Now, draw me a picture of the good life.
June 28, 2012 1:12 PM   Subscribe

"Descartes is smaller than you’d think" in this graph of the history of philosophy

Statistician Simon Raper has used Wikipedia and the open source graph visualization software Gephi to do what took sociologist Randall Collins 25 years to do in his book The Sociology of Philosophies, that is, map the relations of influence in the history of philosophy. Via Drunks&Lampposts.

Previously, related to Collins' efforts. Speaking of Collins, here is his blog, The Sociological Eye.) (And I would hope it's obvious, but if not, I don't think Raper's graph replaces Collins' work which I find fascinating.)
posted by mc2000 (63 comments total) 20 users marked this as a favorite
 
The data source is the "influenced by" sections in their Wikipedia articles' infoboxes? And we're meant to take this seriously?
posted by RogerB at 1:16 PM on June 28, 2012 [8 favorites]


I agree that this is only as illuminating and reliable as the input data, which in this case is Wikipedia entries. But a neat graph. Would be cool to see this done for an input data set on philosophical influences prepared and vetted by experts.
posted by LobsterMitten at 1:23 PM on June 28, 2012


Joseph de Maistre is larger than Augustine of Hippo. A little bit hard to take seriously.
posted by pdq at 1:23 PM on June 28, 2012 [3 favorites]


Which is not to say that it's not quite interesting. I bet it would be more accurate if they figured out a way to get data from the Stanford Encyclopedia of Philosophy.
posted by pdq at 1:24 PM on June 28, 2012 [1 favorite]


Where's Zizek?
posted by Golden Eternity at 1:33 PM on June 28, 2012 [2 favorites]


The data is easy to absorb programmatically from wikipedia, and it's high level data that doesn't (in my semi-expert opinion) require expert opinions to get 'right enough'. Wisdom of the crowds models in sports simulations do well. Not quite as well as the best expert modelling, but well. I think this is an analogous case. Is there a realistic better suggestion for the data?

As for how seriously we're meant to take it, I imagine that no one should base their history of philosophy comprehensive exams on it, but it's worth taking at least slightly seriously, as seriously as we generally take wikipedia. How seriously do you think we're supposed to take it?

Cool post, thanks, sorry these comment threads are a race to say how the thing sucks so often these days.
posted by Kwine at 1:36 PM on June 28, 2012 [14 favorites]


Ah. There he is by Freud and Derrida. Pretty small compared to Chomsky.
posted by Golden Eternity at 1:36 PM on June 28, 2012 [2 favorites]


If this was accurate it would be one big circle with Aristotle written across it, surrounded by a bunch of dots.
posted by Nahum Tate at 1:39 PM on June 28, 2012 [4 favorites]


For anyone trying to view that big transparent png in Firefox's latest version of their image viewer (with the black background), there's an add-on you can install that will return the white background to the image viewer.
posted by Thorzdad at 1:42 PM on June 28, 2012


I don't think anyone's saying it sucks. In fact most people have said it's neat or cool but has limitations based on the data, which the linked site itself notes.

The point is, if there's a surprising feature that we observe by looking at this graph (for example that Descartes is not as big as we might have thought), there are two possible explanations:
1. it turns out he isn't as influential as we might have thought, or
2. the data misrepresents how influential he is.
All I meant to be saying is, let's be careful not to jump to 1, because 2 is very much a live option here.
posted by LobsterMitten at 1:42 PM on June 28, 2012


Obligatory
posted by Chuffy at 1:45 PM on June 28, 2012 [2 favorites]


it's high level data that doesn't (in my semi-expert opinion) require expert opinions to get 'right enough'.

Have you looked at those "influence" slots in the infoboxes? They're often bizarrely arbitrary, and perhaps the least carefully written and maintained parts of the articles. There's no clear criterion for inclusion or exclusion, they're routinely spammed by cranks promoting their pet minor thinkers, and they're not as carefully discussed or vetted even to the patchy standard that Wikipedia articles themselves are. As a consequence, they convey approximately no information. It's as bad a data set as you could choose.

Is there a realistic better suggestion for the data?

Why would anyone not at least use the articles as the data source? Links and name mentions within the articles would yield a much richer and more reasonably weighted set of links.
posted by RogerB at 1:45 PM on June 28, 2012


Then it should be very easy to find spots where the chart has gotten things grossly wrong. Can you find those spots? I haven't found any yet, and I'm pretty knowledgeable.

I'm sure that there are spots where it's wrong. But the difference between the wrongness of it and any idealized data set you can name has been drastically overstated for the kind of task we're talking about-at least, it has been if it's anything like wisdom-of-the-crowd models in other fields. We're not flying rockets into space here, and we'll be closer than you think. At least, that's where I stand. I sent this along to Kevin Scharp, mentioned in the previously link. If he agrees with you I'll change my tune.
posted by Kwine at 1:52 PM on June 28, 2012


While the data set is questionable, this is interesting because it is a stunning example of why there are so many semantic web evangelists. Being able to make these kind of connections (and have each visualization be scaled according to the magnitude of each datum) is what this whole web thing is about. Obviously, like anything Wikipedia-related, don't use the data for anything mission-critical (or even mission-mildly-kinda-important-sorta).
posted by nonreflectiveobject at 1:55 PM on June 28, 2012


(LobsterMitten, RogerB is saying it sucks and that's who I'm talking to. I should have been clearer, sorry).
posted by Kwine at 1:55 PM on June 28, 2012


Then it should be very easy to find spots where the chart has gotten things grossly wrong. Can you find those spots?

What would it mean for the chart to be wrong? It's such a mess of overlapping lines that the links are illegible; only the clustering and the size and color of the circles are clearly legible even in the full-size PNG.

And no, I'm not at all saying "it sucks," I think it's neato — as a quick and dirty visualization. But its data source sucks and therefore it's not worth taking more seriously than that: if the picture surprised me in any way (e.g. Descartes's circle's smallness, which seems pretty "grossly wrong") I'd pretty much default to assuming it was bad data, not a new revelation about intellectual history.
posted by RogerB at 1:59 PM on June 28, 2012


To cut a long story very short I’ve extracted the information in the influenced by section for every philosopher on Wikipedia and used it to construct a network

This is neat and pretty, but I wish he'd said "every Western philosopher," since Confucius would loom at least as large as Aristotle, as would Lao Tzu and Buddha (assuming you are not defining him as a "theologian" rather than a "philosopher"). I'll forgive the omission of Lao Tzu on the grounds that he may be imaginary.
posted by GenjiandProust at 2:02 PM on June 28, 2012 [4 favorites]


Hannah Arendt's WAY too big. Nailing Heidegger only counts for so much.

wait, is that Marshall McLuhan?

Really?
posted by leotrotsky at 2:12 PM on June 28, 2012


This is interesting in the "something to debate over a pint of beer" sense.
It is completely useless in any other sense.
posted by Flood at 2:12 PM on June 28, 2012


You haven't looked at the svg? You can download the svg file and zoom in on it. That would be a great start on beginning to sort out whether your claims about the graph have any merit.

...

I'll let you all think of your own 'limited data source' joke.
posted by Kwine at 2:22 PM on June 28, 2012


Wikipedia articles' infoboxes? And we're meant to take this seriously?

WP's been around for 10 years now. If -philosophers- haven't vetted their own thoroughly, can life have any meaning? Or are we doomed forever to float like pond scum on the shores of obliviOn?
posted by Twang at 2:23 PM on June 28, 2012 [1 favorite]


You haven't looked at the svg? You can download the svg file and zoom in on it. That would be a great start on beginning to sort out whether your claims about the graph have any merit.

The links are no more legible in the SVG version than anyplace else, no matter how zoomed-in you want to get. There are too many thousands of little lines crossing each other and passing behind many circles on the way to their destinations; it's often quite impossible to see where the lines begin and end. I don't see how this is even debatable.

What is your point, anyhow? The "wisdom of crowds" that you're defending says Derrida was "influenced by" Artaud but not Austin, and Wittgenstein by Spinoza but not Descartes or Hume. If you're arguing that it's a pleasant surprise that anything even approximately meaningful is produced based on that quality of input, I agree. Otherwise, what are you actually trying to communicate here?
posted by RogerB at 2:32 PM on June 28, 2012


P.S. Re WP I see plenty of articles that have needed simple fixes for several years. Surely anyone seasoned in the fires of MeFi could repair them. And yet they persist, as do the complaints. Moral of story?

P.P.S. Re Aristotle: Is that not the personage who avowed that arrows fly because the air pushes them? I believe it is.
posted by Twang at 2:41 PM on June 28, 2012


Murray Rothbard? Seriously?
posted by steambadger at 2:43 PM on June 28, 2012 [4 favorites]


Haha look at Leo Strauss stalking Plato. "Get away from me dude, ugh"...
posted by Potomac Avenue at 2:45 PM on June 28, 2012


Meanwhile the chart is missing the two most famous and important friends of wisdom of all time, for the mere crime of never having written a book. Publish or perish I guess! And if you spin the globe around there's a whole other continent of Asian philosophy where Confucius and Tolstoy look like warring continents.
posted by Potomac Avenue at 2:46 PM on June 28, 2012 [1 favorite]


I shrink, therefore I am.
posted by Decani at 3:00 PM on June 28, 2012 [2 favorites]


Seconding the "Murray Rothbard" WTF reaction.
posted by sy at 3:11 PM on June 28, 2012


James Joyce? Yeah, Marshall Mc Luhan?? Lots of WTF and no Buddhists.

But it is fun to see this from a high level, and maybe be reminded there's whole areas and worlds of thinkers you forgot, or never heard of.
posted by stonepharisee at 3:20 PM on June 28, 2012 [1 favorite]


"Then it should be very easy to find spots where the chart has gotten things grossly wrong. Can you find those spots? I haven't found any yet, and I'm pretty knowledgeable."

The absurdly large size of Strauss's circle is a very convincing counterexample as far as I'm concerned. It's sad.
posted by Ivan Fyodorovich at 3:23 PM on June 28, 2012


So ... Wittgenstein was not influenced by Augustine or Ramsey (although Ramsey influenced and was influenced by Wittgenstein according to his Wikipedia page). Quine didn't influence Goodman or Strawson or Follesdal or Harmon or Stich (who isn't even on the graph, as far as I can tell). Peirce's entry has no "influenced" or "influenced by" categories. The Vienna Circle is itself a philosopher. Descartes' entry under "influenced" reads "Most philosophers after including:" and then gives a pretty bizarre list. And who is Murray Rothbard and why is he more connected than Bertrand Russell?

Sheesh. And those are really just the first few I felt like checking.

The graph is pretty. And also pretty useless.
posted by Jonathan Livengood at 3:46 PM on June 28, 2012 [2 favorites]


Crap. Should be "Harman" with an "a".
posted by Jonathan Livengood at 3:47 PM on June 28, 2012


Also, Confucius is on the graph. He's a small dot on the far left-side, not too far from that great medieval philosopher Scholasticism. I am not joking. Sigh ... I'll stop now.
posted by Jonathan Livengood at 3:58 PM on June 28, 2012 [2 favorites]


Kwine, as others have pointed out in this thread, the thing is riddled with, if not errors, then really bizarre connections and philosopher "sizes" that are stunningly different than what one would expect (Augustine of Hippo, for instance).

The problem is, echoing LobsterMitten, what use are these surprising connections and assessments? Are they really showing us truths about connections between philosophers and their influence on each other that goes against what many believe? If so, it's fascinating. But if these odd results are due to the use of a flawed data set, then these really aren't revelations at all and don't really tell us anything useful since they don't depict reality.

It's a very interesting exercise, at the very least.
posted by Sangermaine at 4:20 PM on June 28, 2012


For the very lazy like me, Murray Rothbard is apparently a libertarian economist.
posted by LobsterMitten at 4:21 PM on June 28, 2012


Also, Confucius is on the graph. He's a small dot on the far left-side, not too far from that great medieval philosopher Scholasticism.

That may be even worse. Now they can't hide behind "Western Philosophers...." Unless, I suppose, the set is still Europe-centric, since Confucius has had relatively little impact in Europe and was influenced by, I am willing to assert, virtually no Western philosophers....
posted by GenjiandProust at 4:24 PM on June 28, 2012


Okay, I can't stay well enough alone. In the entry for Kant, under "influenced" you see that Kant influenced "virtually all later western philosophy" and then a big list. That's fine, except that it leads to the graph adding a node for, you guessed it, Western Philosophy.

I've seen at least six nodes that are not philosophers by any stretch of the imagination (since they are schools or groups or movements):

Western Philosophy
Scholasticism
Vienna Circle
Frankfurt School
Neoplatonism
Pre-Socratic Philosophy

And I suspect there are more that I just haven't spotted yet. Now, maybe those extra connections don't make much difference. I don't know. I can't tell if the size of the nodes takes into account both those influencing it and those influenced by it or only the latter. And I can't tell how the size is produced. It would be nice if there were a key of some sort that indicated roughly how size corresponds to influence.
posted by Jonathan Livengood at 4:46 PM on June 28, 2012 [1 favorite]


Otherwise, what are you actually trying to communicate here?

Work intervened (and continues to), but I am responding to the thoughtless snark in the very first comment, an exemplar of a continuing problem in Metafilter threads.

I am happy that you eventually looked seriously at the graph, though I think that it's thoroughly fair to say that Wittgenstein was influenced by Spinoza but not Descartes or Hume. I don't know anything about Derrida.

Most of what Jonathan Livengood says seems right to me. The data needs to be massaged. At the least, the non-people nodes should be culled and the Descartes node needs an overhaul-the way that his influences are listed relative to others. But I think that this dataset is THE place to start from a 'getting the project off the ground at all' perspective. I'll certainly back off my claim that the wisdom of the crowd approach is good enough in this case. Whether that makes the project useless or not to be taken seriously is a different question that depends on a lot of other factors. That we default to useless and not seriously is a common problem with the culture here.
posted by Kwine at 4:50 PM on June 28, 2012 [1 favorite]


I would like to see a weighted graph of all wikipedia articles based on all the links in them, not just the side bar.

I nominate someone with more than 4gb of ram to do it.
posted by jonbro at 4:52 PM on June 28, 2012


also maybe weighted graphs like these are really disguised mechanisms for us humans to see which wiki articles need editing.

I am reminded of the History flow tool.
posted by jonbro at 4:56 PM on June 28, 2012


For me the most underwhelming aspect of this - and I agree there is something tiresome about enumerating the ways in which this sucks - is how it sets up an expectation of awe and wonder - something about the thrill of big data and cybersomething and rhizomatic whatever - when there is nothing to wonder about. It's just a standardized tool belching out graphs in a way that has inspired awe and wonder in the past. That's not to say that the map is worthless or uninteresting, but what is incongruous is that while the map derives its appeal from the subject matter, the subject matter in this case seems wholly subordinate to the process imposed on it.
posted by deo rei at 5:22 PM on June 28, 2012 [2 favorites]


Kwine,

If we wanted to do the project seriously, I think we would have to have some starting ground rules about which philosophers get to play the game. The way this one is done, you get some philosophers -- I'm looking at you Nietzsche -- who get way over-valued because they influenced a large number of very minor philosophers. I will grudgingly say that Nietzsche is a big deal, but he shouldn't seriously be in the same conversation as Descartes or Kant.

The problem, though, is that once you start pre-selecting the people who get to play the game, the project has a lot of potential to go off the rails. And it can do so in lots of ways. For example, Isaac Newton shows up in the graph (as do Shakespeare, Einstein, and Darwin). Should Newton count as a philosopher? I think the answer is yes, but lots of professional philosophers give me the old incredulous stare when I say that. And if we include Newton, shouldn't we include Boyle, Laplace, Maxwell, Feynman, and so forth?

Another problem comes in when you ask how much influence one philosopher had on another. Plato's influence on Aristotle? Enormous. Very different from, say, Parmenides' influence on Aristotle. But in the graph, both get counted the same. Similarly for contemporary philosophers. Carnap's influence on Quine was huge. William of Ockham's influence (yes, he's counted as an influence by the Wiki standard)? Uh ... shaky at best. And it seems worth noting things like which philosophers had large impacts on later philosophers who were really good in their own right, which leads me back to Nietzsche, I guess, so I'll stop there.
posted by Jonathan Livengood at 5:34 PM on June 28, 2012 [6 favorites]


"Descartes is smaller than you’d think"

This is interesting because all my previous estimates of his size were based on this portrait of Rene Descartes, which is only 30 inches high. I suspected that Descartes was actually quite a bit bigger, but now I wonder.
posted by twoleftfeet at 6:08 PM on June 28, 2012 [4 favorites]


It really diminishes the time-based component of influence. A modern philosopher has inevitably been influenced by Plato, Aristotle, and Descartes; he can't help it. It happens, purely because the people who directly influence a newcomer were themselves influenced by earlier philosophers. There is latitude here—a devout Catholic is more likely to be influenced by Augustine than Spinoza—but influence grows by accretion as well as by direct action. (Of course, I can still go back and read Augustine and Spinoza and be influenced by them afresh, which also contributes to their influence.)
posted by sonic meat machine at 8:09 PM on June 28, 2012 [4 favorites]


Rawr, nice Naqshbandi island to the North. I wouldn't have guessed Ahmad Sirhindi would show up bigger than Imam Al-Ghazali, but I guess they didn't call him Mujaddid Alf ath-Thani for nothing.

Maimonides and Rene Guenon shown without Islamic philosophy connections at all? Guenon died a Muslim!
posted by BinGregory at 9:33 PM on June 28, 2012


All the philosophers I remember studying in college showed up with big bubbles (and I was a philosophy undergrad).
posted by Veritron at 11:04 PM on June 28, 2012


The data source is the "influenced by" sections in their Wikipedia articles' infoboxes? And we're meant to take this seriously?

I take Wikipedia entries on philosophers more seriously than I take these graphs.
posted by John Cohen at 12:26 AM on June 29, 2012


It really diminishes the time-based component of influence. A modern philosopher has inevitably been influenced by Plato, Aristotle, and Descartes; he can't help it.

This was bugging me, too. In the interest of trying to say something constructive (rather than just "this is dumb because the data is dumb*"), how could this be modeled? Would a 3D structure, with time "receding into the background," work? So, Aristotle (and Confucius) have huge spheres, but ones that look much smaller from the "head on" view, since they are thousands of years in the past and much of their influence is the filtered influence of their thinking on other, later philosophers.

I agree that the somewhat random use of schools is also a distraction, but there is certainly a place for that -- the general impact of the paradigms provided by, say, Catholicism, Confucianism, Buddhism, Nihilism, and Neolatonism, exists outside of the impact of any given proponent of that philosophy. Zhu Xi, for example, strongly rejected Buddhist and Daoist ideas in an effort to purify Confucianism (and more or less created his own school, both like and unlike "pure Confucianism") -- it's hard to say that Buddhism and Daoism (especially in their more mystical/religious aspects) weren't an enormous influence on him and all the political philosophy of eastern Asia that followed. Is there a way to include the impact of paradigms in a graph like this, maybe as a removable "layer" that could be toggled on and off? Or better use of color (or texture) to indicate affiliation?

Answering some of these questions might make the whole project more useful, especially if a better source of data can be found.

*Which is a valid criticism, but not really a great starting point for a conversation.
posted by GenjiandProust at 4:29 AM on June 29, 2012


The problem here is bigger than most people have realised: it's that most historians of philosophy no longer accept 'influence' as a valid criterion for writing the history of philosophy:
The assumption that successful explanations can be constructed out of tracing influences has tended to result in histories both of ideas and events written merely in terms of the biographies and alleged impact of a conventional canon of leading historical figures. The appropriate perspective for explaining the politics of an age is thus taken to be the biography of its leading statesman; for explaining the political or scientific or philosophical thinking of an age it is taken to be the linked analysis of the most important and influential texts. The approach is misleading not only for the obviously arguable reason that to use biography as a methodological category is naive or at least partial without some consideration also of more general social, economic and ideological conditions; it is also misleading simply because it involves an identifiable mistake. The mistake lies in supposing that the history of an idea or event can ever be adequately written in terms of its leading actors. (Quentin Skinner, 'The Limits of Historical Explanations', 1966; my emphasis)
Or as Skinner put it more pithily some years later: 'at best the whole business of studying influences is nothing more than a scholar's game; at worst, moreover, it clearly leads to the assertion of many claims which there is no reason to suppose are true, and which are very likely false'.

So the problem with the map is not just that it overstates the influence of one philosopher and understates another. It's that the whole concept of a 'map of influence' is inherently flawed. Tweaking the underlying data is not going to make the slightest difference.
posted by verstegan at 5:17 AM on June 29, 2012 [1 favorite]


You are vastly overstating Skinner's argument. His argument is that this kind of direct influence is thought to be comprehensive and determinative while, in reality, there are a huge number of other cultural influences. Your argument is that these direct influences are minor and unimportant and uninteresting (which is the only way in which a "map of influence" between philosophers would be flawed and useless, and probably not even then). That's not that Skinner wrote nor, in fact, is that true.

"...it's that most historians of philosophy no longer accept 'influence' as a valid criterion for writing the history of philosophy" (my emphasis)

No, that's not true. That's such an overstatement that it's simply wrong.
posted by Ivan Fyodorovich at 5:25 AM on June 29, 2012


Who the hell is Murray Rothbard?
posted by outlandishmarxist at 7:07 AM on June 29, 2012


Rothbard was an Austrian School economist, and is much beloved by libertarians and anarcho-capitalists.

I was going to say that the chart seems to overvalue thinkers with a culty appeal -- Rothbard, Chomsky, Strauss, Kropotkin, and even Nietzsche -- but then I noticed that Ayn Rand is a tiny little dot just north of de Maistre, so there goes that theory.
posted by steambadger at 7:45 AM on June 29, 2012


Should old Aquinas be forgot.....
posted by temporicide at 7:47 AM on June 29, 2012 [1 favorite]


To be honest, pace the framing of the post, I think what's interesting here isn't the relative size of the little circles but the fact that the program he used identified different groups of philosophers, defined by being more connected to each other than to the philosophers outside the group, and those actually correspond pretty well to the groupings that philosophers themselves use. Now, some of this is going to be because out-group influences will be less intuitive and so less likely to be recorded on Wikipedia, but it's very striking and pretty nifty.
posted by Acheman at 8:30 AM on June 29, 2012


Christopher Hitchens?
 
posted by Herodios at 8:40 AM on June 29, 2012


Groucho Marx?
posted by temporicide at 8:58 AM on June 29, 2012


I do not see Groucho Marx on the graph.

I do see Christopher Hitchens on the graph.

I found that odd.
 
posted by Herodios at 9:21 AM on June 29, 2012


temporicide: Should old Aquinas be forgot.....
Aquinas is on the graph, way off to the left near Confucius. Below, but not connected to, Scholasticism.

At first, I thought Socrates had been omitted altogether, but he's there above Plato, smaller than Allan Bloom. I was glad to see Baudelaire on the chart, who appears to have been as influential as Voltaire!
posted by dilettanti at 9:24 AM on June 29, 2012


I thought Socrates had been omitted altogether, but he's there above Plato, smaller than Allan Bloom.

Publish or perish, ol' Soc.

Publish or perish.
 
posted by Herodios at 9:59 AM on June 29, 2012


"Descartes is smaller than you’d think"

Of course. That's why you don't put Descartes before the horse.
posted by jonp72 at 10:03 AM on June 29, 2012 [3 favorites]


twoleftfeet: ""Descartes is smaller than you’d think"

This is interesting because all my previous estimates of his size were based on this portrait of Rene Descartes, which is only 30 inches high. I suspected that Descartes was actually quite a bit bigger, but now I wonder.
"

... something something projection upon a plane...
posted by Reverend John at 7:44 PM on June 29, 2012




Mother Jones blog: Who Is Edmund Husserl?
posted by the man of twists and turns at 10:51 AM on July 9, 2012


« Older 40 years of arcade gaming   |   Revenge of the Fan Newer »


This thread has been archived and is closed to new comments