Michael Wolraich's picture

    And Now for Something Completely Different

    A few years ago, a friend of mine suggested that it would be nice to have a reference tool to find stuff related to other stuff. Type a movie title, and get the actors. Type a state, and get the capitol. Type in golf, and get a list of famous golfers.

    So I came up with an idea for a new kind of reference tool for free association. It just connects words with words--no order or structure of any kind. It employs crowd-sourcing to create the associations on the assumption that whatever you associate, plenty of other people do to. My hope was it would enable people to quickly find that word or bit of information they're looking for without having to do any searching or categorizing.

    I wrote up the tool pretty quick and found it fun to play with, but there's one problem: no data. It's not very useful without a crowd adding the data, and I can't attract a crowd to add the data without giving them something useful to start with.

    So I just let it molder in a corner of the web with a few hundred words in it and no users. But I still believe in the concept. It may just need some data to come alive. So I hope that the good folks at dag will take a moment to check it out at www.free-associate.com. Click around, enter a few free associations, and tell me what you think. You don't need stick to politics or any other category; just start free associating, and see where it takes you.

    Update: If you use StumbleUpon, please stumble: http://www.stumbleupon.com/url/free-associate.com/

    Topics: 

    Comments

    cool beans


    Question

    Is Free Association the right term for what  you want? For example Type golf and get a list of  golfers.

    Suppose I type Glennbeck and what comes into my mind is Glen Miller?  Is that what you're looking for? I don't think so. Or maybe it is.


    Free association is what I had in mind, even random connections like Glenn Beck-Glen Miller. My assumption is that most people will think more or less the same way so that the associations that get emphasized will be the most relevant ones. But I'll see how it plays out. In some ways, the random connections are more valuable b/c those are the ones that you can't find on other references sites.


    Actually, I could even see that kind of association as useful -- a whole list of notable Glens or Glendas could settle many a barside issue.


    It's in there now, Flavius.  Check it out.


    I love this!  Google already does something similar, of course,  but you need to know the right question, or better yet, the right way to ask the question.  I can't tell you how many times I've sat around with friends -- dinner party, beers, watching the game, even road trips with the tech available now -- and gone to the google to settle a friendly dispute or answer a nagging question.  I'd love to make FA the first, quick stop.  Thanks for posting this.  I look forward to keeping tabs and contributing.  This could take off like wildfire. 

    ps  The associations with gangs were my random contributions.  Don't know how that word popped into my head.  I'd entered 4 or 5 words before I figured out the program was storing them.  Some instructions and/or guidelines might be a meaningful additions, FWIW


    Cool. I'm glad that you like it. If I get serious about promoting it, I should probably make it a facebook app.

    The gang stuff is right kind of random. There was a bug in the program that was causing the additions not to show up. I fixed it now.

    Some instructions are probably a good idea.


    Okay I have to think about this also.

    You are obviously looking for something more worthwhile than shitehead: beck, rush....

    Just to give you some idea how much I love Wiki & Google, I can type Inhofe & oil and get all sorts of sites telling me about his recent energy votes, his contributions from the oil oligarchy...

    Google now has some message: do you feel lucky today?

    I can type Boehner & Orange and get something, I am sure.

    Anyway I am confused, but what's new?


    I figure that there will be plenty of rush - shithead type stuff, so if that's what comes to your head do it. I just added in boehner and orange.

    PS Google had the "I feel lucky" button from the very beginning


    I think I may have broken it

    =(

    So it's like a Pandora for words?


    It's more like wikipedia for streams of consciousness.

    And I don't think it's broken.


    Wow, there are twice as many associations in there already. Nice work folks. Keep it up.


    Here are some doodles--by the way why not doodle instead of google?

    Google & wind. http://www.huffingtonpost.com/2011/04/18/google-wind-farm-investment-oregon_n_850707.html

    tandem & condom.

    Condemn & condom

    Condemn & condor

    Canon & cannon

    war & price

    word & ward

    tortoise & turquoise

    turkey & turquoise




    Great, but don't write them here. Load 'em into http://free-associate.com.


    Obviously you need to work it up to the point where Google buys it from you and connects it to their system. Then we will be able to "word-associate" to our heart's content and you will be worrying about all the taxes you'll have to pay on your new millions (billions?)


    I just loaded "dork" and "computer programmers." 

    That site makes me happy. 


    I resemble that remark!


    I looked up Genghis and didn't find Ghengis, so I added it.


    I think you're on to something big. The word "exponential" comes to mind.

    Actually I added the word "piano".

    The word came to mind because the other day I was trying to find who locally sells a certain type of piano in my area and I was so frustrated with Google that I went searching for an old copy of the yellow pages.

    I think your idea is significant just because it is anti-Google and because it is essentially creative and artistic and not just another way to channel our lives into hallways of advertising.


    (...) just another way to channel our lives into hallways of advertising.

    If it doesn't do that Genghis won't be able to sell it for billions and run for office on the money.


    And you're suggesting that's a bad thing? Look at all the mischief Genghis has created with the mere millions he's earned off Blowing Smoke.


    A wonderful idea!  It needs to be loaded up, of course, but the idea is tremendous.  You need to hire interns who will sit around all day this Summer adding content.


    Given my vast budget, I think that this would have to go viral to get more content.

    I added a facebook like button, but I'm not sure how else to seed it.

     


    A suggestion:

    add it to StumbleUpon, and then ask people from around here to give it a thumbs up.

    It should be the kind of thing that appeals to people who like free associating sites across the web which is the philosophy behind StumbleUpon anyway, so that might crank up the hits and get more content.



    My first thought, G-man, was why add a "facebook like" button? Why not add an actual facebook button? I'm slow sometimes.

    I don't think your main problem will be lone assholes like quinn (who knew what he had down his pants?) but corporate and political spammers with an incentive to screw with the site. Wikipedia keeps its self-appointed contributors honest by inviting self-appointed referees to jump in. No one (not even you) has the time or inclination to police a site like this.

    So far, though, it's provided a chuckle or two. I like the ever-changing random selection of terms that pops up each time you access the site. Perfect way to seed the site! And the number of associations already exceeds one per cent of the entries in a typical dictionary. Damned good, considering your so-far-limited contributor base. Put a permanent link in a side column.


    I entered 'cheating' early this morning, but my associations were too linear, not so much free-associations, I think..  It ended up sending me to 'the Google',  like the Woodshed of Free-association failure.   ;o)


    Oops....how to fix typos and other mistakes?

    3rd Rock not 3rd ...whatever.

     


    It is fun adding content. I have a question though. How do you correct a mistake?

    I typed in Howard Hughes and since there was nothing, I started by adding the word Aviator ... except, I accidentally spelled it Aviatior. Is there a way to correct spelling errors?  (Sorry. I'll try to check my spelling in the future.) 


    I don't have a good solution for that. There are no logins or user ids, so no way to verify the contributor of the phrase in order to authorize an edit.

    My thought is that when there is enough data, the one-offs will be submerged. I count the number of hits on down each association, and presumably, most people will follow the links towards the correct spelling, so that will be weighted higher. Right now, I show all the associations but make the popular ones bigger. Eventually, I may hide or purge the unpopular associations which should make the misspellings go away.

    (I tried to put in a google suggestion feature to check spelling as you type, but it was too complicated.)


    One could also argue that the misspellings are valid streams-of-consciousness, and that the frequency of those misspellings will show up in that data.


    If nothing else you'll find out how many people can spell.


    I seem to remember that Google had a "stream of consciousness" thingy in their "lab" section... Rummaging around I found this: http://labs.google.com/sets


    Great idea, G. Seriously impressed. 

    You could see it growing into a map of the human un/subconscious, an absolutely wonderful thing for poets etc. AS LONG AS they started at the bottom, the least-often-mentioned connections! Maybe these things could be shown graphically, as trees. Some would go to the roots, others to the tips of the branches. Anyway, truly marvy.

    That said, you need ways to block out assholes like me. Or more precisely, assholish activities someone such as I would engage in. Like... answering "quinn" repeatedly, to all sorts of major words or ideas. Some feature whereby a person couldn't enter the same word more than once in a row or something. 

    I suspect the teenage idiot brigade would probably overrun it, and make it a bit more "urban dictionary" than it should be... but who knows. You'd need to get it into lots of adult hands just to get some entries in, then maybe make the site itself less sexy than it could be. Then... just delete like mad from time to time. 

    But kudos on a good, big, idea.


    Thanks quinn. I'm not concerned about assholes like you, at least not in the long run. As I indicated in my comment about about errors, with enough data, the efforts of lone assholes should be submerged by the rest.

    The ones that might hold their own are public figures, but that's a OK. It would be interesting to see which link in "asshole" becomes most followed.

    (In the short run, I did remove one association out of sensitivity to the small community that is currently using the site.)

    I any case, thanks for the enthusiasm. The functionality can be improved, but I think the site seems to be relatively sticky already. A relatively small number of users have actually added quite a bit of content.

    The only trouble is it hasn't proved viral so far. StumbleUpon didn't run with it, and there are some 17 facebook Likes that aren't growing. Maybe it just needs someone more connected to discover it, but there's also not much incentive to share it.


    I suspect it needs some sort of enormously over-hyped hook to sell it to the bloggers and promoters out there. Like...

    "We're putting the world on the couch, and seeing what they think of." 

    "We're mapping the global unconscious."

    Then you can promote it to people by asking if they want their unconscious dreams and desires to be left off the global map.

    Or, every day, throw out a word or a few that you want people to free associate TO.  

    P.S. Though to obtain 43 regular users, you could also pitch it at poets and bloggers searching for topics and terms. Naw. I'd go with mapping the global unconscious! 


    I like global unconsciousness. Better than my stupid blurb about crossword puzzles. And more likely to get people to share it.


    Not that my advice is very useful, but fwiw, I thought you may want to tweak how people interact with it.

    Contra Q, I don't think it necessarily needs a catchy title or anything. Its the way it's set up, imo.

    Free association maybe feels most interesting when you're getting a word thrown at you and then you see what word comes out of your subconscious. There is something odd about the way this is set up where you

    1. think of a word/expression

    2. see what other word/expression percolates to the surface of your mind

    There is something less spontaneous or 'free' about having to do both steps. It is hard to 'spontaneously' react to yourself. So my suggestion would be to try to set it up so that

    1. the program gives you a word/expression

    2. you type in the first word/expression that comes to mind.

    Of course, then it becomes artificial in another way - you have to find an algorithm for what words/expressions to propose. You could however just let it grow organically - pick words already in the association cloud as it stands at random.

    I think people would love to contribute content to such a project , and it is intrinsically fun to see what your own mind comes up with sometimes.


    Sure pal, take my idea - of a word for the day - and make it better.

    Just go check out what word pops up now when you punch in "troll" eh? 

    That'll learn ya.


    Gonna turn the world subconscious against me, eh?

    WAR .. of the worrds...?


    Great ideas. I think it's fine to pick words from the cloud at random. There are enough seeds in there now, and there will be more as people add associations.

    Let me think about how to do it. The search needs to be on the page, but maybe I can move it somewhere else and emphasize the challenge word.

    A related thought--to really achieve stream of consciousness, after you add a word, you should perhaps see the page for the new word. Otherwise, it's more like a delta of consciousness.

    These thoughts still address stickiness, however, rather than viralness. I can get people to add more content, but how do I get them to share with their friends? One thought is to allow people to share their associations on FB and Twitter. This would fit well with moving from a delta to a stream because you could share your entire stream rather just a word or a page.


    Glad you like the ideas. Here are a couple of more developing on the first ones:

    To make the page more fun, you could have different activities

    1. Take a ride on the stream (have the program give you a series of random words to associate)

    2. Add words to the database (from which challenge words are picked)

    3. Search function: Ask for expressions associated to a given word

    So (1) is just for fun, (2) is when you have a word and want to see what others associate it with (you come back a couple of weeks later to see what word-associates it has), and (3) is to see the association map around a given word. Actually, the current setup, with the function where you insert both the challenge word and then its associate, might not have any role to play at that point.

    As regards viralness, if you want people to share, it might be worthwhile thinking about how the pages for words are represented, and what is represented therein.

    What would people want to share on FB or elsewhere? Imho, maybe something like the search results page for a given word, where the results may be funny, intriguing, or just somehow curious, right? So it helps if that page looks pretty good. And then there is the information therein. It might be fun to see not only the first-degree associations, but also the associations of associations, and maybe even third-degree associations. Kind of a picture of Hebbian pathways of connection like those in neural networks.

    So representationally, you could have a pseudo-three-dimensional space of words, as seen from the perspective of the 'home' word. (i) The largest (nearest) words will be the first degree associate words, even varying in size according to popularity or frequency, then (ii) smaller (more distant) words that are the second degree associates each linked to one or more of the first degree associates. To make the difference between first degree and second degree associates clear, you could highlight the former in color, and leave the latter in grey, or something.

    The screen then would look a bit like a word cloud, the 'home' word in large bold in the middle, and then large first-degree associates around it with smaller second-third-degree associates scattered in between in grey. It would be like looking at the night sky with nearer stars and planets and more distant stars and galaxies. The idea would be that interactively, the user could 'surf' through this space of word associations, clicking on associate-words and thereby 'hopping' to that word's 'home' page from which the perspective on the broader word-space changes. That could then be a fourth optional activity on the site - just surfing around the existing word-universe, hopping from word to word through already established associations, without necessarily adding any new connections nor seeking out any particular associations either.

    Hm, my enthusiasm is getting the better of me. So do ignore if this is getting impracticable, or not what you were looking for.

    But it would be literally like flying around the global subconscious, and the graphic representation should reflect that fact. It could be exhilarating stuff, which I'm sure lots of people would ... 'like'.


    Ok, listen up, as I'm kinda tired of this crap. 

    MY idea was... a tree.

    Whereas YOUR version is some goddamn acid trip through all 11 (or 13) dimensions of Brian Greene's mind, flying Cutter John's wheelchair through a graphic representation of the global subconscious. 

    Clearly, yours would be more popular, draw in a lot more participants, generate more cash for Genghis and enable us earthlings to rapidly form a planetary consciousness, thus achieving world peace.

    Still, man, think about it. A tree! 


    Great ideas which would be excellent in the future, but I'm going to keep it as simple as possible for now. Please check out the new and improved version that emphasizes riding the flow. I would live to hear your thoughts and those of the canaidan tree-guy and everyone else who has intelligent thoughts.

    My plan, if this version is good, is to add a Share My Stream option for Facebook and Twitter. The share link would display the whole stream and encourage ohters to generate their own streams.

     


    Like the updated version.  I think it's much better being given a word to respond to at the outset helps.

    One addition maybe to think about - since it the global unconscious - and once there is a larger database of associations is to pop up randomly what others associate with the word one first types into the initial word.  For example, if the initial word is "fork" and I put in "pie" then what pops up next for me to free associate is "chart" which someone associated with pie at one point.


    That would bias the results, no?


    I guess it would bias it in a different way.  I would see where one could have the option of creating their stream of consciousness string or one which is feeding off other people's associations.  I kind of like the latter in regards to interacting with on the web because I get to see what other people are associating to something I am thinking about and then feeding off of that.  So the string that is created is more a "collectively" developed string.  Whereas in the former option, aside from the beginning word, is just my string which is then added to pot afterwards, which in the end does create some interesting collective data.


    WTF?! Turnips?!!!!


    Hey. It's FREE association, man. Free you mind, Bwak. The turnips will follow.

    Latest Comments