The Semiotics of Consensus

Impact of Network Topology on Communication Strategies in Spanish Language Interaction Networks

by Tobias Gretenkort (Author)
©2023 Thesis 316 Pages


This book draws from graph theory and a semiotic comparison between language and distributed ledger technologies (also known as Blockchains) to motivate three experiments on language and network structure. The work explores the importance of this concept in different areas of linguistic research and establishes elements of a tentative linguistics of networks. Its empirical investigation is based on data from threads posted to the imageboard Hispachan, which often displays radicalized language and hate speech. The experiments (based on topic modeling and sentiment analysis) reveal an impact of the network structure of interaction on the interaction itself as well as the use of ingroup signalling and emotionally charged vocabulary to expand the network of interaction.

Table Of Contents

  • Cover
  • Title
  • Copyright
  • About the author
  • About the book
  • This eBook can be cited
  • Table of Contents
  • List of Abbreviations
  • Introduction
  • Linguistics and Technology
  • Enrichment of the Linguists’ Toolbox
  • The Emergence of New Linguistic Fields
  • Technology Becomes Theory – The Computer Metaphor
  • The Computer Metaphor in Linguistics
  • Artificial Neural Networks
  • Generative Transformational Grammar and Automata Theory
  • Approach and Structure
  • A Semiotics of the Blockchain?
  • Terminology – What Is the Blockchain?
  • Distributed Ledger Technology – DLT
  • Blockchain
  • Cryptocurrency
  • Bitcoin
  • Double Spending
  • Proof-of-Work (PoW)
  • Consensus as a Problem of Communication
  • Semiotic Properties of Consensus
  • Representation in Language and Distributed Systems
  • Perlocution in Language and Distributed Systems
  • Self-Containment of Information in Language and Distributed Systems
  • Graph Theory and Network Analysis – Formal Definitions
  • Graphs
  • Graphs
  • Nodes
  • Edges
  • Paths
  • Ledger
  • Ledger Implementation – Network Topology
  • Linear Topology
  • Centralised Topology
  • Decentralised Topology
  • Fully-Connected Topology
  • Distributed Topology
  • Network Properties
  • Degree Centrality
  • Graph Centrality
  • Network Density
  • P-Regularity
  • Diameter
  • Achievement of Consensus in Distributed Networks
  • The Byzantine Generals Problem in Fully-Connected Networks
  • The Oral Messages Algorithm
  • The Signed Messages Algorithm
  • Non-fully Connected Networks
  • Network Theories in Linguistics
  • Language Typology
  • Cognition and Linguistic Networks
  • Semantic Networks
  • Syntactic Networks
  • Limitations of Linguistic Networks in Individual Cognition
  • Social Cognition and Social Networks
  • Networks in Language Development
  • Cultural Transmission
  • Experimental Semiotics
  • Concluding Remarks on Social Network Theories
  • Online Social Networks
  • A Linguistics of Social Networks
  • Language as a Complex Adaptive System
  • Language as a Solution to the Consensus Problem
  • Parallel Fixation of System and Information
  • Network Topology as a Typology of Interaction
  • Degree Bias
  • Identity Bias
  • Degree Distribution Bias
  • Content Bias
  • Concluding Remarks and Experimental Predictions
  • Data
  • Hispachan – Anonymous Linguistic Interaction Networks
  • Data Collection
  • Data Pre-processing
  • Node-specific Dataset
  • Subgraph Dataset
  • Dataset – Descriptive Statistics
  • Thread Length and Diameter
  • Representation of Communities
  • Graph Topology
  • Node Degrees
  • Graph Centralities
  • Experiment One – Influence of Network Topology on Consensus
  • Materials and Methods
  • Analysis
  • Results
  • Discussion
  • Experiment Two – Identity Bias and Secret Handshakes
  • Materials and Methods
  • Analysis
  • Results
  • Discussion
  • Experiment Three – Content Bias: Expressivity and Local Engagement
  • Materials and Methods
  • Analysis
  • Results
  • Discussion
  • General Discussion
  • Adding to the Network Approach to Language – Ecological Validity
  • Internal and External Validity – Adequacy of the Blockchain Semiotics
  • Global Graph Topologies and Lack of Distributed Control
  • Local Effects and Alternative Communication Strategies
  • Conclusion
  • Bibliography
  • Appendix
  • List of Tables
  • List of Figures

0. Introduction

Sie erweisen mir die Ehre, mich aufzufordern,

der Akademie einen Bericht über mein

äffisches Vorleben einzureichen.

Ein Bericht für eine Akademie – Franz Kafka (1917)

Language and technology have a longstanding cultural relationship, which is now more relevant than ever. While technological innovations such as scripture, printing, telegraphing, and the telephone have continuously sped up communicative processes, the Internet supposes a more fundamental change in communicative practices in the 21st century. This is also abundantly accounted for in linguistics, most notably by the description of linguistic innovations in both English (e.g., Crystal 2004; McCulloch 2019) and Spanish (e.g., Yus 2001, 2008; Tascón 2012). This doctoral dissertation aims to investigate more specific mechanisms which lead to this linguistic variation. As the Internet has led to globalised practises in communication and the creation of a dense network of communicative pathways, it is the aim of this study to investigate the role of linguistic and social network dynamics in the usage and variation of language. Changes in network properties of communication as a result of the expansion of the Internet are best illustrated by one of the most striking contradictions in modern media. Human beings in the industrialised regions of the world now have considerably more opportunities to interact with one another and have arguably gone from being mere receivers of information to being more powerful promoters of information. At the same time, much communicative power is concentrated in the hands of relatively few famous individuals and companies, as is reflected in their huge numbers of followers on online social media. Compared to this, most users have close to no following at all. It would thus appear that communication on the Internet has become at the same time more centralised and more decentralised. This paradox lies at the heart of all the following considerations.

The main goal of this doctoral dissertation is to investigate the consequences of this contradiction for the use of language and communication itself. Approaching this dynamic entails a variety of deliberations and methods. Methodologically, this work is based on three observational experiments (Chapters 6, 7 and 8) on data from a self-compiled corpus of comments on an anonymous online social media platform (Chapter 5). But to produce meaningful hypotheses for such a study, the consideration of exclusively linguistic literature is insufficient. To account for the effects of network properties, some mathematical resources from graph theory will be introduced (Chapter 2) as they provide the basis for the formal analysis of networks. The application of such formal resources shows that the linguistic literature, even when explicitly applying the concept of network, lacks a cohesive understanding of what a network is and how it is best applied to the study of language. As a review of the literature (Chapter 3) shows, contributions are either focused on the individual language capacity, on the description of linguistic variation in a dialectological or typological sense, or on the social configuration of interaction. Networks are thus applied to a variety of phenomena and levels of linguistic inquiry. Consequently, a careful exploration of different approaches is in order to identify what is meant by the term network in each case. As a baseline for comparison, another social and mathematical theory of networks will be explored and characterised with semiotic means, namely Blockchain technology (Chapter 1). As the chapter shows, lots of literature are available to establish a connection between the Blockchain’s graph-theoretic makeup – most notably its “decentralisation” and the informational and social implications it has for Blockchain-based services. It is argued that informational processes on the Blockchain can, in part, be conceptualised as communicative processes. By identifying commonalities and differences between language and the Blockchain as semiotic systems, in conjunction with the literature reviews of graph theory and linguistics, hypotheses for data analysis can be generated (Chapter 4). After reporting on the findings of three experiments (Chapters 6, 7 and 8), their significance in the context of centralisation and decentralisation is discussed in light of the reviewed literature (Chapter 9).

In this way, the current study aims to contribute to both the consolidation and clarification of the use of networks in the analysis of linguistic and semiotic phenomena and to contribute to the identification and description of some of these phenomena. Before the description of Blockchain technology, however, some remarks on the relationship between linguistics and technology are in order. These serve as justification for the inclusion of a technological description in a linguistic dissertation on the one hand and as a justification of the problem of technology and its impact on communication – and by extension language – on the other hand.

0.1. Linguistics and Technology

While the relationship between language and technology is quite apparent, this is less true of the relationship between linguistics and technology. But the systematic study of technologies from a linguistic standpoint is useful for a variety of reasons. The Blockchain is a technology that, both philosophically and mathematically, provides great insights into the properties and the design of network architectures. It is apparent that graph theory, in turn, has a great deal to offer for the study of linguistic interaction within groups and the emergent properties of these interconnected networks. This ultimately amounts to studying how linguistic systems are affected by the interactions in networks of speakers. For the sake of argument, it will be assumed that networks are the locus of language and that languages exist primarily as an emergent property of communication that takes place in communities which can adequately be described and analysed as networks. Thus, a semiotics of networks, as inspired by the Blockchain will provide predictions for communicative networks of human communities, particularly online communities.

While such an approach is eclectic in nature, preoccupation with a specific technology is nothing unusual in linguistics. This study is not the first to invoke technological descriptions for the analysis of linguistic phenomena, nor is it the first to use networks as a means of linguistic research. In point and fact, such ponderations have a long-standing tradition in linguistics. On a general scale, the question of what (a) language is has often spawned comparisons between humans and machines, as well as between humans and animals. Thus, before moving on to the description of the Blockchain, this relationship will be explored to argue why the preoccupation with (communication) technologies is a suitable endeavour for linguists.

To understand this, it is important to point out how fundamental the criterion of language is in the history of human exceptionalism, that is the belief that human beings are fundamentally different from other living things, such as animals. Philosophers and linguists have contributed to this belief by promoting the idea that only humans are capable of using language. The pertinence of this idea can be traced back at least to the book of Genesis, where God tasks Adam with naming the animals and accepting the name that Adam gives each animal. It would appear that, in Abrahamitic tradition, language is not handed down by God but that it exists through and for humans alone. Therein lies ultimately the idea that language is proper to humankind and acts as a clear-cut border between humans and all the other things in the world.

In the more recent history of thought, this idea still finds astonishing resonance. Descartes’ Discours de la Méthode, first published in 1637, draws the line between automata and human beings on language. Descartes states that no machine, however well-constructed, could ever react to human speech in a satisfactory way (Descartes 1673). Some decades later in 1746, Diderot makes a similar argument stating that human-like intelligence could be attributed to any parrot that responded intelligently to anything a person said to it (Diderot 1746). Here, Diderot defines a criterion for human intelligence and at the same time dismisses the possibility of it being found elsewhere. That is, he doesn’t expect any parrot to be found able to sensibly respond to human speech.

Even in the 20th century, much academic activity in humanist departments was devoted to highlighting that humans are special in their use of language. After Chomsky’s important reply to uprising behaviourist tendencies in linguistics, mainly represented by B.F. Skinner (Chomsky 1959a), the generative framework took on the most important role for this task in the linguistic landscape. Importantly, one of the core elements of Chomsky’s theory, as it is thought of today, is nativism (Chomsky 1965, 1968), which is the idea that language is hardcoded into the human brain. There are several degrees of strength to this claim, with the strongest version advocating that language is a capacity that arose within a fairly short period in human evolution through genetic mutation. The identification of the FOXP2 gene has cemented this idea in popular perception because it was diagnosed to be relevant to the development of human language. Specifically, its variation across different species seems to suggest that its fixation in the human organism is in alignment with the appearance of anatomically modern humans some 200,000 years ago (Enard et al. 2002). On the other hand, it seems to be neither a sufficient nor necessary condition for language as species incapable of language have this gene, while pathological mutations to this gene seem to mainly affect speech production and aspects of motricity (Fisher et al. 1998; Lai et al. 2001) of language, while cognitive abilities seem to be unaffected.

Independently of this neuroscientific debate, linguists and psychologists in some schools of thought still assume that animals do not have language. Titles such as Why Animals Don’t Have Language (Cheney & Seyfarth 1997) and The Language Instinct (Pinker 1994) are reminiscent of this tendency. The question of what distinguishes human language from other kinds of communication in nature appears to be much more prevalent in the scientific community than the question of whether this distinction is relevant at all.

But contrary positions to this stance can be found at least since the 18th century. The first relevant philosophical defiance of this idea can be found in the works of De La Mettrie, who makes a structured attempt at a description of human beings as sophisticated automata in L’Homme Machine in 1747 (De La Mettrie 1747), paving the way for materialist philosophy in the 19th century. Here, De La Mettrie argues that animals such as birds already have sophisticated language in their songs and even entertains the possibility to teach human language to an ape. At the same time, he speculates that an automaton could be endowed with language. And although he thinks that this would be achieved in the far future, his idea reduces the difference between humans and machines to an issue of degree rather than an issue of category.

However, human exceptionalism in linguistics has proven to be rather resistant to the argument that universal tendencies in language, such as word order, need not be the consequence of innateness but could be the result of cognitive (Culbertson et al. 2012), functional (Tomasello 1995) or even environmental (Nölle et al. 2020) constraints. Examples of highly advanced communicative systems in animals, such as cephalopods – octopi, cuttlefish, and squid – (Mäthger et al. 2009a, 2009b) and great apes (Bohn et al. 2016), don’t seem to compel a broader public or the scientific community to call the uniqueness of human language into question either.


ISBN (Hardcover)
Publication date
2023 (October)
Soziale Netzwerke Sprachwissenschaft Kommunikationstheorie Onlinekommunikation Verteilte Systeme
Berlin, Bern, Bruxelles, New York, Oxford, Warszawa, Wien, 2023. 316 pp., 4 fig. col., 60 fig. b/w, 10 tables.

Biographical notes

Tobias Gretenkort (Author)

Tobias Gretenkort holds a PhD degree from RWTH Aachen University and currently teaches academic writing and research methods at the Faculty of Law and Political Science of the Universitat Oberta de Catalunya. Having published contributions on the use of addressing forms in Spanish and linguistic politeness, his current research activity lies at the intersection of linguistics and criminology.


Title: The Semiotics of Consensus