All Roads Lead to “Philosophy”

There was an idea floating around that continuously following the first link of any Wikipedia article will eventually lead to “Philosophy.” ¹ This sounded like a reasonable assertion, one that makes a certain amount of sense in retrospect: any description of something will typically use more general terms. Following that idea will eventually lead… somewhere.

It also sounded like an idea that would be easily examinable with basic client-side scripting tools, using the Wikipedia API and a good graphing package. I put something together here based on JQuery and the JavaScript InfoViz Toolkit. It makes use of the HTML5 <canvas> element, so support for Internet Explorer is provided by the Google excanvas package.

I still have a lot of tweaking to do but the results so far are pretty nice.

Multiple titles can be added using a comma-separated list. JSONP requests are made to Wikipedia asynchronously, so more terms can be added while it is accumulating results.

There are some circumstances where a loop is detected up the chain. This is relatively rare. If it finds that it moves to the next link in the chain. One good example is “Telecommunication”.

¹ See the tooltip by hovering over the cartoon at xkcd which is said to be the source of this observation. Though this posting on reddit.com appears to predate that by about a month.

» Posted: Tuesday, May 31, 2011 | Comments (86) | Permanent Link

Comments

Glad I bookmarked this site.
This tool could actually be useful if we gets to change the ending term as well- great for waffling in essay, linking together pieces of information for counter-argument/debate, or just the curiosity in finding how “your mum” is related to… i don’t know, Hitler?

» Posted by Dave on June 1, 2011 02:59 AM

Good show!

» Posted by Ralph Bean on June 1, 2011 09:39 AM

Can this work to stop whenever a duplicate article is found rather than Philosophy specifically? As in, to determine whether there are any other tree roots in Wikipedia?

» Posted by Some Person on June 7, 2011 05:48 AM

This application is hard-coded to look for the “Philosophy” article specifically, but changing it would a trivial matter of updating the root article title and id in the code itself. This is all client-side code, so anyone can download it, change it and run it locally. While I could have added options to allow that to be done on the fly, I was specifically trying to make something very simple and aesthetically pleasing.

Though I haven’t explored it, I’m sure there are other “roots” and in fact, probably a ball of them among a group of very general terms.

» Posted by Jeffrey Winter on June 7, 2011 07:48 AM

Nice work !

» Posted by Tgoc on June 7, 2011 05:24 PM

1) What does your program mean by “first link”?

I’ve heard “the game” being described as the first link NOT in parens AND NOT in italics. I can assume it’s also meant to ignore brackets as well (obviously ignoring superscript, ie citations, etc.)

2) Regardless of such semantics, I cannot see links between certain articles at times. A great example that your program provides: Human > Extant Taxon > Biology > […]

How are Human and Extant Taxon related? There’s no link on the Human article to Extant Taxon…at least not a member of the first few that could be determined “first” in any context of the “rules.”

3) Very interesting work! Well done.

» Posted by Whammy! on June 7, 2011 08:19 PM

Seconding the option for being able to modify the root - when playing 6 Degrees of wikipedia, technique was to find the fastest way through to the Christianity article.

» Posted by Kieran on June 7, 2011 09:23 PM

1. It ignores just what you’ve outlined. It actually is parsing the text as returned by the Wikipedia API which in some cases is slightly different than what is seen on the page itself, but in ways that make sense. For example, the API may show a language code which is rendered on the page as a link to, say, “Greek”, (even outside any parentheses rendered on the page) but this code is marked by parens and can be skipped. It prevents detours along purely etymological lines.

The other deviation is if a link would produce a loop backwards along the single chain of links it is currently walking, it will skip to the second link, but this is very rare.

2. The first link on the “Human” page is to “Existing Taxon” it just happens to be bound to the word “living”, so that does follow the rule. (Notice the link when you hover over the word “living.”)

This isn’t perfect, and I’m sure there are plenty of corner cases I haven’t captured quite correctly, but I’ve tried to clean them up as I encounter them.

3. Thanks!

» Posted by Jeffrey Winter on June 7, 2011 09:26 PM

It works! Fantastic!

» Posted by Nakiska Shaikh on June 8, 2011 11:05 AM

@Jeffrey Winter Thanks for the clarification!

I can’t believe I missed the [[Extant taxon|living]] link - doh!

In the very rare cases, does your program note that it does so (ie the “game” doesn’t work)? That would be a great feature. Out of curiosity, are there any words you’ve seen do this?

» Posted by Whammy! on June 8, 2011 01:56 PM

Hahahahaha, I see Mathematics was until 00:29, 6 June 2011.

In all fairness, it’s a dually beneficial update.

» Posted by Whammy! on June 8, 2011 02:29 PM

This is a great execution of a neat idea. Thank for making this. I have spent too much time and had too much fun playing with it.

I did have one question. Would it be possible to add a “random” button? I thought it might be interesting to enter a bunch of random articles and see what happened. Some interesting groupings appeared. It would be great if this was possible to do without manually entering each random item.

» Posted by Patrick on June 8, 2011 02:36 PM

Yes… :)

One that comes to mind is “telecommunication” that loops back. I had thought of marking those nodes in blue or something, but it seems sort of incidental and it generally is in the spirit of the original assertion to just move along.

There is one bug that currently isn’t handled at all, and that is with “Exogeny” where the loop is formed as part of a redirect. Currently that is left hanging in the air, disconnected from the “root”. I haven’t had time to deal with that particular circumstance yet. I tend to fix issues as I encounter them. As I said, this isn’t perfect by any means, but by this point it handles the vast majority of cases.

» Posted by Jeffrey Winter on June 8, 2011 02:39 PM

@Patrick. That’s actually an interesting idea. Let me see how easy that is to execute. I’ve purposefully tried to keep the interface as minimalistic as possible, to the point of making sure there aren’t even any scroll bars to detract from the layout.

I also wonder if that would expose all sorts of issues that I haven’t encountered yet. In my testing I just randomly punch in the first topic that comes into my head, but they are perhaps relatively general items that don’t “stress” the proposition quiet as much as a random article would.

» Posted by Jeffrey Winter on June 8, 2011 02:47 PM

The reason this works is that typically the first sentence in an article for, say, XXX is to say something to the effect of “XXX is an item belonging to category YYY” or similarly the first link in the article is a link to some type of abstraction of the article’s subject. This isn’t strictly true, obviously, but it generally is. So if you proceed from any topic to the topic one degree of abstraction higher and onward, eventually you arrive at the most abstract of all: philosophy.

» Posted by Mike N. on June 8, 2011 04:28 PM

Boy as I re-read that it occurs to me that with spam filters as aggressive as they are I should really pick better placeholder names in blog comments :-)

» Posted by Mike N. on June 8, 2011 04:29 PM

Awesome idea and execution. I’m having fun finding connections (ever see that Discovery Channel show?) and then reading about things I never knew that contributed and filled holes for things I did know.

» Posted by Kevin on June 8, 2011 05:04 PM

This is a nice and shiny looking program. But you might want to make it clearer that if the program actually finds a loop, an article linking back to another already encountered article, that it then selects the next possible link.

Otherwise, this might make people believe that the “philosophy hypothesis” is true in all cases, not just in most of them. And there are stable loops in wikipedia that never get to philosophy at all.

» Posted by Christoph on June 8, 2011 07:48 PM

The first link for Philosophy is Reason, so really anything on the chain from Reason -> Philosophy could be considered something that all roads lead to. As you put more and more unrelated terms on the same chart you can see this “spine” appearing.

» Posted by Neil H. on June 8, 2011 09:37 PM

More importantly, every article circles back to Science in the same manner. Actually, there is a circular loop between Science and Philosophy that uses Mathematics as an intermediary. There’s not really an end point in the loop, but based on your work it shows that more things are linked directly to Science and that Philosophy is just part of the ‘foundational thinking’ articles within that loop.

» Posted by Nic on June 8, 2011 11:02 PM

Mary Magdalene is a bug, it is leading directly there

» Posted by bob_bobberson on June 9, 2011 02:09 AM

I have a slight problem with this kind of reasoning.

If all roads/articles/whatever lead to Philosophy, then by nature Philosophy also leads to these.

In fact, everything leads to everything else.
The only difference, and would be most interesting, is how many ‘hops’ it takes from any particular article to another.

» Posted by Vincent on June 9, 2011 05:47 AM

Hm, there seems to be a problem with nouns of people (Marc Aurèle, for example, returns missing).
FWIW, I’ve put online a French version of your tool, here: http://w3c.org.free.fr/wikipedia/
Thanks anyway, it’s a very interesting tool!

» Posted by w3c on June 9, 2011 05:57 AM

@w3c Nice! I like that there is a french version. I suppose I should have allowed different versions to begin with but I was just trying to be simple.

I just tried “Marc Aurèle” myself and it redirected to “Marcus Aurelius” and then was on its way, so I don’t know if you established that redirect at Wikipedia, or what, but it works now.

This application is obviously completely dependent on the data returned by the Wikipedia API. If they have an article with a given name (or it redirects to another one) then it will parse the data. If there is no data - or for some reason the API decides not to return any - then it just posts “Not Found” on the side scroll.

Sometimes it’s a typo and sometimes there’s really nothing there. Another thing: for names it seems to be sensitive to capitalization, so “Marcus Aurelius” is found but “Marcus aurelius” is not.

On the other hand, “Steve jobs” redirects to “Steve Jobs” but “Steve ballamer” does not redirect to “Steve Ballamer” so I suppose it depends on whether someone took the time to add the redirection to Wikipedia itself.

» Posted by Jeffrey Winter on June 9, 2011 08:00 AM

Super cool! Unfortunately I think I’ve caused it to collapse in on itself, twice, with the combination, “air conditiong, broccoli.” Also, would love it if the vis rescaled so that you can still see the full extent of the big ones can.

» Posted by jessykate on June 9, 2011 10:20 AM

You didn’t even mention somewhere that it directly comes from the xkcd comics of May 25: http://xkcd.com/903/ (in the text that appears when leaving the mouse pointer over the image: “Wikipedia trivia: if you take any article, click on the first link in the article text not in parentheses or italics, and then repeat, you will eventually end up at “Philosophy”.”).
In addition, the loops that do not lead to Philosophy are not such rare, many loops were manually found: http://www.explainxkcd.com/2011/05/25/extended-mind/#comments

» Posted by Ethaniel on June 9, 2011 10:50 AM

cool program! the chain starting at ‘philosophy’ ends up at ‘mathematics’ excluding a loop at ‘rationality,’ so mathematics can equally be considered a root.

» Posted by H.G. on June 9, 2011 11:03 AM

@Ethaniel I had heard about it second hand I guess. :)

» Posted by Jeffrey Winter on June 9, 2011 02:33 PM

Regarding the request for adding a random article, there is the http://en.wikipedia.org/wiki/Special:Random link which gives a random article. I don’t know if this is what could be used as the “first” term.

In (manually) putting several (more than I care to admit) of these random pages in to see what sort of patterns emerged, I found two issues:

1) If the first (as defined) link does not exist (is red as displayed by Wikipedia), the chain immediately ends (“Schola Cantorum de Venezuela” being the one that popped up or me); in this case, it seems within the spirit of the rules to go to the next link (that is, the first link to another existing wikipedia page).

2) Pages with commas in their names are not handled correctly. http://en.wikipedia.org/wiki/Rockyford,_Alberta was the random link that I got that had this problem.

Thank you for this entertaining tool!

» Posted by Brian on June 9, 2011 03:21 PM

1. Random button. I added the random button just now. I’ve been reluctant to make many changes because I want to keep the interface simple, but this seems reasonable. I’m sure it will expose all sorts of problems.

2. “Schola Cantorum de Venezuela” I’m not sure there is much I can do about this one. The problem is the first link in the article is to a non-existent article, but the Wikipedia API does not mark it as such in the data. It is only marked that way when rendered for display. This application uses the Wikipedia API to load the raw data not the page itself. I was actually surprised to find out that non-existent articles aren’t labelled as such in the raw data.

3. Commas in titles. Thanks for pointing this one out. I’ll have to think about how to delimit things better. I probably should have recognized that commas are legal in the title and worked around that. I’ll see what I can come up with.

» Posted by Jeffrey Winter on June 9, 2011 04:06 PM

This is fantastic, by the way Yahoo.com, Google and Bing. Google and Bing are on the same graph. Yahoo.com seems to have gone in a separate tangent! haha

» Posted by Gaurav Agarwal on June 10, 2011 10:07 AM

Thank you very much for adding the “random” button. It is a great addition. I have tried it out quite a bit and I haven’t noticed many problem at all. All of the problems I have noticed have been articles with dead first links.

One thing that I have noticed over the last couple of days is that several of the main trunks leading to philosophy have been edited. A couple prominent ones, like the trunks leading through “Natural Science” and “Mathematics” seem to have been shorted and simplified. It is amazing to seem changes to Wikipedia over a short period of time mapped out like this. The changes cause me to wonder if this idea has now become a self fulfilling prophecy. If the people that edit Wikipedia believe this to be true then it may become more ingrained that it was when it occurred naturally.

» Posted by Patrick on June 10, 2011 05:58 PM

This is great!

But what does it do if the first link is an internal link? It seems not to work: e.g. “Noon”.

» Posted by Anton on June 11, 2011 03:00 PM

@Anton - that would be a bug. I’ll have to work around this particular case and will update the code later on today. Thanks for finding that one. (It’s now fixed.)

» Posted by Jeffrey Winter on June 11, 2011 04:12 PM

Stumble onto this and suddenly it’s been 45 minutes. Fantastic find. Longest I’ve managed to find so far is Cthulu with “backblast” being a surprise at only one hop through the G.I. Joe character.

» Posted by Tycho2501 on June 12, 2011 05:34 AM

thanks for putting this together, I really love it. I was thinking it might be fun to copy a large number of these links and make a picture of them a bit like the great bear map of the Underground: http://bit.ly/iWpElF but I can’t seem to copy the image and/or spread out the lines to make the text easier to read. Is there a simple way of doing this? Thanks again.

» Posted by Slicktony3000 on June 12, 2011 06:51 AM

There is a message for you on the Wikipedia page “Wikipedia talk:Get to Philosophy”

http://en.wikipedia.org/wiki/Wikipedia_talk:Get_to_Philosophy#A_plea_to_the_authors_of_the_tools_at_ryanelmquist.com_and_xefer.com

» Posted by Wikipedian on June 12, 2011 02:02 PM

@Slicktony3000 I have code to rotate and spread the graph that will be going in shortly. I’m trying to maintain the simplicity of the page but allow some level of customization.

» Posted by Jeffrey Winter on June 12, 2011 04:44 PM

http://en.wikipedia.org/wiki/Tom_Stenvoll

» Posted by Anonymous on June 13, 2011 03:08 AM

@Vincent,

The claim that “all roads lead to philosophy” doesn’t entail that philosophy leads to everything else. The relationship here is directional, taking you from page A to page B, but not from page B to page A.

As I’ve heard other people say regarding this (very cool!) widget, “everything is connected to everything” in Wikipedia. But the point here is that everything is connected to philosophy *in a very special way.*

» Posted by David on June 13, 2011 02:30 PM

Hey, I like the interface, but it seems that there are many articles where there is in fact no link to the “Philosophy” page but still the graph would make it appear like that.

Check with:

“Side of the road (disambiguation)” : I could not find a link to Ph. from http://en.wikipedia.org/wiki/Driving_on_the_left_or_right as indicated.

Same is true for http://en.wikipedia.org/wiki/DNA_viruses

Yet, I like it :-)

-eWald

» Posted by Ewald on June 14, 2011 05:43 AM

@eWald I’m not sure what you saw, but those articles seem to get to the root following the established rules - as of now anyway. Perhaps there was an issue when you were looking at them. These relationships are obviously not permanent so it’s possible they weren’t there when you were looking.

» Posted by Jeffrey Winter on June 14, 2011 07:36 AM

Hitler Dosen’t seem to work properly as austia-hungary is linked twice

» Posted by Stoks on July 4, 2011 10:17 PM

In the spirit of the rules (http://en.wikipedia.org/wiki/Wikipedia:Get_to_Philosophy#Rules), it might make sense to make a version where one of the other “end” conditions shows differently - ie loops show as loops separately from the main tree, pages with no links end separately and turn red, or something like that.

» Posted by Karlan on July 11, 2011 04:23 PM

Hello,

I tried the word Forward :
Forward -> Forward (surname) -> (surname) Not found
So forward does not lead to Philosophy.

Thanks

» Posted by Jarzac on July 21, 2011 01:55 PM

6 Degree of sepration

» Posted by Meysam on July 25, 2011 09:21 AM

This is fantastic but is there a way to add more languages? I’m a Farsi speaker and a developer. Is there a way to commit code to this cool project so we’ll be able to check same thing in Farsi (Persian) ?

» Posted by jadi on July 31, 2011 09:56 AM

Another point… what happens with “bakcpack”?

» Posted by jadi on July 31, 2011 10:11 AM

It’s simple to add new languages; the problem is that the vast majority of them don’t have enough coverage to end up linking to “Philosophy”. Even Japanese and Russian don’t make it and they have close to a million articles each. Farsi has only around 150,000. See: http://meta.wikimedia.org/wiki/List_of_Wikipedias. If you want to try it yourself, you can add a new language to the “root” object in graph.js.

I’ll have to look into “backpack”. It’s probably a bug in my parsing code which doesn’t surprise me :)

» Posted by Jeffrey Winter on August 1, 2011 07:25 AM

Thank you for making this piece of brilliance. Your work inspired me and I’m looking for the most comprehensive documentation on the Wikipedia API (please forgive me if I missed it). Your help (and mad skills) are much appreciated.

» Posted by Preston on August 3, 2011 05:04 PM

Hi Preston, I actually only know the very basics about using the Wikipedia API. This site really only needs to do two things: query a specific page and query a random page.

Everything I know I gleaned from looking at the API documentation:

http://www.mediawiki.org/wiki/API:Query

For me the real challenge was handling the data once it came back but that’s not the scope of the API of course. JQuery made using the AJAX interface easy.

The only thing the API didn’t specify explicitly (or at least I didn’t see it) is the “callback” query parameter it uses for JSONP. That’s the standard one, so I just tried it and it worked. Other than that everything else I needed was pretty well specified in the doc; it just takes a but of testing. I’d recommend using curl.exe for those purposes.

Good luck.

» Posted by Jeffrey Winter on August 3, 2011 07:11 PM

nice post! I believe that everything starts from philosophy and comes back to it

» Posted by eve on August 6, 2011 07:26 PM

This is fantastic. Congratulations on creating something fun, interesting and easy to use. Could you provide an option to change philosophy to something else….so we can search the connections from x to y rather than just x to philosophy.

» Posted by Booker on August 12, 2011 05:55 AM

Is it bug? When I enter “community”, the next item it shows “interacting”. But “interacting” is not the first link, nor the second, nor the third. It is the fourth link in the page about community.

WHY?

» Posted by Max on August 13, 2011 01:05 AM

@Max: Errrr… stupid bug. I’ll fix it soon

@Booker: This follows a specific rule of only following the first link. Only a few specific titles would ever be reached using that rule. What can be done is to click on the dot itself over the title which will realign the graph on that title. It will still walk towards Philosophy but the chosen title will be at the center. You have to click the dot itself, not the title which forms a link to Wikipedia.

» Posted by Jeffrey Winter on August 14, 2011 10:34 AM

I found out that if you start at “Computer” you don’t get to Philosophy.

» Posted by Anonymous on August 15, 2011 03:58 AM

This tool went from “http://en.wikipedia.org/wiki/Television_in_the_United_States” to “http://en.wikipedia.org/wiki/Media_of_the_United_States” however, that isn’t the first link. The first link happens to be the first word, but it seems to have been skipped.

» Posted by Justin on August 29, 2011 02:37 PM

Excuse me, but how can someone download the code?
Great application by the way!

» Posted by william on September 12, 2011 01:36 PM

Hi there! I think I’ve found a bug: Try “Switzerland” and it will tell you it found Philosophy as the next article… which is wrong. Philosophy doesn’t even appear in the article (only the word “philosopher” somewhere in the middle of the article, but it’s not even linked).

» Posted by Markus on October 17, 2011 08:17 AM

@Markus: the ‘Switzerland’ article had an unbalanced parenthesis that tripped up my simplistic article parser. I actually corrected the article on Wikipedia itself and it now works as expected. Thanks.

» Posted by Jeffrey Winter on October 19, 2011 01:53 PM

Mathematics is not a bottleneck, nor required to get to phylosophy.

» Posted by Anonymous on November 1, 2011 07:36 AM

Check
December 2010 in sports
http://en.wikipedia.org/wiki/December_2010_in_sports

doesn’t lead to philosophy

» Posted by Anonymous on November 10, 2011 09:59 AM

That actually works for me. Things can definitely go wrong with this application. You’re connection could time out; that data at Wikipedia my be temporarily unavailable; or there may have been an edit made that causes a problem.

This is looking at live data so anything can go wrong. Of course, there are also parsing errors in the code that have yet to be revealed, but I try to fix them as they are encountered.

» Posted by Winter on November 10, 2011 10:45 AM

“Bill Clinton” appears to produce a bug - it is presented as a direct link to Philosophy, but should either link to “Disambiguation” or, if disambiguation links don’t count, to “American”.

» Posted by Jeff on November 10, 2011 11:00 AM

Any path that leads to “architecture,” leads to “buildings,” which leads back to “architecture.”

» Posted by Brad on November 10, 2011 11:13 AM

Would be neat if you could tweet your particular set of branches. I enjoyed seeing my random Wikipedia search pages converge in a single graphic. Cheers!

» Posted by fejalish on November 10, 2011 07:11 PM

Oddly, the longest path I generated out of about 25 tries was “religion”. Searches that generated comparatively much shorter paths included “needle”, “kitten”, and “uveitis”.

» Posted by speedwell on November 11, 2011 03:48 PM

There is something beyond Philosophy.

» Posted by The Line on November 12, 2011 10:29 AM

Interestingly, the longest chain I was able to create was “Existentialism”.

» Posted by Anonymous on November 16, 2011 10:13 AM

Existentialism lead through Danish. Yes! My home country.

» Posted by Dane on November 18, 2011 12:55 PM

You should leave out redirects, since the redirect link is always going to be the first link in the chain. See the chains for, case sensitive, “Death cab for cutie” or “san francisco.”

I’ll second finding commas awesome btw.

» Posted by Anonymous on November 28, 2011 09:01 PM

to the person who said it was ironic that “Existentialism” was the longest chain they found, what’s even better is that “Plato” is even longer ;) I’m looking forward to someone sharing one longer than that, as I have yet to find one.

» Posted by Anonymous on November 29, 2011 04:54 PM

As of december 17/ 2011 property (philosphy) leads back to mathematics rather than modern philosophy

» Posted by Anonymous on December 17, 2011 07:38 PM

It connects everything with Knowledge, then goes to Philosophy. Not that extraordinary, but the tree builder script is really cool. I would like to see a version which omits Philosophy and all it does, is to connect certain keywords with each other.

» Posted by ashnur on December 21, 2011 05:32 AM

Sorry if I missed it explained elsewhere, but, how are you doing the graph search? Just basic Depth/Breadth First Search? It seems surprisingly quick. Excuse my lack of web programming clue. Very impressive.

» Posted by Rustin Bergren on December 22, 2011 12:41 PM

@Rustin Bergren: it uses the standard Wikipedia API to extract the given article. JQuery makes that very simple It then parses the raw contents and extracts the first link, then fetches that article. It just keeps going until it reaches the target (in this case “Philosophy”)

» Posted by winter on December 22, 2011 12:49 PM

LIGHTHOUSE!!!!

http://en.wikipedia.org/wiki/Lighthouse

DOESN’T WORK HAHAHAHA!!!

» Posted by Hudson on January 17, 2012 01:46 PM

Not to be annoying or anything (I’m sure you appreciate people pointing out flaws in the program!), but “Leipzig” leads directly to Philosophy when it shouldn’t. There is a link to Philosophy within the article but it is nowhere near the first.

Awesome program! It’s great to see the tree unfold.

» Posted by C-PM on February 3, 2012 04:25 AM

@C-PM thanks for pointing that out.

The reason why that particular article failed was because it was missing a parenthesis. This program uses a very simple parser to extract the first link to follow. Wikipedia obviously uses a more forgiving renderer that must auto-correct any issues such as these.

I actually went into the article and added the missing parenthesis and it works as expected now.

» Posted by winter on February 3, 2012 07:35 AM

Interesting hypothesis. I tested it starting at “Jerry Lewis” - it rapidly ended up in a loop beween the NATO and International phonetic elements.

I changed the rules slightly - but I think it keeps the spirit. I followed the first non-italicised link in the article proper (ignoring disambiguation sections). Then whenever I hit a link I had already touched I took the link immediately after it. (After all, in a maze if I came back to an junction where I had turned left I would naturally go right the next time.)

Jerry Lewis, did indeed lead to philosophy. But only after about 52 links.

I decided to substitute the objective with some equally fundamental concepts.

I would have to say that the hypothesis that all wiki articles eventually lead to “physics”, “culture” and “language” holds just as well as for “philosophy”.

My conclusion is unexceptionally - language and concepts operate in a very dense and complex network of relationships. Some are ‘hot nodes’ - they are very heavily connected. These junctions are so crowded with connections that even if you are randomly traversing the network there is a very high probability that you will cross one quite early in your travels. But there are many - of which “philosophy” is only one. There is no master node - many times I crossed “physics” and “language” on my way to “philosophy”.

If I had selected them instead the alternative hypothesis that all paths lead eventually to “physics” would appear valid - and the rich connectivity of philosophy would have been ignored by the initial assumption.

» Posted by RJP on February 29, 2012 06:20 PM

If you start with Latin it doesn’t work.

» Posted by Sid on March 3, 2012 04:55 PM

Actually, it seems you have to click only on those links outside of parenthesis and in the text of the article itself. Those seem to be the rules. If you take that path with Latin, it leads to Philosophy.

» Posted by Sid on March 3, 2012 05:01 PM

This is beautiful! Thanks for your effort.

Henrik in Sweden.

» Posted by Henrik on June 14, 2012 04:47 AM

This is a FUN thing to play with when you’re in work-avoidance mode. 8-) Thanks!

Would it be possible to add the chain’s length after the first element on each chain? This would make it easier to compare lengths to see which is longer. For instance, if one entered “Village”, the chain from Village to Philosophy would have “Village (11)” at the start of the chain. The “(11)” would not be part of the link to the article, of course.

Another question: is there any way to remove a chain from the graph, if the graph starts getting too cluttered? If not, could one be added?

» Posted by Chris W. on June 24, 2012 08:50 PM

Entirely consistent with an infinite series of ‘Why” questions: all lead to Schopenhauer’s Will, the thing-in-itself.

» Posted by Infonomics on September 15, 2012 02:29 AM

Swingline does not end at Philosophy. It ends at Law

» Posted by Anonymous on October 12, 2012 01:12 PM