evolOT:  Simulating language evolution with Optimality Theory
Gerhard Jäger
University of Bielefeld

evolOT is an implementation of the iterated Bidirectional Gradual Learning Algorithm for simulating language evolution.

How to Download and Install the Software

evolOT runs under various Unices and Windows. There are precompiled binaries for Linux and Windows. You can compile the sources yourself if you the Qt libraries are available on your system (standard for Linux and the OS Darwin for Mac).


You need Qt (mandatory) and gnuplot (useful). Furthermore standard tools like make and awk are necessary (which are most likely already installed at your system).

In the following instructions, I assume that you use the Bourne Shell. If you use another shell, type


at the command prompt (or adapt the commands below to the syntax of your shell).


click here to download evolOT

and use your Web browser to save the file.  The file you are downloading is called evolot.zip.

Unpack the archive by cd-ing to the directory where you put evolot.zip and typing

unzip evolot.zip

A directory "evolot" will be created, containing the sub-directories "evolot/bin",  "evolot/sources" and "evolot/data" (containing the executables, the source code, and the files from the example that is discussed in the manual respectively). Type


to start the binary program. If the evolot main window comes up, you are done. Goto the section on gnuplot. If you get an error message, you will have to recompile the sources.

To recompile the sources... (applies also to Mac/Darwin)

... figure out whether qt2 or qt3 are installed on you system. Type

echo $QTDIR

The command should display the home directory of qt at your system. If it doesnt, qt is probably not installed. You can download the sources from here for free and compile it yourself, and most Linux distributions have it as a package. (At present evolOT is optimized for qt2, but it should also work with qt3.) If the command just gives you something like "/usr/lib/qt" (without a version number), try

ls $QTDIR/lib

You will get a list of libraries, containing either "libqt.so.2" or "libqt.so.3".

  • If you use Qt2:

    Check whether tmake is installed by typing
    which tmake
    If not (i.e. if the command does not give you the path to tmake), you can get it from here (and probably from you Linux CD).
    Follow the installation instructions for tmake. Change directory to "evolot/sources".  Type
    tmake -o Makefile evolot.pro
    You should have a "Makefile" now. Type
    This will produce an executable "evolot". Try it by typing
    If the evolot-window appears, click "Quit", and type
    mv evolot ../bin
    make clean

  • If you use Qt3

    Download the source distribution. Unpack it with
    gunzip evolot-0.2.tar.gz
    tar xvf evolot-02.tar
    Cd into the source directory and follow the standard linux installation steps, i.e.
    cd evolot-0.2
    Finally, put the executable into evolot/bin and delete the sources (I assume that you unpacked evolot.zip and evolot-02.tar.gz in the same directory).
    mv evolot/evolot ../evolot/bin
    cd ..
    rm -rf evolot-0.2*



    To check for gnuplot, type

    which gnuplot

    If you get the path to the executable, you are fine. Otherwise, it is recommended to install gnuplot, either from your Linux distribution, or by downloading it from here.


    Download the windows distribution and unpack it. Start the program by clicking on the file "evolot.exe" in the folder "evolot\bin".

    The manual refers to the Linux version. It mostly applies to the Windows version as well. Instead of calling "otmovie", you have to click on the file "gpldisplay.bat" after you started a simulation. The gnuplot graphics does not update automatically, but you have to click "okay" on the gnuplot dialog window for an update.You can store the gnuplot graphics by clicking on the upper left corner of the gnuplot window, copying the graphics to the clipboard, and pasting it into some office program.

    User's manual

    The manual is available in html,postscript, and pdf format.


    Papers using the algorithm

    Last modified 08/08/2005

    Back to Gerhard Jäger's Home Page