Jan 13, 2017

Visualizing Networks of Philosophical Influence

A recent piece in Daily Nous, A Visualization of Influence in the History of Philosophy, highlights a new visualization project for figures in the history of philosophy - Philosopher's Web, created by Grant Louis Oliveira.  I'll lay my proverbial cards on the table with this one immediately.

While the project has a very worthwhile goal, and could provide a much needed resource, in its present form it is highly inadequate as what it purports to be, for reasons I'll discuss below.  Worse, precisely because of its serious flaws, and because it is being presented as merely "incomplete" but on the whole a decent visualization, it will end up providing in some cases - and reinforcing in others - mistaken conceptions about the history and contemporary state of philosophy.

The crux of the problem is that ALL - yes all - of the information used in the project comes from one online source, English-language Wikipedia entries on philosophers.  Actually, it all comes from one section of those Wikipedia pages, the one summarizing "influence".  This isn't something new, by the way.  A precursor visualization project - Simon Raper's Graphing The History of Philosophy - did precisely the same in relying solely on English-language Wikipedia entries, and generated a visualization roughly on par in terms of its inadequacy to Oliveria's

Wikipedia Has Come A Long Way But Still Shouldn't Be The Source

Lest people mistake this as a typical professorial anti-Wikipedia screed, let me note from the start that I'm not a critic of Wikipedia as such.  I do remember the early days when the quality of the entries ranged from not bad at all to quite simply awful (giving occasion to humor such as this Professor Wikipedia skit), but those were a long time ago, really an entire academic generation past.

Eventually, while never recommending Wikipedia as a starting point or resource for research to my students, I came to be all right with their inevitable use of the site as a go-to.  I'd simply try to steer them to better resources - which often were listed near the bottom of the Wikipedia pages - told them to take the information provided in Wikipedia with the proverbial grain of salt, and didn't allow them to cite or paraphrase Wikipedia in research papers.

I do understand the attraction that Wikipedia exercises for non-experts.  It purports to gather all the world's information in one place.  It has the whole crowdsourcing, "wisdom of crowds", aura going for it, coupled with the promise of a rigorous editorial process.  Perhaps even more importantly - just looking at it as a medium - there's the sense of a seamless, standardized, comprehensive system.

The problem, however, is that unless you know about the topic - that is, have at least an intermediate level of expertise in it - or do the labor of comparing a Wikipedia entry to one from a better-researched source (in my field, some examples might be the Internet Encyclopedia of Philosophy or the Stanford Encyclopedia of Philosophy) - you're not able to determine much about the quality of the Wikipedia article.

Whatever people's hopes for Wikipedia are, and whatever merits there may be to their platform and its processes, when you do possess a high level of expertise in the history of philosophy - or even just small portions of it - many Wikipedia entries at this point are just not very good.  I'll give you one example that I looked at as a test case when examining the structure of Oliveria's and Raper's visualizations - Maurice Blondel.

If you look at that entry in English-language Wikipedia, you'll notice it is very short, and doesn't contain much information.  None of the information it does provide is technically incorrect, with the exception of the claim that his work "integrated classical Neoplatonic thought with modern Pragmatism" - there's a whole story that could be told about how both parts of that are wrong! - but what is there tells you very little about Blondel's vast network of connections (even after his "retirement") within the 20th century philosophical milieu.  The only names mentioned in terms of influence are Émile Boutroux and Henri de Lubac.

Now, compare that - if you read French - to the French-language Wikipedia entry.  Or the German-language Wikipedia entry, or the Spanish-language one.  (By the way, checking these other entries is as simple as clicking a language-link on the left side of a Wikipedia page).  You'll notice that there's a good bit more information provided than in the English-language entry.  Not a lot more, but some. You get a slightly fuller picture - not a perfect or even adequate one to be sure - but a somewhat better one.

What Else Would a More Adequate Visualization Use?

In answering this, one would need to take pains not to allow the best to become the enemy of the good, as the old saying goes.  Academic research in the history of ideas - when it is done right - inevitably opens up more rabbit holes than one would ever have the time to dive into and explore. And crowdsourcing research projects - even when that crowd is composed of acknowledged experts (which is not actually the case for Wikipedia) - typically still introduces its own blind spots, omissions, closed-off lines of discussion.  It's not a question of being absolutely and uncontroversially comprehensive.  Instead, it is a matter of mapping a network of influences that would be a good deal closer to what experts on those philosophers would say is more or less adequate.

Simply expanding the Wikipedia net more widely into the entries in other languages would be a start, but just that.  Again, to use the Blondel entry as an example, a number of names would be added to those who influenced Blondel - interestingly, the German entry is better in this respect than the French one - but there would still be some important names missing - Blaise Pascal for instance.  And there definitely would not be enough figures who Blondel influenced added to the list.  So where else might the researcher go?

Unfortunately, the Stanford Encyclopedia of Philosophy - which, unlike Wikipedia, is a high-quality, peer-reviewed online resource - does not have an entry on Maurice Blondel (though he gets a few mentions in some other entries).  Another excellent online source is the Internet Encyclopedia of Philosophy, though the article on Blondel needs to be updated and expanded.  How do I know that?  I wrote that entry, years back, and just haven't gotten around to updating it.  Reading back over it, there's not a lot there that will clear up the issue of who Blondel influenced, since I summarized that issue in writing:
Blondel's importance has largely been in theological and Catholic philosophical circles in France, Germany, Spain, Italy, and Quebec. Among many other important authors in the 20th Century, Blondel is responsible for the "new theology", that played such a great role in the deliberations and arguments of the Second Vatican Council.
Notice that there is a link there to another entry I authored, on the 1930s Francophone Christian philosophy debates.  Following that link and reading through it would indeed yield several names of figures who Blondel clearly influenced.  Notice that this would require additional follow-up, the kind of research that someone who doesn't work in the history of ideas or the humanities more generally might not realize needs to be done.  (And that of course, is assuming that no laziness, time-constraints, or other factors preclude following out those leads!)

Where else then would one have to go to get a more adequate picture of which other philosophers Blondel influenced?  Here's where we run into a dilemma, one that inevitably arises with these sorts of data-driven visualization projects that attempt to give us a view of incredibly complex intellectual and historical matters.  And it arises precisely because there aren't a set of standard sources that one might go to for the History of Philosophy itself - let along for the question of influence - available online.  There are online sources to be sure, that would provide that needed information, but you'd have to do a lot of work, informed by some degree of expertise, to find, vet, and draw upon them.

The sources that are available, easily accessible, and which would appear - key term there, appear - to be more or less comprehensive - turn out to be decently informative about some of the figures they are supposed to inform one about, and uninformative about many others.  To acquire a fuller, much more adequate range of information about which philosophers influenced each other - or even, in many cases, about who was a philosopher of some importance - requires going to a much wider range of sources, wading through them, compiling and comparing information.  It also tends to require the expertise required to tell the chaff from the wheat, the bullshit from the real information.

So, either you go with the easy source - acknowledging that it makes one's visualization a bit incomplete (since you've got no idea just how incomplete it really is) - and now you've got a manageable project that requires no real expertise in field one is visualizing.  Or, you do the hard work - the first bit of which is planning out how one might feasibly acquire a much fuller, more complete, and more accurate range of information - and then produce a much better, less misleading, more adequate visualization.

Why Call The Visualization Misleading?

That is certainly a valid question to ask.  Granting that the visualization draws upon and represents data that is for some figures less than entirely complete, and for other figures clearly inadequate, why not just say that the visualization is a work in progress, the best available at the present time, and understandably incomplete?  Why go further and call it "misleading"?

The main reason is that as I've pointed out above, only those who actually know quite a bit about a given topic will have any sense of how incomplete Wikipedia is as a source - let alone how selectively incomplete that site is at present (and is likely to remain in the foreseeable future). Absent that understanding, I suspect that the average, literate, educated person is unlikely to see anything off when looking at the visualization, or when reading (and reposting) about what looks on the surface like a great new resource for philosophy.

And don't get me wrong - I think this indeed could be a wonderful resource for philosophy.  It would just require a much more complete set of information than what is currently being used.  But then, you run into that dilemma I mentioned above.  Absent anything like that, this is about as good as it gets for visualizing influence relationships across the history of philosophy.  And when "as good as it gets" is all right in some respects and quite poor in others, for whatever good it may do, it also has an effect of reproducing intellectual mediocrity, and reinforcing ignorance about the field of philosophy.

Artifacts like this in the digital age don't simply reflect or replicate - they reproduce and reinforce - in this case, whatever good and bad points, whatever inclusions and exclusions, the Wikipedia entries possess.  The Daily Nous piece, and the visualizations, will be shared and reshared - along with the verdict "a work in progress. Impressive work, nonetheless!", or variations thereof.  In the process, the Philosopher's Web visualization will come to be presented by at least some as a definitive representation of this key aspect of the History of Philosophy.

And that's unfortunate, since for those who aren't already well-informed about at least a portion of the History of Philosophy - or better the broader History of Ideas (since there are also a number of omissions of influential thinkers just based on how Wikipedia classifies figures, whether as philosophers or as something else) - if they take their cues about who is worth focusing upon in the limited time they possess from this visualization and from similar sources, they'll never suspect what they're missing out on.

Two things to say as last notes.  First, though I used one particular figure as a test-case and an example to illustrate the weaknesses and gaps in Wikipedia and Philosopher's Web, there are hundreds, perhaps thousands of other omissions scholars could bring up by way of criticism.  A few have in the comments section of the Daily Nous piece.  Second, I really do love this idea of visualizing relationships of influence (though that notion "influence" requires some clarification as well).  I'm hoping sometime in the future to create visualizations, much less ambitious in scope, but much more adequate in content, for domains such as the existentialist movement, or the people involved in the 1930s Christian philosophy debates.