Although network images depicting the development of species and languages can be traced back to the 18th century in both linguistics and biology, models of reticulation instead of the powerful tree representations have only recently gained growing interest in the sciences and humanities and are now widely used on a formalized basis. In biology, research in procaryot evolution suggests lateral gene transfer as a major feature in the development of bacteria. In the field of linguistics, mutual lexical and morphosyntactic borrowings between languages as well as the wave-like distribution of innovations seem to be much more central for language evolution as the family tree model is likely to concede. In the humanities, networks are employed as an alternative to established phylogenetic models to express the hybridization of cultural phenomena.

The international conference ‘Bridging Disciplines: Evolution and Classification in Biology, Linguistics and the History of Sciences’ was held at Schloss Reisensburg (Günzburg), Science and Convention Center of Ulm University, from June 24th to June 26th, 2011, to encourage interdisciplinary exchange between scholars from the sciences and humanities. The conference was organized by Heiner Fangerau (Ulm), Hans Geisler (Düsseldorf) and William Martin (Düsseldorf) as part of the research project ‘Classification and Evolution in Biology, Linguistics and the History of Science’1 which had been generously funded by the German Federal Ministry of Education and Research (BMBF) since 2009. Biologists, linguists, historians and philosophers of science from eight countries explored in detail whether networks might provide a more appropriate or at least complementary model describing processes of evolutionary development than pedigree models.

The first session was dedicated to the field of biology (’New ideas of evolution: Networks in Biology’). In her introductory lecture about “Phylogenomic networks,” TAL DAGAN (Düsseldorf) explained the fundamental significance of networks for the study of evolutionary processes in gene and species evolution. As Dagan stated, current approaches to genome phylogenies usually operate within the framework of phylogenetic trees, but several evolutionary processes in nature would not be tree-like, e.g. hybridization, genome fusions, and lateral gene transfer (LGT). Because networks comprise entities (vertices) connected by pairwise relations (edges), Dagan argued, approaching genome evolution with networks rather than trees would enable the reconstruction of both vertical and lateral gene transfer events. Yet phylogenomic networks have mainly been used to study genome evolution in prokaryotes and bacteriophages where lateral gene transfer is a common mechanism of natural variation. This important limitation of this approach, which was also mentioned by Dagan, was emphasized repeatedly in the course of the following presentations and discussions.

In his case study on “Lateral gene transfer (LGT) in prokaryote evolution,” OVIDIU POPA (Düsseldorf) examined the nature and magnitude of barriers to lateral transfer. To investigate donor-recipient events of recent lateral gene transfer among sequenced prokaryote genomes, he had implemented so-called directed networks. His presentation led deeply into the methodological and technical foundations of molecular genetics.2

“What genes can tell us about the evolutionary history of species,” asked INGO EBERSBERGER (Vienna), shifting the emphasis to the level of eukaryotes as ‘higher’, more complex organisms than eucaryots. According to Ebersberger, distinct phenotypes in individual species are determined by the repertoire and the functional interplay of genes in individuals. With regard to evolutionary processes it is essential that genes and their respective functions are subject to permanent change. As key problem Ebersberger outlined “the accurate identification of genes whose genetic lineages split due to a speciation event, the interference of population genetic effects when species split in close temporal succession, and methodological artefacts during tree reconstruction due to model violations.” In the discussion following his talk, Ebersberger argued in favor of repeatedly checking the methodological approach and the consistency of data in genome research. In Ebersberger’s opinion, phylogenetic structures representing evolutionary processes are often too hastily abandoned in favor of network models due to missing or insufficient data.

Talking about ”Network genes: how can they inspire linguistics”, PHILIPPE LOPEZ (Paris, together with ERIC BAPTESTE) initiated the discussion on methodological transfer between biology and linguistics. Based on the network approach used in the study of microbial evolution in E.coli, he showed how the same method can be used to explain the evolution of hundreds of words of the Chinese vocabulary, how this approach could also be used to foster original hypotheses about how many words in multiple dialects evolved.

The second session adressed ‘Phylogenetic classifications and network approaches in linguistics’ and was started by the well-known historical linguist GEORGE STAROSTIN (Moscow) who continues working on the famous ‘Tower of Babel’ project initiated by his father.3 In his methodological contribution ”Lexicostatistics as a basis for language classification,” Starostin argued for improvements of this method originally proposed by Morris Swadesh to establish a genetic classification of languages giving time-horizons for branching processes (glottochronology) based on lists of basic vocabulary.

Maybe even more controversial as the lexicostistical approach in linguistics is the application of phylogenetic models to cultural evolution that SIMON GREENHILL (Auckland) presented. Greenhill is a member of the research group of Russel Gray that pioneered the application of computational evolutionary methods to questions about linguistic prehistory and the sequence and timing of the peopling of the South Pacific. In his paper on ”The shape and fabric of human history” he examined ’the extent to which lexical evolution is tree-like in different parts of the world’ and tried to ‘evaluate the coherence of cultural and linguistic lineages’, discussing the capacity of current computational methods derived from evolutionary biology to address these issues.

In a similar way, but reaching quite different results, NORBERT ENDRES (Greifswald), MICHAEL DUNN (Nijmegen), DAVID ERSCHLER (Tübingen) and SHIJULAL NELSON-SATHI (Düsseldorf) examined the possibilities and limitations of phylogenetic methods and network approaches in comparative linguistics. Endres (”Evaluating phylogenetic methods for the purposes of historical linguistics”) gave an overview on commonly used methods, putting his emphasis on data input and the underlying mathematical principles of calculation techniques. Dunn, in Language networks and reconstructed protolanguages, presented a revised phylogeny of the Indo-European family based on a Bayesian phylogenetic analysis of the substantially corrected version of the Dyen et al. (1992) word list. Erschler (”On feature transmission in linguistic areas: The case for co-Evolution”) analyzed the parallel emergence of certain rare grammatical features in Georgian and Ossetian (two geographically neighboring but not genetically related languages spoken in the Caucasus), focussing on the question whether shared morphosyntactic features ’are literally transmitted or may co-evolve in contacting languages’. Using the frequency of hidden borrowings among over two thousand cognates in the ’basic vocabulary’ of 84 Indo-European languages, Nelson-Sathi (”Networks uncover hidden lexical borrowing in Indo-European language evolution”) could show that borrowing in languages forms a significantly non-tree-like process similar to horizontal gene transfer in genome evolution. In both cases, this process cannot be reconstructed from a pedigree-model.4 Concerning the study of the syntactic - rather than the lexical – evolution of languages, a quantitative approach was presented by GABRIELE RIGON (Trieste) in his paper “Increasing accuracy of syntactic distances in the biolinguistic perspective”.

JELENA PROKIĆ (Munich, together with JOHN NERBONNE, Groningen) demonstrated the successful application of multi-sequence alignment techniques taken from bioinformatics to dialectromety in her paper ”What can dialect geography tell us about diachronic variation of language?” She presented a study of phonetic data from different dialects in Dutch and Bulgarian in which ‘alignments are used to extract sound correspondences automatically’. Closely related to Prokić’s approach, JOHANN-MATTIS LIST (Düsseldorf) presented a new sound class based algorithm for “Multiple sequence alignment in historical linguistics”.5

The third session ‘Scientific concepts and investigative practices - Networks in science‘ highlighted the transfer of ideas and methods between biology and linguistics in a historical perspective and finally extended the study of interdisciplinary transfer to the bio-linguistic discourse. THIERRY HOQUET (Paris) emphasized the ‘power of words’ in his talk on the “Translation of the term ’natural selection’”, mostly in French and German. He showed why Darwin had maintained the term ‘natural selection’ despite the significant dispute the term had provoked after the first publication of the Origin of Species (1859) and subsequent translations. In accord with Hoquet, LÁSZLÓ KOVÁCS (Tübingen) expressed the view that the change in meaning of terms might form a base for their conceptual understanding. He examined the evolution of the term ’genetic information‘ and stated that ‘Applying natural selection to genetic information we can prove or refute the theory of evolution of linguistic terms.’

Adressing the methodological transfer between biology and archaeology, KATJA RÖSLER (Freiburg), in “Metaphors and artefacts: Archaeological practice between biology and history”, reflected the methodological proximity of these two disciplines which is based on the that they both deal with things that are withdrawn from human grasp (nature) or have been so for a long time (culture without tradition), as Rösler stated.

In his contribution ”Common traditions in mapping human linguistic and biological diversity,” FRANK KRESSING (Ulm) emphasized the long tradition of the idea of human biological and cultural co-evolution in European-based intellectual thought covering both ‘man’s physical and cultural dimensions.’ Based on methods of the digital humanities, MATTHIS KRISCHEL (Ulm) could, in ”Historical network analysis can be used to construct a social network of 19th-century evolutionists,” present empirical evidence for the existence of a social, intellectual, and institutional network of biologists, linguists and anthropologists who contributed to the development of a unified theory of co-evolution of human languages, populations, and cultures in the 19th century.6

In a kind of meta-analysis (“Creation of concepts in interdisciplinary groups”), HANNE ANDERSEN (Aarhus) reflected on basic mechanisms of interdisciplinary approaches, analyzing ‘how scientists involved in interdisciplinary collaboration link concepts originating in different disciplines or research fields to each other and subsequently develop new concepts that cut across disciplinary boundaries’.

In the summarizing discussion, Tal Dagan, Hans Geisler (Düsseldorf) and Heiner Fangerau (Ulm) reflected on the contextualization of theories of evolution in history, linguistics and biology. They argued for the transdisciplinary testing and adaptation of different concepts of evolution but at the same time they stressed the necessity to keep in mind the specific objects of investigation of each discipline. The objects of investigation limit the application of an overarching concept of evolution and demand for disciplinary adjustments.

However, the value of the method of network approaches to examine and explain evolutionary processes in history, linguistics and biology was demonstrated by most of the contributions. Nevertheless, it has to be stressed that network approaches are more a method than a theory and that their methodological character again force the researcher to adapt them to his or her object of research.

Conference overview:

William Martin, Hans Geisler, Heiner Fangerau (Düsseldorf / Ulm): Welcome and introduction

First Session: New ideas of evolution: Networks in Biology
Chair: William Martin (Düsseldorf)

Tal Dagan (Düsseldorf): Phylogenomic networks

Ingo Ebersberger (Vienna): What genes can tell us about the evolutionary history of species

Philippe Lopez (Paris): Gene network: how can they inspire linguistics

Ovidiu Popa (Düsseldorf): Directed networks reveal genomic barriers and DNA repair bypasses to lateral gene transfer among prokaryotes

Second Session: Phylogenetic classifications and network approaches in linguistics
Chair: Hans Geisler (Düsseldorf)

George Starostin (Moscow): Lexicostatistics as a basis for language classification: increasing the pros and reducing the cons

Simon Greenhill (Auckland): The shape and fabric of human history

David Erschler (Tübingen): On feature transmission in linguistic areas: The case for co-evolution

Michael Dunn (Nijmegen): Language networks and reconstructed protolanguages

Shijulal Nelson-Sathi (Düsseldorf): Networks uncover hidden lexical borrowing in Indo-European language evolution

Jelena Prokić (Munich): What can dialect geography tell us about diachronic variation of language?

Johann-Mattis List (Düsseldorf): Multiple sequence alignment in historical linguistics - A sound class based approach

Gabriele Rigon (Trieste): Increasing accuracy of syntactic distances in the biolinguistic perspective

Norbert Endres (Greifswald): Evaluating Phylogenetic Methods for the purposes of historical linguistics

Third Session: Scientific concepts and investigative practices - Networks in science
Chair: Heiner Fangerau (Ulm)

Thierry Hoquet (Paris): Translating natural selection

László Kovács (Tübingen): Evolution of the term "genetic information" - following Darwins theory?

Katja Rösler (Freiburg): Metaphors and Artefacts: Archaeological practice between biology and history

Hanne Andersen (Aarhus): Bridging disciplines: Creation of concepts in interdisciplinary groups

Matthis Krischel (Ulm): Historical network analysis can be used to construct a social network of 19th-century evolutionists

Frank Kressing (Ulm): Common traditions in mapping human linguistic and biological diversity – a result of bridging borders between scholarly disciplines and personal networks of scientists

Hans Geisler, Heiner Fangerau (Düsseldorf / Ulm): Summarizing discussion: Genes, words, ideas: Is a unified evolutionary theory reasonable?

Notes:
1http://www.evoclass.de
2 Ovidiu Popa, E. Hazkani-Covo, G. Landan, William Martin, Tal Dagan, Directed networks reveal genomic barriers and DNA repair bypasses to lateral gene transfer among prokaryotes, in: Genome Res 21 (2011), pp. 599-609.
3http://starling.rinet.ru/main.html
4 Shijulal Nelson-Sathi, Johann-Mattis List, Hans Geisler, Heiner Fangerau, Russell D. Gray, William Martin, Tal Dagan, Networks uncover hidden lexical borrowing in Indo-European language evolution. Proc Biol Sci 278 (2011), pp. 1794-803.
5http://lingulist.de/lingpy
6 Matthis Krischel, Frank Kressing, Heiner Fangerau, Die Entwicklung der Deszendenztheorie in Biologie, Linguistik und Anthropologie als Austauschprozess zwischen Geistes- und Naturwissenschaften, in: Matthis Krischel, H.-K. Keul (eds.), Deszendenztheorie und Darwinismus in den Wissenschaften vom Menschen. Stuttgart 2011, pp. 107-121.