Sir

Your Briefing on electronic journals asked “Why do researchers download and print?” rather than reading articles on the screen (Nature 397, 195–200 1999). The author suggests that the explanation lies with the poor resolution of text displays on computer screens. I think the real problem lies much deeper, in the way our brains process information, and will not be susceptible to a simple technological fix.

Reading requires us to pile one level of abstraction on top of another. We must recognize letters from the shapes of marks on the page, and the meanings of words from those letters and from the sounds our minds associate with them, and the information in sentences from the meanings of words and from grammatical rules, and still-more-abstract concepts from the sentences. One feature of reading that makes this achievement possible is that it is grounded in a physical object, the page or book.

One example of the importance of physical context in reading is our positional memory for text — frequently we can easily remember the location of a piece of information on the page and in the book, even though we've forgotten its textual content and context. I suspect that this physical framework is a key part of our reading ability, especially for difficult material.

When we read online we forfeit this level of perception — all text is ephemeral. It flows like water across the screen, and appears and disappears with the click of a mouse. For serious reading of the scientific literature, an abstract may be all we can absorb from the screen without becoming disoriented. It's as if the ideas we're reading can't be given locations in our minds if they don't have locations in space.