ROB SEWARD   ::   WORKS   BIO   CONTACT   BLOG

Word Association Apps

Highlight App

Six months ago I became obsessed with a project at the University of South Florida. In 1973 the psychology department began performing word association tests. A word association test is where you say the first thing that comes to mind when given a word. For example, if I gave you “Cat” you might say “Dog.” Over 25 years USF performed around 750,000 of these tests on almost 6,000 individuals. The results are online here. In their current form the results are hard-to-read text files, so I wrote programs that could slice and dice the data and produce readable output. Eventually, I hope to find an algorithmic definition of repression. In the meantime, here are some of the apps I made.

SEMANTIC WEB VISUALIZER
Enter words separated by commas, press go, and get an image showing the connections between your input and other connected words.


COMPARE
Compares each word in the database with respect to two words you enter. Enter “Picasso” and “Mona Lisa” and you will receive a list of words that are closer to Picasso (“abstract”), closer to Mona Lisa (“portriait”), and equidistant (“painting”).


PATHS
This app shows what words are most closely associated with the word you enter. Rolling over the words will show you the associative leaps required to get to your word.


LOCAL CENTER
Type in a series of words separated by commas, and see the results most closely associated with that series. For example, if you enter “fangs, blood, coffin” you are shown “Vampire, Dracula,” and so on.


HIGHLIGHT
This app lists words closest to your “list-order word.” It will then change the size of the results according to a “magnifying word.” For example, one looking for evidence of an oedipal complex might enter “mother” and “sex.” Words will be listed in order of their association with “mother.” Words more closely associated with “sex” will appear larger than those less associated.


HIGHLIGHT NEGATIVE
This is very similar to HIGHLIGHT, except font size is changed according to distance from a group of “unpleasant” words [links to pdf]. As a result words are magnified according to their “unpleasantness.”


CHARLIE KAUFMAN INSPIRATION MACHINE
I noticed that words not directly related to an input word, yet still somewhat related, started to resemble elements of a Charlie Kaufman screenplay. This app lists elements that are 1 degree away from an input word. Enter “murder” and you get “accident, bear, tomahawk, drugs, television”–all you need to start your screenplay.


[Update]
For those of you who want to mess around with this, here is the study’s data as a a sql database.

6 Comments

  1. Posted June 15, 2009 at 1:21 pm | Permalink

    Hmm… interesting. Do you know what license the data can be used under (can it be placed under MIT or LGPL or something?).

    It would be good to integrate this with a natural language library. I know there are some good apps of this sort for nodebox/shoebot around.

    MontyLingua seems to be along similar lines -
    http://www.fslog.com/2008/09/20/montylingua3-gpled-fork-of-montylingua/
    …but I reckon it would be possible to get different stuff from your data.

    If the api is suitable, it might be nice to have a similar one for your data (or integrate your data with it somehow).

    Nodebox linguistics stuff
    http://nodebox.net/code/index.php/Linguistics

  2. Posted June 15, 2009 at 5:55 pm | Permalink

    Thanks for these links. The Nodebox stuff especially. I had an idea to take random text and swap out words that were in the association database with words that were 1° away. This didn’t work because there were too many parts of speech and congregation issues.

    As for the license, I don’t know. I wrote the PI of the project, and he said that their funding required them to make the data publicly available. I think the result is what you see on their website.

  3. Posted September 11, 2009 at 12:11 am | Permalink

    Great site…keep up the good work.

  4. Q
    Posted June 25, 2010 at 8:28 pm | Permalink

    Hi. Nice website and excellent list. I would like to know if you know whether there exists a list/database of word associations (excluding the SQL database mentioned above) for the English langauge online? I wanted to use the one you provided but opening it in SQL Server 2008 doesnt reveal anything I understand. I am probably doing something wrong. If you could give me a clue as to how to get access to the database you provided thatd be great. Thanks

    Regards

    Q

  5. Posted June 28, 2010 at 7:58 am | Permalink

    To interpret the database, you need some programming skills and this link, which explains the columns in the DB:

    http://web.usf.edu/FreeAssociation/Intro.html

    As for online lists, there’s always Visual Thesaurus.

  6. g.
    Posted January 28, 2012 at 4:09 pm | Permalink

    the HIGHLIGHT app is so beautiful. thanks for that

One Trackback

  1. [...] on the Internet that you can use. One of these is Rob Seward’s word app. You can find it here: http://robseward.com/blog/2009/02/23/word-association-apps/. This tool will not only give you synonyms but also related words that can help you craft better [...]

Post a Comment

Your email is never published nor shared.

Code
News
Project Related
Stuff I Like

Add to Google Reader or Homepage