I started teaching writing in graduate school 20-plus years ago, and it did not take me long to start looking forward to the pile of research papers at the end of the semester. Unlike much of the writing earlier in the semester, done from assigned readings and carefully crafted prompts, the research papers tackled broad, open-ended questions. Students developed their own ideas and went to the library to research topics of their choice. It was exciting to see how they made sense of what they read.

But that was in the old days, before the ease and precision produced by the Internet. Now students hardly ever use books in their research, and their papers have become as predictable as those they write from assigned readings.

When I started teaching, books were easier to find than articles, whose references were buried deep in voluminous, thin-paged indexes. Students took different paths in their research and came up with wildly different sets of texts. Some checked out the better books early, leaving the others to scrounge for what was left. Sure, there was overlap, but students often ended up with individualized research materials, exercising their critical abilities to integrate what they found into a coherent, cohesive discussion.

As periodical-search engines blossomed, students, ever adaptable, started using more articles. While the electronic card catalog remained more or less static, the search engines became increasingly user-friendly. It became so difficult to get students to use books in their research that I started stipulating that they use a minimum number in my assignments.

Then the development of Google and of electronic journals essentially converged. Why bother with books and the stacks when you can search full-text articles online? The process has become even more alluring with database products like Discover (which our libraries enthusiastically characterize as \”the scholarly version of Google!\”). It searches millions of entries, including all of the library catalog, the most-used journal databases, and local historical collections. Like Google, Discover ranks findings according to relevance. With the aid of our reference librarians, students easily set up their searches to obtain exactly what they think they\’ll need, usually in the form of full-text articles.

Consequently, my students hardly ever consult books. Circulation statistics support this impression. In 2005 our libraries checked out or renewed 86,807 books or other media. That number has been steadily declining. By 2012, the number had dropped to 45,394, down 48 percent in seven years.

Why am I bothered by these developments? Well, partly because modern library design mirrors student preferences. Increasingly, libraries are social spaces—with Wi-Fi, study nooks, coffee shops, chat areas, and movable furniture—and not homes for books, which are relegated to off-site repositories, save for a few recent acquisitions. If a student wants a book, she can requisition it. I cannot imagine students already deterred by the stacks having much patience for the repository.

But more important, I am bothered because I think there is pedagogical value in getting lost in the stacks. When I was a student, the stacks filled me with fear but also with awe—they contained so much learning! Today we applaud students not for exploring the stacks but for being efficient, making research quick and easy.

But should research be quick and easy? Yes and no.

If we think of education as the gaining of intentional knowledge, then yes: The students had a question, and they found an answer. Search engines lend themselves well to the learning of that kind of knowledge.

But as the late philosopher Michael Polanyi and others have argued, most of the knowledge we possess is not conscious and intentional; it is incidental, or tacit, acquired as a byproduct of performing some other activity. Practically all the knowledge we acquire as children up to the time we start school is incidental, but the process does not end there. Incidental knowledge continues to play an important role in our adult lives. The library stacks are a mine of incidental knowledge.

Imagine a first-year student in a course on global poverty who is trying to determine if a policy of economic \”degrowth\” can alleviate the plight of the extreme poor. She can find excellent full-text articles on the topic using Discover (which has, for example, papers from the first and second International Conference on Economic Degrowth for Ecological Sustainability and Social Equity, in Paris and Barcelona in 2008 and 2010, and a collection of articles from a 2010 issue of the Journal of Cleaner Production). She could be done with her research.

But what if she decides she wants to look at a book, perhaps Serge Latouche\’s Farewell to Growth, an accessible, impassioned introduction to the degrowth movement. After tracking down the call number, she sets off. On her way, she walks past a wealth of scholarship, some of which may catch her eye. Given her interest in social justice, she might stop and look at the section of books on Brazilian history and culture, which contains prominent titles on social justice. These opportunities for incidental learning happen multiple times on the way to her book.

At her destination, Farewell to Growth is on a top shelf. As she searches for the call number, she sees titles like Democracy: The God that Failed and Contested Sudan and Economics and Psychology. Even if she does not open those books, she can learn from their titles—that some people think democracy has failed, that the country of Sudan is contested, that psychology and economics are not completely unrelated. Freakonomics and its sequel, SuperFreakonomics, may catch her eye as books she has heard about but maybe not read.

Before she spots the thin single volume of Farewell to Growth, she will probably notice Paul Wachtel\’s hefty The Poverty of Affluence: A Psychological Portrait of the American Way of Life, a title that will resonate with the conversations we have been having in class. If she scans the table of contents, she will see chapters like \”The Illusions of Growth,\” \”The Cultural Context of Growth Ideology,\” and \”New Alternatives.\” She may not know of the popularity of Wachtel\’s book when it first appeared, but she would probably understand its intriguing argument: More is not necessarily better, and growth does not necessarily bring happiness. If she notices the publication date—1983, apparent in its call number—she might realize that the degrowth movement may be new and edgy, but that its roots predate her birth. She would learn all this before reading Wachtel\’s book, but she might decide to read it and use it in her paper, which would make for a rather original discussion.

These library journeys feel like a waste of time to students, but they are of immense value. Incidental knowledge may not be of immediate use, but it will become the fuel that powers acts of creativity and discovery to come. Students can see the holdings and make decisions for themselves instead of allowing an algorithm to decide for them.

Students\’ lack of interest in books bothers me for another reason. We know that reading skills in general (and book reading in particular) have been declining over the past 20 years at all grade levels, and that this trend is associated with a loss in the ability to do close reading and to sustain reading for long periods of time. The concern is particularly acute today, as evidence mounts that hyper-reading (the reading of link-heavy online texts) is actually changing the way our brains function.

Traditional, linear reading is less taxing than hyper-reading, and as a result allows us to comprehend and retain more of what we read. We can make a similar argument about research. Compared with digital-database mining, traditional stack-based research is slower, less overwhelming, forces closer attention, and as a result, allows students to build sophisticated knowledge in a measured manner. Books also encourage close reading and rereading (in my experience, at least). Stack-based research produces more-original, creative work, as students make individual decisions about which texts to consult.

I am not making a plea for a return to the good old days. Discover, Google, Wikipedia—these are all wonderful tools. They bring us benefits I could not have imagined 20 years ago, and they should have an important place in our lives. But at what cost?

We need a greater awareness of what we are losing in overprivileging digital tools, and a better balance of digital practices and traditional ones. We must preserve the slower, more thoughtful approach to reading and writing. Part of our mission as teachers is to counteract the preferences that students bring with them and to help them adopt those that enable them not just to gather and scan information efficiently, but also to pursue their interests more purposefully—to encourage them to think and write more deeply, more reflectively, and more creatively. Only when that happens will education be truly transformative.

Author Bio: Julio Alves is director of the Jacobson Center for Writing, Teaching and Learning and of the Writing Program at Smith College.