Special Issue "Complexity of Human Language and Cognition"

Guest Editor
Dr. Ramon Ferrer i Cancho
TALP Research Center, Departament de Llenguatges i Sistemes Informatics, Universitat Politecnica de Catalunya, Campus Nord, Edifici Omega Jordi Girona Salgado 1-3, 08034 Barcelona (Catalonia), Spain
Website: http://www.lsi.upc.edu/~rferrericancho/
E-Mail:
Interests: applications of information theory and network theory to biology and linguistics; quantitative linguistics; corpus linguistics; origins and evolution of language

This special issue is devoted to shed light on the complexity of human language and cognition from a transdisciplinary perspective. Much recent progress on the complexity of human language and cognition from network and information theory is scattered in the literature of physics, cognitive science, neurology, biology, quantitative linguistics...It is time to bring together efforts from different areas in a more coherent environment to show that new insights into different areas of human language and cognition can originate from common theoretical approaches.

There is no a priori constraint on the aspect of human language or cognition addressed. There is no constraint either on the background of the contribution: physics, cognitive science, neurology, biology, quantitative linguistics... The only condition is that the insight comes from a theoretical discipline such as information theory or network theory. Theoretical insights from other theoretical disciplines might be considered.

Ramon Ferrer i Cancho, Ph.D.
Guest Editor

Submission

All papers should be submitted to entropy@mdpi.org with copy to the guest editor. To be published continuously until the deadline and papers will be listed together at the special websites. Both, research articles and review articles are invited. For planned papers, a title and short abstract (about 100 words) can be sent to the Editors for announcment on this website.

Submitted papers should not have been published previously, nor be under consideration for publication elsewhere. All papers are refereed through a peer-review process. A guide for authors, sample copies and other relevant information for submitting papers are available on the Instructions for Authors page. Entropy is an international peer-reviewed quarterly journal published by Molecular Diversity Preservation International.

Please visit the Instructions for Authors page before submitting a paper. Open Access publication fees are 800 CHF per paper. English correction fees (250 CHF) will be added in certain cases (1050 CHF per paper for those papers that require extensive additional formatting and/or English corrections).

• complexity
• language
• cognition
• information theory
• graph theory
• network theory
• quantitative linguistics
• statistical mechanics

Type of Paper: Article
Title: Information And Uncertainty: Some Applications In Experimental Psychology
Author: Donald Laming
Affiliation: Department of Experimental Psychology, University of Cambridge, Cambridge, CB2 3EB, UK; E-Mail: drjl@cus.cam.ac.uk
Abstract: This paper sets out the relationships between information (statistically defined), statistical hypothesis testing, channel capacity in a communication system, uncertainty, and the concept of entropy in thermodynamics. Illustrative applications are grouped under
(a) the human operator viewed as an ideal communications channel;
(b) the human operator treated as a physical system, and
(c) Bayes’ theorem as an algorithm for combining information from different sources.
While the idea of the human operator as an ideal communications channel has long since been abandoned, treating the human operator as a purely physical system provides a platform for the exploration of many aspects of human performance by analogy with systems analysis. Finally, the use of Bayes’ theorem to calculate the effects of prior probabilities and stimulus frequencies on human performance is probably misconceived.

Type of Paper: Article
Title: A Graph-Theoretical Model of Alignment in Communication
Author: Alexander Mehler
Affiliation: Faculty of Technology, Bielefeld University, D-33594 Bielefeld, Universitätsstraße 25; E-Mail: Alexander.Mehler@uni-bielefeld.de
Abstract: Evolution has made dialogical communication one of the most efficient means of human interaction. A procedural characteristic of this communication is alignment (Pickering & Garrod 2004), that is, the structural coupling of interacting agents in terms of their functionally coupled or similarity enhanced representations. Alignment is one of the resources that affect the speed, cognitive ease and memory efficiency of communication. Unlike coordination, alignment is an automatic and resource-sparing process that operates on several linguistic levels (e.g., phonological, syntactical and semantic). In this paper, we present a graph-theoretical model of alignment in dialogical communication. We introduce the notion of a bipartite time-aligned network series to measure the alignment of interacting agents. This is done with a focus on dialogue lexica as the time-aligned output of their communication. The paper introduces several entropy measures of such lexica - modeled as linguistic networks - to make alignment a measured variable in quantitative linguistics.

Type of Paper: Article
Title: Keyword Detection in Written Texts Based on Statistical Mechanics Consideration
Authors: K. Koroutchev ¹ and E. Korutcheva ^2,3
Affiliations: ¹ Escuela Politécnica Superior, Universidad Autónoma de Madrid, 28049 Cantoblanco, Madrid, Spain
² Dpto. Física Fundamental, Universidad Nacional de Educación a Distancia, c/Senda del Rey, No 9, 28080 Madrid, Spain; E-Mail: elka@fisfun.uned.es
³ G.Nadjakov Inst. Solid State Physics, Bulgarian Academy of Sciences, 1784, Sofia, Bulgaria
Abstract: In this article we present a model of human written text based on statistical mechanics consideration. The text is conditioned to the language in which it is written, being represented by a large text corpus. As a first approximation we use the words as elementary units composing the text, which interact with the language. We find that the Gamma distribution for the word's distribution is a reasonable hypothesis that is well supported by the data. By using general principles from the statistical physics, we derive empirically the potential energy for the different parts of the text and we calculate the thermodynamic parameters of the system. The derivation shows that the "specific heat'' parameter reflects important differences related to the function of the words and effectively separates the function words (the grammar) from the specific terms (the keywords) used in the text. This property preserves even if multi-word ensembles are used as a unit. Our approach gives advantages when used in a text searching mechanisms as it does not use any empirically introduced probabilities measures and one does not need any a priori knowledge about the words class separation. Finally we give an example of a possible application of the method in information retrieval query disambiguation, and briefly discuss some future directions of the present investigation.