After re-watching the infamous silent episode from the fourth season of Buffy the Vampire Slayer, "Hush," I am compelled to conclude that Joss Whedon loves him some tracking shots. Unless the camera is tightly framed on someone's face, it's either panning or tilting or both. If this were my only encounter with his work, I'd probably draw the conclusion that because the majority of the episode contains no spoken dialogue, Whedon felt the camera work had to be dynamic enough to sustain audience interest. But that's not the case: Whedon's camera is consistently active, only a bit hyper in "Hush." For example, when Buffy finishes her ablutions and enters the hallway, the camera pans with her:
The camera's attention here is mimicking Buffy's curiosity, so as soon as she disappears, the camera whip pans back to our protagonist:
The reason for keeping this in a single shot—besides the obvious one that Whedon always prefers tracking shots—is to create the impression that Buffy has had a realization. At this point, it consists of "something's wrong," or maybe the more innocuous "something may not be right," but the audience connects Buffy's attention to the anonymous student's flight via the tracking shot here.
As noted above, the only time Whedon's not tracking is when he's tightly framing faces. That technique makes sense in an episode in which all the information about the characters' respective mental states is going to be non-verbal. In his excellent post on non-verbal facial cues in The Social Network, David Bordwell argues that the "intensified continuity" in modern cinema requires actors to "be maestros of their facial muscles and eye movements," and though "Hush" is an episode of a television show and not a film, the same applies here. For example, immediately after the above, Buffy enters her dorm and begins to "converse" with Willow:
Because even shows like Buffy prefer some sort of realistic acting, this almost qualifies as over-doing it: this is intensified continuity intensified, but it remains naturalistic in the hushed context of the episode. The only way the characters can communicate is to over-act. Looking concerned no longer communicates being worried unless, as per the last frame above, that concern is exaggerated. (I could demonstrate that this dynamic is operative when Xander and Spike are conversing or when Riley's trying to enter the Initiative, but in the interest of space and bandwidth, just take my word for it.) In sum, Whedon is setting the audience up by having it pay closer attention to facial expressions than they otherwise would. Why would he do that? Meet the episode's antagonist:
They would be the Gentlemen, and creepy as they would be otherwise, their creepiness is heightened by the fact that in an episode that keys the audience to pay attention to the plasticity of faces, theirs don't move. The above also represents the only moment after the initial dream sequence in which Whedon frames them in the conversational medium close-up that he employs for the all other players in the episode. Otherwise they are perpetually in motion and the camera's tracking them:
In other words, outside of that one shot, the audience is going to be discomfited by their paralyzed smiles and the blocking as they hover about and odd angles at which they're shot. I choose the word "discomfit" for a reason: the Gentlemen don't horrify the audience so much as unsettle it, and the techniques Whedon uses to unsettle it are intended to be disconcerting, not frightening.