Jonathan McCollum and Dr. David Bowie, Linguistics
Language is an ever changing method of communication. Linguists, in their systematic studies of languages, attempt to explain the changes that modify and sometimes eliminate languages altogether. Much research has been employed to discover the methods that keep language continually evolving.
Mary S. MacKeracher, at the University of Toronto, has performed important research in regards to the spread of language change.1 She has produced mathematical models that parallel the aspects of language change. These models describe language change and display the rate at which modifications are spread.
Under Professor Bowie’s guidance, I attempted to formulate computer models that mimic the mathematical models of MacKeracher. To create the computer models I worked with the program Netlogo 1.1. Netlogo 1.1 is a variation of the computer program Starlogo, created by Mitchel Resnick. In the 1980s, Resnick, interested in decentralized computer modeling, developed a new version of the Logo programming language known as Starlogo. One might recognize Logo as the “turtle” computer program used in many elementary schools. Resnick’s idea for Starlogo was to allow the programmer to explore the interaction between individual turtles (the small two dimensional characters of the program). Thus any number of turtles could be made to interact in a nearly infinite number of ways. Resnick has used his program to model everything from traffic jams to the spread of diseases.2 Using the updated Netlogo 1.1, I began to explore the possibilities of tracking the change of a language.
My first priority was to learn the Netlogo language. Although not as complex as Java or C++, Netlogo presented some challenges in learning all its capabilities. Turtles can receive innumerable traits and can act in a variety of ways. On top of learning all the commands for the turtle, the environment of the turtles can also be programmed in a variety of ways. Furthermore there are many options that help track and calculate the results of each model.
Having learned the different options I could implement, I set about programming a model that would provide the viewer with a visual representation of the language change process. It was easy to facilitate the spread of a change in the language of the turtle community. I chose to use color as the modifier that spread amongst the turtles. A change in the color of a turtle represented the adoption of a difference in speech. This modifier could symbolize the pronunciation of certain words or the usage of new words. It was not simple to develop a model that would simultaneously mimic the mathematical models of Mackeracher and effect a change that would spread rapidly throughout the community.
The first difficulty I encountered was matching the mathematical models as I spread a color throughout the community. Mackeracher purports that change in language follows an S-shaped curve as it spreads. Initially the modification of a language infects few speakers. As time progresses the expansion accelerates peaking and tapering off leaving few unaffected by the change. The initial slow progression and the subsequent boom in change was easily produced in my early models, nevertheless, it proved time consuming to conceive a model that would leave some turtles out of reach.
Ultimately Spain brought the inspiration necessary to program a model that would spread rapidly and then taper off. I recalled that the stubborn Catalans of Northeastern Spain who refuse to give in to the pressure of adopting the Castilian language (what we know as Spanish). I programmed a model mimicking this situation. I created two locals that drew the Turtles to them. Once within the boundaries of these two areas the turtles would change color. One area, Madrid, was more attractive than the other, Catalunya, pulling in more turtles. Madrid changed the turtles from their normal green to yellow as Catalunya changed them to blue. The yellow turtles grew slowly then more rapidly, however; they were unable to overrun the stubborn blue turtles of Catatlunya, which remained a tiny minority within the population that resisted the change to yellow.
After successfully creating an S-curve model I encountered some difficulties with Netlogo 1.1. I attempted programming a trait that would infect all turtles through contact with an infected turtle. Unfortunately the Netlogo system was unable to calculate the strict turtle to turtle interaction. This programming experience demonstrated a great weakness of the program and the need for improvement to achieve a completely decentralized method to programming. Although Netlogo did have its drawbacks it proved a powerful tool in examining language change. The programming process improved my understanding of how language can be adopted and changed by a group. Individuals, however, proved more difficult model