Deep focus and realism

In his latest blogpost, Do filmmakers deserve the last word? (October 10th, 2007), David Bordwell uncovers fascinating insights about the relationship between filmmaker's talking points and what the audience and critics make of them. In particular, the contextualization for the birth of the deep-focus critical concept, coming from Welles and Wyler's cinematographer, Gregg Toland, is incontestable, as Bazin appropriates the same talking points almost word for word. Gregg Toland lays out the principle of his revolutionary technique, "pan-focus", in a 1941 article. And Bazin re-uses it, under the name "profondeur de champ", in his essay "L'évolution du Langage" which dates from 1955, where Toland is never mentioned.

But I'm not sure Bazin would accept all Bordwell's implications as is :

1. Bazin is a "plagiarist"
2. Bazin's critical theory is shaped by publicity talking points
3. Some "deep-focus" scenes from Citizen Kane were actually forged, thus disproves Bazin's theory of realism
4. "Deep-focus" existed before Toland in pre-1920 cinema

I'm not arguing with (1), the precedence closes the case, and Bazin should have at least cited the article, as his duty of journalist would command. It's unlikely he would have phrased it exactly the same way without knowledge of Toland's speech. It's really odd though that Bazin would intentionally resist to mention the cinematographer's name at the origin of this invention...

(2) However, I would like to moderate the interpretation of this case as critics being subject to plagiarism and influence. Critics never invent technical or aesthetical devices themselves. Their job is to spot them, analyse them, understand them, trace their genealogy and explain them to the public. Conversely, it's not enough for a filmmaker to spell out a theory to earn a landmark in history.
Critics either find out by themselves by looking at the pictures alone, or talk with the filmmakers to learn from their practice. But in the end, the critics make the decision to validate or to dismiss whatever is purported by filmmakers' or publicity's talking points. I mean great critics there, more precisely, theoreticians and historians, not the reviewers of course.

So Bazin cherry picked one of many claims championed by their auteurs out there, found it credible and fruitful, and added his credential to it by publishing it under a more elaborate theory. Like Bordwell says, we can't listen to everything filmmakers claim they do if the screen disproves it.
Toland could not make history by himself if nobody out there was listening. He couldn't trumpet his own glory alone either.
By the way I would like to know what were the repercussions of his article in the USA. Did American critics understand it like Bazin did, 14 years later? Did the public opinion receive Welles and Wyler as geniuses like they were after Cahiers celebrated them, once these films made it across the Atlantic after WW2? I think the appropriation of a cinematographic device by a critic is what makes all the difference. It took Bazin to transform a publicity stunt into a critical landmark. Inventors of form could go unnoticed if they are not endorsed by a critical authority. Sometimes the filmmakers aren't even aware what they do unconsciously is truly revolutionary.
Like Bordwell reminds us, Greengrass claims he revolutionized cinema language... but it's the critics job to validate or invalidate this talking point.
Bazin's theory of deep-focus and realism goes well beyond whatever Toland proposed, which was mainly practical issues.

"That's why deep focus is not a cinematographer's fad like the use of filters or lighting, but a capital gain for mise-en-scene : a dialectical progress in the history of filmic language.
And it's not just a formal progress! Well mastered deep focus is not only a mere economical way, simpler, subtler to emphasize the event; it affects, with the structures of filmic language, the intellectual relation of the spectator with the image, and thus modifies the meaning of the spectacle."
Bazin (L'évolution du langage)

(3) André Bazin (L'évolution du langage) :

"It's obvious, to whoever can see it, that Welles' plan-sequences in The Magnificent Ambersons are not at all mere passive "recording" of a photographic action within a single frame, but to the contrary, that the refusal to break up the event in bits, to analyse over time the dramaturgic space is a positive operation which effect is superior to one produced by traditional cutting."
"(...) deep focus places the spectator in a relation to the image closer to the one (s)he experiences in reality. It is thus right to say, that independently from the very content of the image, its structure is more realistic."

In a footnote of his essay "Montage interdit", he describes the scene from Where No Vultures Fly (1951) where, after a parallel montage, a little boy with a lion cub in his arms and the mother lioness meet in the same frame, which constitutes the recreation of reality for the spectator. But he acknowledges that the lion is obviously tamed and that the boy's life is never threatened unlike the shot suggests. So to Bazin, it's not so much that whatever happens on the set should be the reality handed over to the spectator, but that the mise-en-scene should recreate the conditions of reality (which would be otherwise negated by heavy editing), that the filmic language, with its technical devices and tricks, should not betray our perception of the time-space continuum on screen. We know cinema is an illusion, in so many ways. But the mise-en-scene may choose to betray reality or to reinforce it, which determines the realistic approach of the filmmaker.

Thus the post-production tricks of transparency for Citizen Kane doesn't negate the theory of realism, as long as the frame gives the impression of something inherently plausible on screen. Besides the foreground and background added (for aesthetical composition purpose) into the shots described by Bordwell, do not alter the main dramatic action within the frame. There is no direct interaction between the drama unfolding in each separate shot of the double exposure. Which is very different from the deceiving interaction suggested by CGI tricks where the actor actually interacts with a green vacuum on set. The green screen superimposition pretends two characters talk to each other while they never had a lifelike experience together on set.

"It's not that Welles refuses to resort to the expresionnistic devices of montage, but precisely their episodic use, between "plan-sequences" with deep focus, gives them a new meaning. (...) In Citizen Kane, a succession of superimpositions contrasts with the continuity of a single-take scene, it is a different modus operandi, explicitly abstract, of the narration. Accelerated montage cheats with time and space, but Welles doesn't attempt to fool us (...) Thus the "quick editing", "Attraction Montage", superimpositions that talky cinema hadn't used in 10 years, acquire again a possible use in relation to the temporal realism of a cinema sans montage."
Bazin (L'évolution du langage)

(4) Bazin acknowledges that the wide shot with deep focus existed since the origins of cinema. The focus of early cinema lenses was designed to capture pretty much everything in front of the camera (like the cheap disposable cameras today).

"Agreed, like in the case of Griffith's close up, Orson Welles didn't "invent" deep focus; everyone in primitive cinema used it, logically so. Image blur only appeared with editing."
Bazin (L'évolution du langage)