"Users’ expectations are different for these two. When you search books you expect an answer; when you search journal articles, a scholar expects a list of things to read. A book represents late-stage work, not the early-stage work of journal articles."
Thus Anurag Acharya, one of the founders of Google Scholar, on different kinds of reading, at the ALPSP Annual Conference in October 2015 (a presentation helpfully summarised by John Sacks in a Scholarly Kitchen post).
“You can easily see the difference between these two modalities. Do a search in Google for “San Francisco weather” and the answer pops up: the current temperature and conditions, and the forecast. But for the “weather scholar” there is also a list of sites below the forecast that you can go to if you want to study the topic by reading web pages about San Francisco weather.”
You can see why Google (and any other search engine) is keen to disambiguate natural language and to make searching more precise. But I’m not sure if the distinction between books and journals is as Anurag states it. I think the statement might hold true for scientific discourse, but not for much of the humanities, especially for the arts.
Do books provide a late-stage topic? What about historical works? They can just as easily provide a list of things to read. What about a book-length work of literary or artistic criticism? It is perfectly possible that a study of a major artist or writer might find their work full of contradictions and so leave the reader with questions rather than answers.
For searching purposes, the most obvious difference between books and journals is the lack of an abstract for individual chapters, or even, for many books, of an abstract for the book as a whole. Anurag Acharya is perfectly correct in calling for chapter-level abstracts, and for abstracts in intelligible language (as described in this post).
Another fundamental difference with books is the curious relationship of chapters with each other. One of goals of Google Scholar, and that of anyone involved in the scholarly communication process, is to try to facilitate the journey of the researcher to content related to their topic of interest. Over 20 years or more search engines have discovered that the combination of a search engine and journal articles (that is, articles typically between 3,000 and 10,000 words) can be quite precise and successful in retrieving related results. But searching in a book, typically between 70,000 and 100,000 words. is not only more difficult because of the overall length, but because of the lack of signposts. Of course, you might say that books were never designed to be read as individual chapters without having read the preceding content. But, like it or not, that is the way academic reading takes place - researchers, students, and many others as well, look at a book for a specific answer, not to read the book in its entirety.
Now, no author of a journal article would dream of referring to other articles in the same issue of the journal by “see next article” or “see last article”. But a book does just that. In other words, a book is not just a collection of several self-contained articles. As soon as you start to break up books by chapter, you run into problems. This is because the way books are structured is not as clear-cut as the way articles. One chapter may omit any reference to other chapters, but then again, may be full of links to other chapters. Even the notion of what a chapter is doesn’t have full consistency. Book authors sometimes use chapter divisions, sometimes section divisions, sometimes “part” divisions. There is no consistency.
How common is interlinking in books? Well, a search on Google Scholar for “as I stated in the last chapter” produced 90,400 hits when I searched on Google. Anyone retrieving that text will have problems working out what the last chapter was, or where the reference was. Here is an example selected at random from a textbook:
Now, if wages are flexible (as stated in the last chapter), then they will soon fall and the SRAS curve in Exhibit 2 will shift to the right. In time, the economy will ...
And talking of chapters, here is a challenge. Look again at the contents list displayed in the image at the top of this post. Can you identify the title of the book? Do you have any idea what the book is about? It is the contents list for Robert Hughes’ excellent history of Barcelona (1992). While not a monograph, it includes a bibliography and index, and would regard itself as a reasonable authoritative work. Yet the author feels no need to make any of the chapter headings intelligible to anyone who has not already read the chapter. To confirm the point, take a well-known academic history title – say, The Roman Revolution by Ronald Syme (1938). Some of the chapters have intelligible titles, e.g. “The Consul Antonius” or “The First March on Rome”. But what about chapter XXI, “Dux”? Or chapter XXII, “Princeps”? Whatever the function of these chapter headings, it is not a signpost to new readers. Trying to present individual chapters from books produces a guessing game rather than a thought-through solution to the scholar’s journey.