Deep Reading
Deep reading is sustained, immersive engagement with long-form text, holding an argument or a narrative in mind long enough to form your own inferences about it. The claim Nicholas Carr borrows from Maryanne Wolf is that this is a built capacity rather than a natural one, and a historically recent one. Early Western text was written without spaces between the words, which made reading a decoding task rather than a fluent act. Word separation made fluency possible, and fluency freed attention to be spent on meaning instead of mechanics. On Carr's account the contemplative reader is roughly 550 years old. The historical half of this argument is well supported and genuinely surprising. The further claim, that the internet is destroying the capacity, is not supported, and the only measured evidence in the vicinity is the small and bounded screen inferiority effect.
What it is
Start with the thing the argument is built on, which is that reading is not natural. Human attention evolved to track change, and holding still on a static object for an hour runs against it rather than with it. Every part of the apparatus, the letters, the spacing, the page, the practice, was assembled, and the mind that uses it was assembled alongside.
The historical evidence for this is more specific than a general appeal to culture. In Greek and Latin antiquity the norm was scriptura continua, writing without spaces between words. Paul Saenger's argument in Space Between Words: The Origins of Silent Reading (1997) is that unspaced text, combined with flexible word order, imposed a burden that kept reading slow, effortful and largely oral. The reader's capacity went into working out where one word ended and the next began. Word separation spread through the early medieval period and was established by the thirteenth century, and once decoding stopped consuming the reader, that capacity could be spent on what the text meant. On Carr's account this is the precondition for deep reading, and therefore for the contemplative reader.
There is a famous illustration of how differently reading once worked, and it needs handling with more care than it usually gets. Augustine records his astonishment at watching Ambrose of Milan read without moving his lips. The passage is quoted constantly, including in The Shallows, as though it proved that silent reading was invented in the fourth century. It does not. Historians dispute how much weight the passage can carry, silent reading is evidenced before Augustine, and his surprise may have been at the circumstances rather than at the act. What it is good for, and what this page uses it for, is the more modest and still interesting point that reading practices change enough over time for one era's ordinary act to strike another as remarkable.
Carr's own use of the history does overreach in one identifiable place, and it is worth naming because it is the same move that undoes the rest of his case. He asserts that an ancient reader's entire cortex "would have been buzzing with neural activity". That is an inference drawn from a historian's textual argument, not a measurement. Nobody scanned anyone in the fourth century. The historical claim stands on Saenger's evidence about manuscripts and reading practice; the neural gloss added on top of it stands on nothing, and dropping it costs the argument nothing at all.
The reason this concept matters to a publication about attention is the inversion it performs. If deep reading were a human default, then any difficulty with it would be damage, something done to a person that ought to be reversed. If it is a cultural achievement built over centuries and maintained by practice, then difficulty with it is closer to disuse, and the question becomes what environments sustain the practice rather than what is wrong with the reader. Carr's line, and it survives everything that later happened to his neuroscience, is that anything built can be lost.