Human language defines one of the major transitions of evolution. The emergence of such an elaborated form of communication allowed humans to create extremely structured societies and manage symbols at different levels including, among others, semantics. All linguistic levels have to deal with an astronomic combinatorial potential that stems from the recursive nature of languages. This recursiveness is indeed a key defining trait. However, not all words are equally combined nor frequent. In breaking the symmetry between less and more often used and between less and more meaning-bearing units, universal scaling laws arise. Such laws, common to all human languages, appear on different stages from word inventories to networks of interacting words.
Among these seemingly universal traits exhibited by language networks, ambiguity appears to be a specially relevant component. Ambiguity is avoided in most computational approaches to language processing, and yet it seems to be a crucial element of language architecture. Ambiguity is shown to play an essential role in providing a source of language efficiency, and is likely to be an inevitable byproduct of network growth.
In our lab we have studied the organization of complex language networks including co-occurrence of words within written texts, syntactic graphs, the development of language in children and semantic webs. Among other things, we discovered that language networks are typically heterogeneous, with a scale-free distribution of connections among words: most words have one, two or asmall number of relations with other words,whereas a handful of them are hubs, i. e. have a very large number of links. Heterogeneity is strongly tied to ambiguity, and has been shown to be extremely important in order to make these webs navigable. In our study of language acquisition in children, we observed that there is a sharp transition between a tree-like and a scale-free web around the two-year critical zone. This transition cannot be explained by any simple model of percolation, suggesting in fact that something is "hardwired" in the brain in order to acquire language.
We study these problems using network models, statistical physics and mathematical models. Our main interests are: (1) understanding the transition to language in evolution, (2) the sudden jump in language complexity through ontogeny and (3) how language is stored and retrieved in neural networks.
Ambiguity in language networks. Ricard V. Solé and Luís F. Seoane. The Linguistic Review 32, 5-35 (2014)
Towards a mathematical theory of meaningful communication Bernat Corominas-Murtra, Jordi Fortuny and Ricard Solé. Scientific Reports, 4, 4587 (2014)
Lenguaje, redes y evolución. Ricard V. Solé, Bernat Corominas-Murtra and Jordi Fortuny. Investigación y ciencia 440 (Mayo 2013), 58-67
On the Origin of Ambiguity in Efficient Communication. J Fortuny and B Corominas-Murtra. Journal of Logic, Language and Information, 22(3), 249-267 (2013)
Emergence of Zipf’s law in the evolution of communication. Bernat Corominas-Murtra, Jordi Fortuny and Ricard Solé. Physical Review E, 83(3), 036115 (2011)
The semantic organization of the animal category: evidence from semantic verbal fluency and network theory. Joaquín Goñi et al. Cognitive Processing 12, 183-196 (2011)
Diversity, competition, extinction: the ecophysics of language change. R Solé, B Corominas-Murtra and J Fortuny. J. Royal Society Interface 7, 1647-1664 (2010)
Emergence of Scale-Free Syntax Networks. B Corominas-Murtra, S Valverde and R Solé. In: Evolution of Communication and Language in Embodied Agents. Springer-Verlag Berlin, p.83-101 (2010)
Language networks: Their structure, function, and evolution. R. Solé, B Corominas-Murtra, S Valverde and L Steels. Complexity, 15(6), 20-26 (2010)
The ontogeny of scale-free syntax networks: phase transitions in early language acquisition. B Corominas-Murtra, S Valverde and R Solé. Advances in Complex Systems, 12(03), 371-392 (2009)