Reading for the Structure

Thinking more about the problem of reading philosophy, which I raised yesterday, I realize that there is a parallel with the issues of grammar that I raised when I was trying to develop my argument about "accountability to the music itself." In that post I talked about the need for structural explanation in both performing and listening to music; but this reinforces a position that I have taken for many years that music is a text in the same sense as the "garden variety" one we apply to the books we read. In yesterday's post I talked about structural explanation in terms of sorting out the embellishing from the embellished; but, where many difficult texts are concerned, it may also be an issue of segmentation. Let me try to say a bit about each of these perspectives.

The philosopher who probably had the best appreciation for the distinction between the embellishing and the embellished did so in only one text, which is basically the only completed text he ever wrote. The philosopher, of course, was Ludwig Wittgenstein; and the text was his Tractatus. Any other "book" with his name as author is basically a collection of notes, either his own or those of others; and, while they are excellent examples of Kleist's concept of "the gradual fabrications of thoughts" that I cited yesterday, none of them are in a form that Wittgenstein would have felt was suitable for publication. The Tractatus, on the other hand, uses an explicit hierarchical numbering system to establish the embellishment relationship across the individual passages. Given my interest in such structural hierarchies, I am a bit embarrassed to confess that it took Ray Monk's book about reading Wittgenstein to point out to me that the best way to read the Tractatus is in a top-down (breadth-first) "tour" of the paragraphs. In other words, if you begin by reading the "single-digit" passages (1, 2, 3, 4, 5, 6, 7) in order, by-passing everything that comes between them, you get a text that would pass for the abstract of an extended technical monograph. However, every one of those passages requires further elaboration. Passages 1.1 and 1.2 provide that elaboration for Passage 1. After reading them, you can decide to see where Passage 2 is going or how Passage 1.1 is elaborated. You, as reader, can make this choice on the basis of where you want to focus your attention. Whatever tour you decide to take, however, this strategy is likely to serve you far better than beginning at the beginning, going on until you reach the end, and then stopping (with apologies to Lewis Carroll).

There may not be any other philosophers who reveal the organizational structure of their text quite so explicitly, but some put enough effort into the Table of Contents to allow it to serve a similar purpose. Kant seemed to be rather good at using his Tables of Contents as "road-maps" to the actual text; and it seems as if he would approach a subject by first "mapping it out" and then providing the text for all the "regions on the map." As anyone who has read one of the Critiques knows, however, Kant made up for the simplicity of his road-map with the complexity of his sentences! This is particularly evident when a translator feels it is important to be true to Kant's linguistic style, which seems to boil down to the fundamental strategy of cramming as much as he can into each sentence. (Yes, I know I can be accused to the same stylistic problem.) This is where real grammar (as opposed to my attempt to talk about music at a grammatical level) enters the picture. The best way to deal with those sentences is to diagram them: Start with the basic relationship between subject and predicate and then sort out all the parenthetic remarks and adjectival and adverbial modifications. A typical sentence of Kant demands several passes to do this; but a "tour" of one of those sentences is not that different from the "tour" of Wittgenstein's Tractatus. One you get hold of the "primary message," you can then set your own agenda for dealing with the rest of the text.

This kind of mapping also addresses the problem of segmentation: the segment boundaries are equivalent to the "region boundaries" on the map. Not all texts are this accommodating, however. By assuming the style of the prolonged conversation, in the course of which the lessons reveal themselves, Plato saw no need to provide his reader with any explicit segment boundaries. However, most of these texts are too long for a single sitting; and sometimes you need to take a break just to reflect on what you have been reading. Fortunately, there are no end of scholars out there who can provide segmentations of Plato's texts; but you can also view the search for those boundaries as part of the reading experience. I tend to prefer the casual summaries I can get from someone like Edith Hamilton, which serve me in rather the same way as the single-digit passages in Wittgenstein's Tractatus; and, knowing no more than that "bird's eye view," venturing into the text looking for those transitions that take me from one stage of the argument to the next. This is particularly nice because some of Plato's dialogs are actually quite dramatic, framing the argument itself between a prologue and an epilogue, which sometimes turn out to be rather tragic reflections (as on the death of Socrates).

Let me close with two remarks, both of which I may explore in greater depth in later posts:

1. While I chose to focus on philosophy, I suspect that most good texts out there benefit from a similar structural approach to reading. Indeed, that kind of structural comprehension seems to lie at the heart of what is called "explanation" in the hermeneutics literature. I would even go so far as to suggest that the strategy works on fiction as well as it does on non-fiction.
2. On the other hand, there are plenty of texts out there that resist such an approach to reading. Many government reports (the longer the better) are good examples. I might even want to skate out onto thin ice and call these non-texts. My argument would be that these texts were not meant to be written but only for the sake of creating an archival record, just like the transcripts of congressional and judicial proceedings. Texts like these will not benefit very much from good reading strategies and may not benefit from much other than good search engines!