I kinda feel that in these kinds of games, the characters do exist in a vacuum.
The characters don't exist in a vacuum in the setting; thus if they're for whatever reason undertaking a journey it's not like they're not going to look around at what they pass along the way.
This points to my big issue with scene-to-scene jumping - way too much potential gets missed in between, and way too much opportunity for interaction with the setting, perhaps even in ways not directly related to any existing beliefs etc. Also missed is any opportunity for the players/PCs to at some point decide to sidetrack themselves into something else (which I certainly hope is allowed; they're not on rails, are they?) e.g. they stop and do a good deed for some villagers along the way and come to realize there's a bigger issue there, so they stay and sort that before continuing to town.
Here, seeing the abandoned farms might eventually end up causing a player to change a character's beliefs (it's allowed, isn't it?). That opportunity never arises if the PCs are jumped straight into town without any narration of their journey and what they see during it.
Of course the characters don't exist in a vacuum. Play is just tightly focused on the characters.
As far as scene-to-scene jumping, I actually agree that it can be overdone, but I think if I remember
@pemerton's description of how play developed around the scenario in Evard's tower correctly, events leading up to that were pretty granular. Aramina didn't roll her Grand-Wizard Wise and then everyone was at the tower. There was a journey and some diversions along the way. They were all related to player priorities and beliefs, but they were there.
Regarding side-tracking and changing beliefs, they're both possible and usually related (and changing beliefs is absolutely encouraged and expected). A GM can absolutely put a character in a hard place to test their commitment to a stated belief — I've been watching a lot of westerns lately, and in
Bend of the River, they've established that Cole is a former border raider looking to make it rich (keeps on talking about going down to California for gold), but for a while he accompanies Glyn and the settlers up the Columbia with the food and helps them. When he's offered $100k to bring the food to the gold strike instead, he's presented with a hard choice. In a BW sense, he's got a belief about bringing the food to the settlers, helping Glyn, or whatever and maybe the Greedy trait or something. Putting these in opposition, forces his player, Arthur, to make a choice about what Cole will do. And would likely lead to a change in beliefs and a different pattern of play.
David Mamet once said that a screenwriter should get into a scene as late as possible and out of it as early as possible. In some ways, I think the same is true about where BW play should start. From there, pacing can vary, but the GM should be driving to something hitting a belief and hard quickly.
I thought part of the drive in BW was that the players were also responsible for bringing their beliefs into play? That is, it is good if the GM helps, but if the GM is not helpful the players still can work to make the tings the GM narrate into something that affects their beliefs (and hence keeps the artha flowing)
Yes, players should be acting on their beliefs, too, and you're right to bring up artha flow, but I think it's absolutely a responsibility of the GM to target beliefs (and traits) constantly and keep the pressure on. On some level, if players aren't acting on their beliefs from the drop, then it speaks to a situation that's probably half-baked and needs to be ratcheted up a bit. There should be urgency.
Edit: Jimmy Stewart's character is Glyn.