I wasn't saying that a narrative-first style game would produce a railraod. I was saying that I think it's more likely there than in other games.
Perhaps this is a topic for another thread, but I think your intuition here is off. Good contemporary narrative design makes the game reasonably railroad-proof - the players build their PCs (which include the thematic hooks for the GM), the GM builds encounters/situations that engage those hooks, then everyone presses "play" and we see what happens! The whole promise of this school of game design is that it will produce satisfactory stories although
no one at the table is responsible for doing so. The promise is of story emerging out of the players doing their job (ie building PCs with hooks for the GM), and the GM doing his/hers (ie building encounters that bite on those hooks).
This sort of game can fizzle for any number of reasons, including if the players don't build decent PCs (eg they turtle) or if the GM builds boring situations. But railroading shouldn't be one of them.
To bring this back a little bit on topic: one of the techniques that 4e offers for supporting this sort of play is the interaction between the particular powers that players get to choose for their PCs, and the general action economy of the game. This is intended to mean that if the players just do their job - do their best to use their PCs' powers to win combats in cooperation with their fellows - and the GM just does his/her job - builds encouners with an interesting mix of NPCs/monsters and terrain (as per the guidelines in the various manuals) - then dynamic, engaging combats will result. In my experience, the design realises this intention most of the time. Furthermore, in my experience it's fairly easy to build both PCs and monsters/NPCs/encounter settings that have sufficient thematic "oomph" to their mechanical elements that the mechanically dynamic combat will also produce a reasonably thematically dynamic combat.
Healing surges, and the various steps that must be taken to gain access to them, are a key part of these mechanical and thematic dynamics.
That said, I think it is probably fairly easy in 4e to build boring PCs who don't contribute that much to the thematic dynamics, in part because they kill off the mechanical dynamics - archer rangers tend in this direction, in my view, and I would find it easy to believe that pacifist clerics do also. Luckily, there is an easy solution: build more interesting PCs!
As has often been noted, these same aspects of 4e's design - the centrality of powers, of encounter terrain, etc - can create some challenges in relation to fictional positioning. Some think that the game tends to collapse into nothing but dice rolls and moving miniatures around on a battlemat. On the other hand, in my game I haven't had much trouble keeping fictional positioning central. I think the focus on thematic as well as mechanical dynamics helps with that - fictional stakes that the players care about will go a long way to making fictional positioning matter - but there are other techniques that I use as well.
I dislike the automatically scaling bonuses in 4E. The game really seems to be about a characters relative level, rather than their absolute. What matters about an opponent is if they are +2 level, say, as opposed to -2. That's built into the attack bonus and defenses to the point that actual level is largely irrelevant. By my prestated design principle, actual level should be removed, and relative level put in its place.
I can see the force of your point. My own view is that the scaling, in combination with the default monsters from the various published sources, produces a game that, very roughly, tells "the story of D&D" ie the PCs begin by confronting kobolds, and end by confronting Orcus.
I know that others also see the scaling as corresponding more-or-less directly to ingame toughness of characters and monsters, but I personally don't make that assumption. When I'm desigining encounters, I do think just in terms of relative levels, as you describe.
4e players seem to complain about what the healing wand did (allowed PCs to avoid 15 minute adventuring days).
Then 4e players turn around and praise Healing Surges because they do the exact same thing, only better!
I'm not sure why someone can use wands as a negative but then use Healing Surges as a positive.
In my own case, it would be because I find that wands of divine juice on tap are kind of lame, whereas heroically turning the tide by drawing on your own reserves of heroism - perhaps when spurred on by a valiant leader - is kind of exciting and evocative of LotR, Arthurian romance, etc. But then I never played very much 3E, and so may be an outlier here.