range of skills, but will have secodary INT, or secondary WIS, but not both. A warlord can be a good historian (History class skill plus secondary INT) but has no other knowledge class skills.
So maybe because I have a lot of experience with Rolemaster, where casters tend to have narrow spell lists by classic D&D standards but are also generally the best at knowledge, or because I am currently GMing 4e, which tends to resemble RM in this respect, I have a different view of the situation.
I'm from the anti-Find Traps and anti-Knock school, so no disagreement there! (4e fixes both.)It seems the mage having spells like knock or the cleric having find traps, or in 4e all characters having access to miscellaneous magic seem much more niche busting than a sage type.
I'm not worried about the sage's overlap - it's the sheer mechanical effectiveness. If it's clearly better than the casters, they get downgraded in relative terms which means, if the game is tightly designed, also in absolute terms.
I was thinking of this more in terms of system parameters than individual scenario design. (If there is a good Lore resolution system, then its DCs need to be defined with reference to the range of PC build parameters just as monsters, which are central components of the combat resolution system, are designed by reference to those parameters).I really like the idea of how it impacts setting up the DCs, but I think that leads to a whole other area.
I agree with you that scenario design is a somewhat separate area. I think good scenario design permits multiple avenues, and tries to spell out consequences for that. So if the players don't have a scholarly sage among them, they have to learn the info some other way (say, go to the scriptorium and ask politely, bribe, or just shake the place down!). And this other way should matter to how things unfold. (This is tricky in pre-packaged scenario design, but I think not impossible - look at the scenarios in the HeroWars Narrator's Book, or some of the Penumbra d20 modules.)
As far as D&Dnext is concerned, I think this is where background traits could be put to work. The players and GM should both be playing towards traits as natural guides to likely approaches to a scenario, and also sources of consequences.