The javaScript InfoVis Toolkit includes a number of pretty cool libraries that work in the browser and can be customised to your own needs. Some of these visualisations are specifically designed for trees and graphs, so I always wondered how a dynamic tree-rendering of Wittgenstein’s Tractatus would look like.

The Tractatus Logico-Philosophicus (Latin for “Logical-Philosophical Treatise”) is the only book-length philosophical work published by the Austrian philosopher Ludwig Wittgenstein in his lifetime. It was an ambitious project: to identify the relationship between language and reality and to define the limits of science. It is recognized as a significant philosophical work of the twentieth century.
[…] The Tractatus employs a notoriously austere and succinct literary style. The work contains almost no arguments as such, but rather declarative statements which are meant to be self-evident. The statements are hierarchically numbered, with seven basic propositions at the primary level (numbered 1–7), with each sub-level being a comment on or elaboration of the statement at the next higher level (e.g., 1, 1.1, 1.11, 1.12).

The final result is available here (warning: it’s been tested only on Chrome and Firefox): http://hacks.michelepasin.org/witt/spacetree

Some more details

I’ve played around a little with one of the visualisation libraries the JavaScript InfoVis Toolkit makes available, the Radial Graph, to the purpose of transforming the Tractatus text into a more interactive platform. The Radial Graph is essentially a tree-rendering library built over a circular area (hence called also space-tree).

I liked the idea of making the tree-like structure of the text explorable one step at a time, within a framework that suggests a predefined order of the text-units but also allows for lateral steps or quick jumps to other sections. However I’m still trying to figure out what the advantages of looking at the text this way can be, once you go past the initial excitement of playing with it as if it was some sort of toy!

Some of the pros seem to be:

On the other hand, here are some cons:

I wonder how all of this could be developed further and/or transformed into a useful tool.. if you have any comment or suggestion please do get in touch !
I’m also planning to release the source code for the whole app as soon as a I clean it up a little; for the moment, here is the javascript bit that renders the graph:

Prism is presented as

… a tool for “crowdsourcing interpretation.” The concept emerged from a decade-long conversation on categories of textual interpretation which took place at the University of Virginia, and specifically from transparency mark-up games designed by Bethany Nowviskie for her Media Studies students and her colleagues in SpecLab.

The concept of crowdsourcing interpretation is definitely a fascinating one; maybe the end result of Prism a little less so, at the moment. Simply merging different people’s highlights can be visually interesting, but at the same time not very useful from a scholarly point of view I think. Overall, it seems to me like an indiscriminate association of possibly very different viewpoints.

..the end goal of Prism is to produce aesthetic provocations, that is, visualizations which provoke further discussion. Prism expands upon current notions of crowdsourcing to allow for more meaningful interaction with “the crowd.” Users interact subjectively with a text and contribute to a collective interpretive energy that has infinite possibilities beyond the highlighting exercise itself

Despite the fact it’s still early days for Prism, it seems to me that this is potentially the beginning of a new research field (cf. this post too: Future possibilities for Prism), so I’m really curious to see how it’ll develop. I have the feeling that just by adding a few more controls (e.g. filtering highlights/annotations by users group, or mechanisms for tagging/categorising passages) the final visualisation of the text could become a really powerful tool for exploration and learning.

Think for example of a classroom scenario: groups of students could focus on different facets of an author, annotate what’s relevant with respect to that facet, and eventually explore what the other groups have annotated in the text (and how it overlaps with their own annotations). All of this could be achieved just by playing with the interactive controls of the final visualization .
Another interesting scenario could be a single-user context: when you study a text, it’s not uncommon to find new elements of interest each time you read the text, maybe because your goals have changed, or simply because you are in a completely different mood. Using an enhanced version of Prism it’d be possible to see the evolution of your notes chronologically, and maybe re-consider them in association to other contextual information you might have stored elsewhere (e.g. about other readings you were doing at the time).

In conclusion: will crowd-sourced interpretations change the way we read texts? We’re not really there yet, but definitely going in the right direction!

Related work