However I've seen it used in different contexts (eg. a web browser) where it's less easy to defend.
How do you know on an android contact that you can swipe sideways to SMS and the other direction to phone? Or even that swiping horizontally in either direction does anything at all? I could imagine some universally agreed way to indicate this, perhaps very subtle visual hints or gradients or a system where if you press and hold without moving you get a set of icon overlays that show you what you can do.
There are a few different options, but to form a useful language they need to be widely adopted.
One of the articles that triggered these conversations was "No to NoUI" http://www.elasticspace.com/2013/03/no-to-no-ui which laments the modern tendency to hide interactions.
It's almost impossible to know when you can drag and drop things, and the only way to learn is to start trying by dragging on random UI elements. Of course, knowing what happens once you drop them on something that cares is just as impossible as knowing what you can drag in the first place.
Sorry if I sound pessimistic, I hope development and focus is better now, and of course touch screen devices are in front of a much wider audience which might prompt a more rapid development of some standards.
The smartphone market has reached a level of maturity and ubiquity that we can start setting down a baseline for knowledge, and build on top of it.
It happened to the PC before, and it's happening to smartphones now. The original iPhone was criticized (especially in demographics like ours) for being insultingly shallow and toy-like. This is the smartphone becoming more complex now that (nearly) everyone has had a go.
1. Swipe to delete
2. Shake to shuffle
If this hypothetical person has never used a touch device before, they'll be stuck on the "scrolling and dragging around" step, but it won't be long 'til they stumble upon the pull-down.