It's worth keeping in mind that different people view the same rules in different ways. You may see that a failed Nature check caused a gorge to appear, but I would say that the gorge was always there and your failed check caused you the ill luck of running into it. Inspiring healing when you're unconscious... is actually kind of an incredibly standard trope in heroic fiction.
When we make the causal judgement, there are two "locations" where causation is occurring: in the real world; and in the fiction of the game.
In the real world, it is the failed Nature check that causes the GM to narrate the presence of the gorge.
In the fiction, it is the player's poor choices and bad luck that cause him/her to encounter the gorge in his/her path.
Thus, there is a lack of connection between ingame causal processes and action resolution - because the action resolution is not modelling the PC's choices and luck, which are what cause the PC to encounter the gorge, but rather is simply generating the outcome. (For instance: the action resolution does not require the making of any navigational choices, or contain any components or elements that model navigation.)
Inspiration healing is probably harder to analyse, because it takes us into the notoriously murky realm of hit points. But given that (say) Inspiring Word works the same whether the damage taken is ongoing acid damage from being swallowed by a purple worm, psychic damage that follows recovery from having been mind-controlled to attack one's friends, or falling unconscious from a hammer blow, I'm am confident in say that it is not modelling the various causal processes that take place in the fiction. It is simply geneating an outcome.
In case it's not clear: I don't object to this sort of outcome-based resolution. 4e is rife with it!
Was it there on the map the PCs bought sessions ago? Is it on the handout the players have from the area?
In the sort of play resolved in the way I described above - in which the action resolution mechanics do not model the causal processes that are occurring within the fiction - a map wouldn't be a detailed handout on which the players track their PCs' movements. It would be an item on an equipment list that grants +2 (say) to Nature checks used to navigate.
More generally, action resolution that proceeds without trying to model ingame causal processes tends to rely on a degree of abstraction. The same sort of issue you raise about maps would also be raised by trying to mix a hit location system with a D&D-style hit point system.
Did the PCs see the gorge in their previous travels? Will it still be there in a few sessions?
It would seem to me, that once the gorge is established in play, it would always be there.
Nothing has to be set in stone until you have presented it to the players.
What Hussar and Incenjucar say here seems right to me.
But that's not the issue. The gorge appearing is fine. It's that for some reason the gorge disappears after its existence has been established in play.
I doubt hero quest does that.
Heroquesting happens in a mythical realm, where things
can change depending on how the hero-quester engages the myths and heroic figures that underpin that realm.
This is a bit of a corner case, and I think [MENTION=49017]Bluenose[/MENTION] is right that it probably wasn't what Derren had in mind.
The problem with this is that when a class is say, 90% good at social and 10% good at combat, or even 0% good at combat, it means that other classes have to make up for it.
Couldn't the GM just build weaker combat encounters?