Replication data for: Introducing pressure for expressivity into language evolution experiments (doi:10.18710/TJUSQA)

View:

Part 1: Document Description
Part 2: Study Description
Part 5: Other Study-Related Materials
Entire Codebook

Replication data for: Introducing pressure for expressivity into language evolution experiments

doi:10.18710/TJUSQA

DataverseNO

2014-06-16

1

Berdicevskis, Aleksandrs, 2014, "Replication data for: Introducing pressure for expressivity into language evolution experiments", https://doi.org/10.18710/TJUSQA, DataverseNO, V1

Replication data for: Introducing pressure for expressivity into language evolution experiments

doi:10.18710/TJUSQA

Berdicevskis, Aleksandrs

University of Bergen

2011

DataverseNO

The Tromsø Repository of Language and Linguistics (TROLLing)

Berdicevskis, Aleksandrs

2014-06-16

2012

https://doi.org/10.18710/TJUSQA

Arts and Humanities, expressivity, regularity, language evolution, horizontal transmission, vertical transmission, compositionality

Field: other, Time-depth: diachronic, Time-depth: synchronic, Topic: others

48 secondary school students participated in the experiment, all native speakers of Russian, mean age 15.4 years. Participants were rewarded with candies for successful rounds (see below). Participants were asked to learn an "alien" language and to communicate in it in pairs. During the learning stage the alien language was presented to the participants on a computer screen as a set of string-picture pairs. The members of a pair were trained on the same language: the first pair in a chain on the randomly generated one, and the following pairs on the output of the preceding pair (see below). The training took place simultaneously, but in different rooms. The meaning space was determined by three parameters: shape (four possible values); colour (two); background (three). The design was made asymmetrical (4x3x2) in order to keep the meaning space less predictable. The strings for the input language of the first pair were generated randomly in lowercase Cyrillic letters (represented in Roman transliteration here). After the learning stage, the communication stage took place. Two participants were seated in the same room at different sides of a screen, so that they could not see each other. The only thing they knew was that their partner was an "earthian" who had received the same training. Voice communication was forbidden. One of the participants (A) was shown a random picture on the screen of a laptop. She wrote a name for it on a paper card. The card was passed to the participant B who depicted on it a picture that, in her opinion, corresponded to that name. If the depiction coincided with the original stimulus, this round was considered successful and the p articipants were awarded a point. The original stimulus was then shown to B, and the card with the depiction was shown to A, so that they both knew their result. When writing a name for a figure, participants used a black pen, when drawing a picture, they could use any of four coloured pencils (red, blue, green, yellow). For the next round, the roles were exchanged, and the next picture was shown. Such rounds were held for all the 24 stimuli (presented in random order, each stimulus once). Thus, each of the participants performed 12 operations of naming a picture and 12 operations of drawing a picture for a given name. The communication system shared by A and B was considered to consist of the signs used in successful rounds. In other words, if the participants won a round, the name on the card was considered a word of their shared language, its meaning being the picture on the same card. This output language (E-language, as opposed to each participant's I-language) served as input for the next pair. If the number of signs in an E language (n) was larger than 12, then n-12 signs (chosen randomly) were removed. This rule was applied to the first input language as well. If n did not exceed 12, no bottleneck was applied. If n was smaller than 2, then language was considered extinct (evolutionary dead-end), since the pilot experiments showed that in this case the participants were unable to perform the communication task. Thus, the number of signs in the input languages was not fixed and could range from 2 to 12. Three chains (2, 4 and 5) ended up at the first round, since the pairs produced E language consisting of one sign only. This shows that the tasks offered to the participants were really difficult. Since the results of these pairs are not indicative of the cumulative evolution of language, they are not included into further analysis, which is thus limited to the successful chains. Three pairs struggled successfully through the first round, which resulted into three chains (1, 3 and 6) of seven generations each. The results are provided in the csv-file.

Russia

experiment

Berdicevskis, Aleksandrs. 2012. Introducing pressure for expressivity into language evolution experiments. In: The Evolution of Language: Proceedings of the 9th International Conference on the Evolution of Language, edited by T. C. Scott-Phillips, M. Tamariz, E. Cartmill & J. R. Hurford, 64-71. Singapore: World Scientific.

10.1142/9789814401500_0009

Berdicevskis, Aleksandrs. 2012. Introducing pressure for expressivity into language evolution experiments. In: The Evolution of Language: Proceedings of the 9th International Conference on the Evolution of Language, edited by T. C. Scott-Phillips, M. Tamariz, E. Cartmill & J. R. Hurford, 64-71. Singapore: World Scientific.

Label:

Berdicevskis_evolang_data.csv

Text:

Results of a language evolution experiment

text/plain; charset=US-ASCII