How to read non-fiction

Karin Wulf, a well-known history academic, who effectively straddles the divide between academic and popular writing, revealed her methodology for appraising an academic work in an interesting post, a process she calls efficient reading. Her system can be summarised as “TICCN”:

1. Title and structure
2. Introduction
3. Conclusion
4. Chapters
5. Notes

Does this method work? And is it different for digital reading? To some extent. Let’s be clear, at the start, what kind of reading we are doing here. In this case, I assume we are here reading to appraise the work, not to read it in detail: that comes later (if at all). Appraisal is just one stage in the process of interacting with content. There are perhaps three or four main stages to this process, of which I concentrate here on stage two:

1. Identify (you might notice, for example, 25 articles or books, from conversations, from reading, from reviews or wherever)
2. Appraise (you would then reduce the 25 down to 5)
3. Read in detail (perhaps 2 or 3 maximum)
4. Cite and/or critique (perhaps one critique emerges from all this reading).

I compared Karin Wulf’s practice with my method for appraisal, which can be summarised as “theme-criticism”, or “TC3”:

1. Look at the title, then read the first two or three paragraphs. What is the theme? Articles usually (but not always) reveal their theme in the title. But books often have a theme that is not explicit from the title, or even unrelated to the title. Write down the theme – it should be a maximum of a few words.
2. Read the first section – introduction for an article, first few pages of the first chapter, if a book, looking for:
   1. Evidence to support the argument. Check to see if the author has added attributions as necessary, and if any are missed.
   1. points where you disagree.
3. The theme might now have two or three subheads. Now you can start to list your criticism of the argument. Try to find three or four points, actually hypotheses, of agreement or disagreement; most importantly, ensure you find evidence for each. You may subsequently confirm or refute these claims, based on a lengthier exposure to the text, but it is highly valuable to play devil’s advocate at this point. This is essentially what you would do when you carry out a peer review.
4. Look at the chapter headings, for a book, or section headings, for an article. See if the headings affect the theme or send the content in a different direction.  
5. Read the conclusion (usually at the end of a book, but some begin with the conclusion). Unfortunately, many journal articles have a “discussion” section which is not a conclusion – this has to be inferred from the introduction.
6. Decide if the object is worth reading in detail. In my experience, I read in detail perhaps one book or article in ten that I appraise.

If it is a digital version, you can of course search for the theme or words around the theme to confirm your hypothesis.

I don’t make any claims for my own methodology over anyone else’s; my method is only one of several, and it strikes me that getting an overview of a book or article is a problem that has many solution paths. It reminds me of a review I read years ago on how to teach maths to teenagers. Some teachers insisted that maths problems had to be solved in a particular method, and spent hours trying to correct the students to use one heuristic. But, stated the article reassuringly, no method of solving problems is ideal. If your method works for you, that’s good. No need for us all to follow the same methodology.

Much more important is to engage with the content. Now, in the age of generative AI, it is possible to get an instant summary of an article or book with no effort; but you will have learned little in the process. By actively engaging with the text, and looking for errors, you end up with a much better knowledge of that text. It’s a well-known principle that teaching something fixes it in your mind, and so does constructively agreeing or disagreeing with it. I don’t say that AI-produced summaries are unhelpful. I can envisage using them to correct and refine my idea of the theme and critique; but I would be wary of relying on them as the only view of an article that I read, since I won’t have been able to engage with the author so effectively before I come up with my opinion.

To conclude, I have to kick myself to read actively, rather than passively. It helps to disagree with the text in front of you even if only to get your brain working. And that is the process that generative AI doesn’t replace.