Totally understand where the author is coming from, but the notion that text should be a 'first class' object to interact with does not match the reality of how people interact with long form text. Very few people interact with the text on a page, and 100% of people reading a book very frequently scroll/paginate [0]. People are used to swiping on content to see more of it, I don't know that it's as confusing as you might think.
Likewise, it also wouldn't make any sense to have a single tap within a mobile browser bring up the dictionary definition of the word you tapped on. Does it save time? Sure, for one power user use case. But it pollutes your otherwise simple navigation paradigm with stuff popping up all over the place promoting text interaction. People hate stuff popping up all over the place unnecessarily. Better to build your navigation paradigm out of the way that 100% of people interact with the content; it's OK to have a few things buried one layer deeper, power users will find them anyways.
Likewise, it also wouldn't make any sense to have a single tap within a mobile browser bring up the dictionary definition of the word you tapped on. Does it save time? Sure, for one power user use case. But it pollutes your otherwise simple navigation paradigm with stuff popping up all over the place promoting text interaction. People hate stuff popping up all over the place unnecessarily. Better to build your navigation paradigm out of the way that 100% of people interact with the content; it's OK to have a few things buried one layer deeper, power users will find them anyways.
[0] I worked on these products at Amazon.