Task: Analyzing Essay Structure

(Thanks to @archurhh for the inspiration here.)

Let’s suppose that you have a hunk of writing: a John McPhee article perhaps, or something like Emerson’s The American Scholar.

Your task is to read through the text and to learn how it is put together — to understand and explain the structure. You might do this by building an outline or a map, as you prefer.

I think it might be interesting to discuss exactly how one might approach this, either for one’s own use or as a student assignment.


I think I’d start to read the table of contents — especially if it has been made with Latex and printed in a very good font, but it is off topic — and probably use the Outline View to mirror the outline of the book I read. I’d have as many containers as the book has parts, chapters, sections and I’d create a note everytime an idea would retain my attention. I’d use Tags and the Attribute Browser as a way to view my notes from different perspectives. Maybe would I create an agent to gather some specific notes about an important topic.

Fair enough for a conventional textbook. But that might not be what you have to work with.

Suppose, for example, you want to dissect “ L’existentialisme est un humanisme* Existentialism Is A Humanism — not just to read it for comprehension, but to look closely, to see how it’s put together. In school, they told me an essay looks like this:

  • Thesis
  • Objections
  • 3 Supporting Arguments
  • Conclusion

But Montaigne didn’t do that, and neither does Sartre, or at least not quite that. And even a superficial glance suggests some secondary structures — for example, provocations about Christian critique in one paragraph seem often to set up refections on Christianity a few paragraphs later.

But just mechanically, Sartre’s doesn’t give us a table of contents. We’re going to have to extract the framework ourselves.

OK. In fact, puzzling out a table of contents is a common exercise in philosophy. There is a reason why a student is encouraged to do it: the table of contents is in principle builded in a logical way, in a way that demonstrates a main thesis. That’s the reason why I’d try first to mirror the structure of a book in the Outline View, in order to see the logic behind the titles. But, you’re right: there are so many ways to defend a thesis. Reading Derrida’s works is enough to be convinced of this point. Therefore, maybe in a second phase, following your suggestions, I think I’d try to identify the underground strategies used to argue a thesis or to serve a demonstration: examples, thought experiments, hypothesis, controversies, and so on. Doing this, I’d have a lot of categories that the Attribute browser could easily display.

Personally, I woudl start with the “first principles”:

  1. The outcome, what outcome do I want to achieve? Is it a paper, presentation, talk, or simply> I’d want to have a loose idea on the output structure; I’d trust that I could produce any template want to get there, but I want a general guide. In addition to the primary objective, I’d keep in mind that I want to build atomic assets (i.e. standard alone notes) that will help me for future projects.
  2. The elements, what elements do I think I might need (e.g., people, terms, concepts, places, events, notes, media, entities, pages), I’ll turn these into attributes
  3. The attributes, what attributes do I thnk I’ll need, e.g., $Terms, $People, Concepts, $Places, $Events, $CitationKey; as well as, linktypes. I’d add these to the attributes
  4. Action Code and templates, I’d then set up my action code that will enable Tinderbox to automatically generate the element notes as I populate the various related attributes as I’m doing my reading. This will help me get the manual work out o the way. I’d pull form my existing files and efforts to do this.
  5. Execute and Incremental Formalization, I’d then start reading and following where the winds take me. Mostly I’d use an outline, but I’m right to pop into Map or the other views as the need arises. I’d start to explore and increment my elements, attributes, action code, and templates along the way. I would keep all media (images and video) external to TBX, as I can pull them in through attributes and value. If the visual in TBX is really necessary, I’d use $HoverImage and $Fill. I would heavily leverage linked notes and the tear-off window feature so that I can stay in context of whatever topic I’m writing about or taking notes on, rather than wasting time navigating all around the file. If I had original source text in Tinderbox, like @archurhh’s example, I would also use the .highlights operator to extract highlight text.
  6. Side trips, I’d expect to experience some side trips, and speedbumps, like the recent non-breaking experience issues that I raised. I’d look forward to these rather than seeing them as an issue. They are a health distraction and help me learn the fundamentals so computing and knowledge management in the modern age; I’d trust that the learning I gather, if not immediately useful, will be in the near future. When in doubt, I’d leverage my Sangha (i.e., the TBX Forum), my Buddhas (@eastgate, @mwra, forum, and backstage members, and others in the community), and my Dharma (aTbRef, TBX Forum, Backstage).

Most of all, I’d go in with a beginner’s mind and remind myself to have fun!

Breakdown of Draft No4 - ver01.tbx (522.2 KB)


The structure of any single piece of writing might look quite different to different specialists. A philosopher’s perception of an article on philosophy is likely to be different from that of a social psychologist with an interest in social constructionism, for example.

When I did some textual analysis years ago, it was all about pencil and paper. Software (such as Nvivo) was far too laborious and not flexible enough for my needs. Even with other tools available nowadays, I think I’d still be using pencil and paper for much of the work. But I’m glad to say that I’m no longer doing such work. :grinning:

Great! In the thread that inspired this thread, you described a mechanic in some other outliner — Obsidian? — where you could refactor a stream of text into an outline by dragging elements upward.

Is that what you did here?

Not exactly; for the example file here I worked somewhat reverse from the process I was describing, as it’s more of a breakdown/analysis of someone else’s work. Tinderbox really excels at this - my breakdown was limited only by the extent of my Tbx knowledge and the time I spent on breaking down the article.

In contrast, for ‘upward outlining’ - taking the case of the outline I demonstrated in this other post - the raw text block that originated that outline might’ve started out something like this, a stream-of-thought musing:

I have some clarity in how I think about the moon, and how it relates to planets in general. I sense a metaphor in here that I’d like to flush out. Hmm, let me think about whether and how the potential metaphor might scale. Is there a correlation to our solar system vis-a-vis the galaxy? Are we (as in the earth and other planets) simply circling some other larger planetary body? Is that larger body itself simply circling something bigger, and so on ad infinitum? And then everything circling a black hole, that swallows and digests us and centuries later, spits us out as just that much planetary graffiti?
And how does this thought scale downward? Do we - individuals, communities - tend to circle other more prominent individuals, shadowing/eclipsing/swallowing others in our zeal? Do those prominent individuals themselves, seeking a greater significance, circle more prominent individuals who in turn orbit black holes (myths, legends, beliefs)? What does it all mean? Am I just sleep-deprived tonight? Or hungry?

From this musing, I might then extract a variety of metaphors and analogies; as well, I might do a bit of web-browsing and/or scanning my existing Tinderbox data set for other bits and pieces I might want to glom on to the initial thought. Therefore, it would be a process of adding clay as I intuit exactly what will go into the final design. Along the way, I will naturally generate excess - which itself may be useful in a different project; this excess is squirrelled away in Tbx Notes and suitably annotated for future access and deployment. The process of ‘outlining upward’ is that effort; it’s a sort of incremental formalization, but also incremental chaos creation to some extent.

Hope this clarifies.


Asked of that particular McPhee essay, this becomes a trick question.


I see what you did there

Zounds! Detected!

1 Like