TOWARD AN EXPERIMENTAL METHOD FOR STUDYING THE EMERGENCE OF HUMAN 
COMMUNICATION SYSTEMS

Bruno Galantucci, Carol A. Fowler and Michael J. Richardson

Haskins Laboratories and University of Connecticut 
bruno.galantucci@haskins.yale.edu

Ludwig Wittgenstein proposed that we understand language under the
guidance of two related metaphors: The metaphor of language as a game and
the metaphor of words as tools (Wittgenstein, 1953). In the fifty years
elapsed since Wittgenstein's proposal, the metaphors have turned into fully
fledged theories (Clark, 1996; Millikan, 2004). One of the central tenets of
these theories is that our understanding of natural languages should be
grounded in the study of the procedures humans develop to coordinate their
cooperative activities. Methods that allow systematic investigations of the
emergence of human communication systems provide ideal opportunities to test
these theories. Typically, investigators study the emergence of human
communication systems either via natural experiments (e.g., Bickerton, 1981;
Goldin Meadow & Feldman, 1977; Kegl, 1994) or via simulations with
artificial agents (Cangelosi & Parisi, 2001). The latter option has the
desirable feature of allowing experimental control, the former that of
preserving the richness of human behavior. This paper introduces a method
that combines the two features.

In brief, the method focuses on how communication procedures emerge from
the exchange of visual signals between adults. Two participants play a
videogame with two interconnected computers located in different rooms.
Players do not know each other's identity but share the virtual environment
of a game that requires them to communicate for coordinating their moves.
However, the experimental set-up does not support the use of conventional
communication systems. Spoken language is not possible because there is no
audio channel. Visual communication is possible, but occurs through a device
whose parameterization effectively prevents the use of handwriting or print.
More in detail, players communicate with each other by using a magnetic
stylus on a small digitizing pad. The resultant tracings are relayed to both
players' screens and quickly fade. Crucially, while the horizontal component
of the player's movements directly controls the horizontal movements of the
trace on the screen, the trace's vertical component is independent from the
player's movements, moving with a constant downward drift. In such
conditions, the use of letters or other common graphic symbols such as
numbers is practically impossible. (The problem players face is similar to
the problem one would encounter in making a polygraph write words or
numbers.) Hence, success in the game depends on the pair's ability to
generate a novel communication system that befits the available channel.

The paper presents a study conducted with thirteen pairs of 
participants that is designed to assess the viability of the method 
for research. In brief, the study demonstrates that the method is 
viable. Pairs developed novel communication systems in a reasonable 
time-frame, often adopting fairly sophisticated communication systems. 
The results of the study will be discussed in the context of an 
assessment of the potential and the limitations of the proposed method 
for research on human communication.

Bickerton, D. (1981). Roots of language. Ann Arbor MI: Karoma. 

Cangelosi, A.  & D.  Parisi, Eds.  (2001).  Simulating the evolution of
language.  London, Springer-Verlag. 

Clark, H.  H.  (1996).  Using language.  Cambridge: Cambridge University
Press. 

Goldin Meadow, S., & Feldman, H.  (1977).  Development of language-like
communication without a language model.  Science, 197(4301), 401-403. 

Kegl, J.  (1994).  The nicaraguan sign language project: An overview. 
Signpost, 7(1), 24-31. 

Millikan, R.  G.  (2004).  Varieties of meaning.  Cambridge,
Massachusetts: MIT Press. 

Wittgenstein, L.  (1953).  Philosophical Investigations (G.  E.  M. 
Anscombe, Trans.).  Oxford: Basil Blackwell.