Somewhere I have the bones of a super ginormo followup post to my original "fat and simple" one where I discuss causality itself in more depth (for example, I came up with 4 axes for categorising kinds of causality; I can only imagine that some long-dead philosopher has beaten me to it).
Lots of these debates actually resolve around definitions.
Anyhow, the other stuff I've written under the Systems heading may be of interest to you[1].
As for the causality things, I don't know about philosophers, but there is some good writing on the topic by Eliezer Yudkowsky, both the older texts[0] and the new ones about causality, like [1]. I found them very insightful.
Anyway, off to reading systems stuff ;).
[0] - http://wiki.lesswrong.com/wiki/Sequences
[1] - http://lesswrong.com/lw/ev3/causal_diagrams_and_causal_model...
The trick is, what I so far found out, to establish a set of meals that you can actually stick to for a longer time without feeling nauseous, while still maintaining calorie deficit. There are also issues like insulin reaction; so far I believe in this one, as hunger doesn't seem to feel as bad and doesn't disable my thinking since around the time I cut out sugars completely (I used to drink a lot of them). But accounting for all those things only helps you to stick with your diet.
So bottomline, I think that most of the other mechanisms people bring up are important not for weight loss per se, but for the ability to stick with a given diet for longer, while calorie deficit does it work.
FYI: I'm on keto for just over two months (with lot of cheating unfortunately - there were so many parties recently...) and have lost 10kg in that time.