XENOLINGUISTICS

Xenolinguistics brings together biologists, anthropologists, linguists, and other experts

specializing in language and communication to explore what non-human, non-
Earthbound language might look like. The 18 chapters examine what is known about
human language and animal communication systems to provide reasonable hypotheses
about what we may find if we encounter non-Earth intelligence.
Showcasing an interdisciplinary dialogue between a set of highly established
scholars, this volume:

• Clarifies what is and is not known about human language and animal communication
systems
• Presents speculative arguments as a philosophical exercise to help define the
boundaries of what our current science can tell us about non-speculative areas of
investigation
• Provides readers with a clearer sense of the how our knowledge about language is
better informed through a cross-disciplinary investigation
• Offers a better understanding of future avenues of research on language

This rich interdisciplinary collection will be of interest to researchers and students

studying non-human communication, astrobiology, and language invention.

Douglas A. Vakoch is Director of METI (Messaging Extraterrestrial Intelligence), a

nonprofit research and educational organization dedicated to transmitting intentional
signals to nearby stars. He is the editor of more than a dozen books, including Ecofeminist
Science Fiction: International Perspectives on Gender, Ecology, and Literature (2021)
and The Routledge Handbook of Ecofeminism and Literature (2022).

Jeffrey Punske is Associate Professor and the director of undergraduate studies in

linguistics at Southern Illinois University. He is the editor of Language Invention and
Linguistics Pedagogy (2020).
XENOLINGUISTICS
Towards a Science of Extraterrestrial
Language

Edited by Douglas A. Vakoch and Jeffrey Punske

Designed cover image: © Getty Images | WhataWin
First published 2024
by Routledge
4 Park Square, Milton Park, Abingdon, Oxon OX14 4RN
and by Routledge
605 Third Avenue, New York, NY 10158
Routledge is an imprint of the Taylor & Francis Group, an informa business
© 2024 selection and editorial matter, Douglas A. Vakoch and Jeffrey Punske;
individual chapters, the contributors
The right of Douglas A. Vakoch and Jeffrey Punske to be identified as the authors
of the editorial material, and of the authors for their individual chapters, has been
asserted in accordance with sections 77 and 78 of the Copyright, Designs and
Patents Act 1988.
All rights reserved. No part of this book may be reprinted or reproduced or utilised
in any form or by any electronic, mechanical, or other means, now known or
hereafter invented, including photocopying and recording, or in any information
storage or retrieval system, without permission in writing from the publishers.
Trademark notice: Product or corporate names may be trademarks or registered
trademarks, and are used only for identification and explanation without intent to
infringe.
British Library Cataloguing-in-Publication Data
A catalogue record for this book is available from the British Library
ISBN: 978-1-032-39960-7 (hbk)
ISBN: 978-1-032-39959-1 (pbk)
ISBN: 978-1-003-35217-4 (ebk)
DOI: 10.4324/9781003352174
Typeset in Times New Roman
by Apex CoVantage, LLC
CONTENTS

List of Contributors viii

1 Goals of the Volume 1

Jeffrey Punske

2 Many Ways to Say Things: What the Diversity of Animal

Communication on Earth Can Tell Us About the Likely
Nature of Alien Language 3
Arik Kershenbaum

3 Recognizing Intentional Signals and Their Meanings in

Non-Human Communication 15
Catherine Hobaiter, Adriano R. Lameira, and Derek Ball

4 Getting Out of Our Skin: What Decoding Interspecies

Communication and Nonhuman Intelligence Can Tell Us
About Deciphering Alien Languages 33
Denise L. Herzing

5 Communicative Resources Beyond the Verbal Tier: A View

on Xenolinguistics From Interactional Linguistics 39
Heike Ortner
vi Contents

6 How Studies of Communication Among Nonhumans and

Between Humans and Nonhumans Can Inform SETI 51
Irene M. Pepperberg

7 Patterns of Communication of Human Complex Societies

as a Blueprint for Alien Communication 62
Anamaria Berea

8 Interstellar Competence: Applications of Linguistics

and Communicative and Cultural Competencies to
Extraterrestrial Communication 74
Sumayya K.R. Granger, Judd Ethan Ruggill, and Ken
S. McAllister

9 Why Do We Assume That We Can Decode Alien Languages? 87

Con Slobodchikoff

10 Xenolinguistic Fieldwork 96
Claire Bowern

11 Investigating the Foundations of Meaning in a Xenolanguage 110

Andrew McKenzie

12 A Linguistic Perspective on the Drake Equation: Knowns

and Unknowns for Human Languages and Extraterrestrial
Communication 123
Daniel Ross

13 Cognition, Sensory Input, and Linguistics: A Possible

Language for Blind Aliens 138
Sheri Wells-Jensen

14 The Design Features of Extraterrestrial Language: A

Domain-General Approach 152
Darcy Sperlich

15 Universal Grammar 165

Ian G. Roberts, Jeffrey Watumull, and Noam Chomsky
 Contents vii

16 Where Does Universal Grammar Fit in the Universe?

Human Cognition and the Strong Minimalist Thesis 182
Bridget D. Samuels and Jeffrey Punske

17 Learning and Adaptation of Communication Systems in

Biological Life Forms 194
Jessie S. Nixon and Fabian Tomaschek

18 Writing Systems and METI: Off-the-Shelf Encoding of

Human Physiology, Language, Cognition, and Culture 214
Daniel Harbour

Index 230
CONTRIBUTORS

Derek Ball is a senior lecturer in philosophy at the University of St Andrews. He

holds a PhD in Philosophy from the University of Texas at Austin. His research
focuses on understanding the mind and its place in the natural world, with special
focus on the ways in which our thought and language are shaped by the environment.
During his time in St Andrews, he has served as Coordinator of the St Andrews/
Stirling joint Ph.D. program in philosophy, and as Director of the Arche Research
Centre. He has held visiting fellowships at the Australian National University and
the University of Oslo, as well as research grants from the Arts and Humanities
Research Council. He is the co-editor of The Science of Meaning (Oxford Univer-
sity Press), and his work has appeared in Mind, Philosophy and Phenomenological
Research, Philosophers’ Imprint, Inquiry, and elsewhere.

Anamaria Berea is an associate professor in computer science at George Mason

University. She holds Ph.D. degrees in economics (2010) and computational social
science (2012), and her current research is focused on the emergence of commu-
nication in biological and social networks by applying theories and methods from
economics, complex systems, and data science to reinterpret historical, anthropo-
logical, biological, and artistic data regarding the fundamental aspects of commu-
nication on our planet, from signaling in simple biological organisms to complex
human and computer languages.

Claire Bowern is a professor at Yale University. She is one of the leading schol-
ars on language documentation, historical linguistics, and indigenous languages of
Australia. She has authored numerous articles and chapters related to these topics.
She is the author or co-author of two textbooks: An Introduction to Historical Lin-
guistics and Linguistic Fieldwork: A Practical Guide. She is also the author of A
 Contributors ix

Grammar of Bardi and Sivisa Titan: Grammar, Texts, Vocabulary. She earned her
Ph.D. in linguistics from Harvard in 2004.

Noam Chomsky, often considered the founder of modern linguistics, is one of the
most cited scholars in modern history. Among his groundbreaking books are Syntactic
Structures, Language and Mind, Aspects of the Theory of Syntax, and The Minimalist
Program, each of which has made distinct contributions to the development of the
field. He has received numerous awards, including the Kyoto Prize in Basic Sciences,
the Helmholtz Medal, and the Ben Franklin Medal in Computer and Cognitive Sci-
ence. Chomsky is also one of the most influential public intellectuals in the world.
He has written more than 100 books. Chomsky joined the University of Arizona in
fall 2017, coming from the Massachusetts Institute of Technology (MIT), where he
worked since 1955 as professor of linguistics, then as professor of linguistics, emeritus.

Sumayya K.R. Granger is an assistant professor in the Department of Public and

Applied Humanities at the University of Arizona. She is also the Associate Director
of Program Administration at the Center for English as a Second Language. She
earned her M.A. and Ph.D. in linguistics, writing her dissertation on the syntactic
theory of modals, and she has written on other topics, such as plural formation in
Nahuatl and Aktionsart in Navajo. She has taught classes in pedagogical gram-
mar and intercultural competence, and she has presented several times on different
aspects of language teaching/learning and intercultural competence.

Daniel Harbour is Professor of the Cognitive Science of Language at Queen Mary

University of London, where he currently holds a Leverhulme Research Fellow-
ship exploring the impact of grammatical change on writing system evolution. He
received his Ph.D. in linguistics from the MIT in 2003 for work on the complex
number system of a Native American language. Since then, he has conducted in-
depth fieldwork on languages on four continents. He uses broad typological varia-
tion, coupled with language-specific investigation, to build mathematically precise
models of the conceptual building blocks of human language, particularly in rela-
tion to systems of counting and deixis. His publications include An Intelligent Per-
son’s Guide to Atheism, Morphosemantic Number, and Impossible Persons.

Denise L. Herzing, Founder and Research Director of the Wild Dolphin Project,
has completed more than 33 years of her long-term study of the Atlantic spotted
dolphins inhabiting Bahamian waters. She is an affiliate assistant professor in bio-
logical sciences at Florida Atlantic University. Dr. Herzing is a 2008 Guggenheim
Fellow, a fellow with the Explorers Club, a scientific adviser for the Lifeboat Foun-
dation and the American Cetacean Society, and on the board of Schoolyard Films.
In addition to many scientific articles, she is the co-editor of Dolphin Communi-
cation and Cognition, and author of Dolphin Diaries: My 25 Years with Spotted
Dolphins in the Bahamas and The Wild Dolphin Project (2002).
x Contributors

Catherine Hobaiter is a field primatologist who has spent the past 19 years stud-
ying wild primates across Africa. She received her Ph.D. from the University of
St Andrews and is now a Reader in the School of Psychology and Neuroscience.
She spends around half the year in the field and leads a team of researchers explor-
ing great ape behavior. The main focus of her research is communication and
cognition in wild apes. Through long-term field studies, she explores what the
behavior of modern apes living in their natural environment tells us about their
minds and also about the evolutionary origins of our own behavior. Her work
revealed that apes use large repertoires of gestures with specific meanings in a
language-like way.

Arik Kershenbaum is a fellow of Girton College, University of Cambridge, and

a college lecturer. He researches animal communication across diverse species,
with a particular interest in the communication between cooperative predators.
Using field experiments and mathematical simulations, he tries to understand
what this communication tells us about the fundamental principles driving the
evolution of language. Dr. Kershenbaum’s research focuses mostly on wolves,
which he studies in Montana, Spain, and Italy, and on free-ranging human-habit-
uated dolphins in the Red Sea. He held the Herchel Smith Research Fellowship in
Zoology at the University of Cambridge, and prior to this, he was a postdoctoral
research fellow at the National Institute for Mathematical and Biological Synthe-
sis (NIMBioS) in Knoxville, Tennessee. Dr Kershenbaum received his Ph.D. from
the Haifa University in Israel, where he studied the communication and behavior
of the rock hyrax.

Adriano R. Lameira started primate vocal research in the early 2000s in Borneo
with wild orangutans, and quickly expanded his research in the years that fol-
lowed to Sumatran orangutans. Today, he holds the largest collection of orangutan
calls ever assembled, spanning multiple wild and captive populations and several
tens of thousands of observation hours. Since the beginning of his work, orangu-
tans have exhibited a level of vocal diversity, flexibility. And learnability that has
surpassed traditional expectations. He is an assistant professor at the University
of Warwick.

Ken S. McAllister is Associate Dean for Research and Program Innovation for the
College of Humanities at the University of Arizona, where he is also Professor of
Public and Applied Humanities. He holds affiliate appointments in the departments
of English and Teaching, Learning, and Sociocultural Studies, as well as in the
School of Information. He co-founded and co-directs (with Judd Ethan Ruggill) the
Learning Games Initiative and its attendant research archive. His research focuses
on technologically enhanced modes of persuasion, particularly in transdisciplinary
contexts.
 Contributors xi

Andrew McKenzie is an associate professor at the University of Kansas. He is

the co-author of Plains Life in Kiowa: Voices from a Tribe in Transition along with
numerous articles and chapters on natural language semantics and the structure
and documentation of Kiowa. His work has been supported by two National Sci-
ence Foundation (NSF) grants. He has also co-authored work regarding language
change during long-distance space travel for the European Space Agency’s journal,
Acta Futura. Prior to his service at the University of Kansas, he was employed at
University of Texas at Arlington. He earned his Ph.D. in linguistics in 2012 from
the University of Massachusetts Amherst.

Jessie S. Nixon completed her Ph.D. in psycholinguistics at the University of Lei-

den in 2014. Dr. Nixon uses neuroscientific and behavioral methods, combined
with advanced statistical modeling, to investigate speech processing and acquisi-
tion in a variety of languages. Dr. Nixon’s research has demonstrated that listeners’
level of perceptual uncertainty during speech perception is highly sensitive to the
shape of statistical cue distributions in both native and non-native speech and in
both segmental and prosodic cues. Dr. Nixon is currently based in the Quantita-
tive Linguistics Group, University of Tübingen, where she is working on a project
developing a computational model of speech perception and production based on
learning theory and discriminative linguistics.

Heike Ortner is an associate professor at the Department of German Studies, Univer-

sity of Innsbruck, Austria. She received her master degree at the Karl-Franzens-Uni-
versity Graz in German Philology and Applied Linguistics. In 2012, she completed
her Ph.D. studies at the University of Innsbruck with a “Promotio sub auspiciis praes-
identis rei publicae,” the highest honor for academic achievements in Austria. From
2011–2014, she held a postdoctorate position in the division “German Linguistics”
at the Department of German Studies. She became assistant professor at the same
department in 2014. In 2016, she was a short-term visiting professor at the Wirth
Institute at the University of Alberta, Canada. Her research interests include language
and emotion, interactional and multimodal linguistics, linguistic aspects of public
health communication, computer-mediated communication, and discourse analysis.

Irene M. Pepperberg is a research associate at Harvard, and has been a North-

western University visiting assistant professor, a University of Arizona tenured
associate professor, an MIT Media Lab visiting associate professor, a Brandeis
University adjunct associate professor and an MIT senior lecturer. She has received
John Simon Guggenheim, Whitehall, Harry Frank Guggenheim, Selby, and Rad-
cliffe fellowships, was an alternate for the Cattell Award for Psychology, won the
2005 Frank Beach Award, and was nominated for Weizmann, L’oréal, and Grawe-
meyer awards and the Animal Behavior Society’s 2001 Quest and 2015 Exemplar
awards. She received St. Johns University’s 2013 Clavius Award. She authored the
xii Contributors

book, The Alex Studies, on grey parrot cognition and interspecies communication.
Her memoir, Alex & Me, a New York Times bestseller, won a Christopher Award.
She has published more than 100 scholarly articles in peer-reviewed journals and
as book chapters.

Jeffrey Punske received his Ph.D. in linguistics from the University of Arizona in
2012. He is an associate professor in the School of Languages and Linguistics at
Southern Illinois University, With previous employment at Kutztown University
in Pennsylvania and the University of Oklahoma. His primary research is one the
interfaces of morphosyntax with other grammatical, and cognitive components.
He is the co-editor of the book Language Invention in Linguistics Pedagogy. He
has also co-authored work regarding language change during long-distance space
travel for the European Space Agency’s journal, Acta Futura.

Ian G. Roberts is a professor of linguistics at the University of Cambridge. His

research is in theoretical linguistics, more specifically in comparative syntax. His
work is set against the background assumptions advocated by Noam Chomsky:
that there exists a specific human cognitive capacity for language that is present
at birth and requires simple environmental stimulation for linguistic competence
in the mother tongue to develop during the early years of life. He currently holds
a European Research Council Advanced Grant funding a project whose goal is to
investigate a specific hypothesis as to the way in which the grammatical options
made available by universal grammar are organized. Refining and testing this
hypothesis involves looking at languages from all over the world and assessing the
extent to which certain patterns recur. Dr. Roberts received his B.A. from Bangor
and his Ph.D. from University of Southern California.

Daniel Ross is a lecturer of linguistics at the University of California, Riverside.

He has taught courses on syntax, semantics, morphology, and historical linguis-
tics. His dissertation (University of Illinois Urbana-Champaign, 2021) focuses on
pseudocoordination, serial verb constructions and other multi-verb predicates as
instances of form-structure mismatches in syntax; from a comparative perspec-
tive, these constructions are strikingly similar in function and syntactic properties
despite variation in form, and from a theoretical perspective, the data from English
and other languages proves difficult to explain in conventional syntactic theory.
The research in the dissertation is supported by a 325-language comparison of an
array of morphosyntactic features.

Judd Ethan Ruggill is Head of the Department of Public and Applied Humanities
at the University of Arizona, and an affiliate faculty member in the Africana Stud-
ies Program; the Department of English; the School of Information; the School
of Theatre, Film, and Television; and the Graduate Interdisciplinary Program in
Social, Cultural, and Critical Theory. In addition, he co-directs the Learning Games
 Contributors xiii

Initiative, a transdisciplinary, inter-institutional research group he co-founded in

1999 to study, teach with, and build computer games. He primarily researches play
and the technologies, industries, and sociocultural phenomena that enable it. He
has published and presented on topics ranging from game design for second lan-
guage acquisition and teaching to the wicked problem of collaboration.

Bridget D. Samuels received her Ph.D. in linguistics from Harvard University and
is a member of the Center for Craniofacial Molecular Biology at the University of
Southern California. Previously, she has held teaching positions at the University of
Maryland and Pomona College. Her research focuses on the evolution of language
and cognition in the history of humankind, phonological theory, and the interface
between phonology and morphosyntax. She is the author of Phonological Architec-
ture: A Biolinguistic Perspective (2011, Oxford University Press), as well as numer-
ous other publications in evolutionary, biological, and theoretical linguistics.

Con Slobodchikoff is a professor emeritus of biology at Northern Arizona Univer-

sity, Flagstaff. He received B.S. and a Ph.D. degrees from the University of Califor-
nia, Berkeley, and has been a Fulbright Fellow and a visiting professor of Zoology
at Kenyatta University in Kenya. Slobodchikoff has authored two books on animal
behavior (Harvard University Press and St. Martin’s Press); has edited three books
on subjects of species concepts, ecology, and social behavior; and has published
about 100 papers and book chapters in the scientific literature. His work with prai-
rie dogs has drawn considerable interest, and he was the subject of a BBC one-hour
documentary (Talk of the Town) and an Animal Planet documentary (Prairie Dogs).
He has appeared in a variety of media programs such as NBC Dateline, ABC World
News, the BBC, the CBC, Australian Broadcasting, German-Belgian television,
CNN, and Turner Broadcasting.

Darcy Sperlich is an associate professor of Linguistics at Xi’an Jiaotong – Liv-

erpool University. He works both in theoretical and experimental linguistics with
a cross-linguistic perspective, focusing on syntax and pragmatics, and their inter-
faces. He has published a number of articles, and published a book entitled Reflex-
ive Pronouns: A Theoretical and Experimental Synthesis.

Fabian Tomaschek obtained his Ph.D. in 2013 with a study on the neural mecha-
nisms of vowel perception. In 2012, he became a postdoctoral research fellow at
the Department of General Linguistics, University of Tübingen. Ever since, he has
focused in his work on how humans learn to encode and decode messages using
the acoustic signal. He investigates the articulation of vowels and uses advanced
statistical modeling to predict their acoustic and articulatory shape. His work has
been published among others in the Journal of Phonetics, PloS ONE, Brain and
Language, and Linguistics Vanguard. On his website, he also published an intro-
ductory textbook on programming data analyses in linguistic corpora with R.
xiv Contributors

Jeffrey Watumull is the artificial intelligence (AI) lead at Oceanit. He received

B.S. degrees, summa cum laude, in mathematics and evolutionary biology from
Harvard University in 2009, an M.Phil. in linguistics from the University of Cam-
bridge in 2010, and a Ph.D. in artificial intelligence from MIT in 2015. His research
has centered on the mathematical optimality of human cognition, specifically in the
domain of language. He has published numerous papers in peer-reviewed journals
and his work has been covered frequently in the media. At Oceanit, Dr. Watumull
has embarked on research to develop strong AI: computational systems equipped
with human-level intelligence in domains both specific (e.g., language) and gen-
eral (e.g., problem solving). He is currently principal investigator on a Defense
Advanced Research Projects Agency (DARPA) seedling, an Office of Naval
Research Broad Agency Announcement (ONR BAA), and a number of Small Busi-
ness Innovation Research grants (SBIRs).

Sheri Wells-Jensen is a professor in the Department of English and coordina-

tor of the minor in linguistics at Bowling Green State University in Ohio. Along
with various aspects of issues pertaining to SETI (the search for extraterrestrial
intelligence), her research interests include phonetics, braille, language preserva-
tion, TESOL (teaching English to speakers of other languages), language creation,
and disability studies. She serves on the board of directors of Messaging Extrater-
restrial Intelligence (METI). Dr. Wells-Jensen also coordinates BGSU’s graduate
certificate in TESOL and teaches courses in general linguistics, applied phonology,
and applied syntax.
1
GOALS OF THE VOLUME
Jeffrey Punske

To date, there is no serious evidence of contact between humanity and non-terres-

trial intelligence. In fact, as of writing, there is no confirmed evidence available
to us of any form of life that did not originate on Earth. Given the vastness of the
universe and the processes involved in both the development of life and the devel-
opment of intelligence, it is generally assumed that this is due to limitations of our
ability to find the evidence – not necessarily the lack entirely. It is unlikely that
contact with any non-terrestrial intelligence will occur within the lifetimes of any
of the contributors to this volume.
This raises the question of the purpose of a volume like this. This volume seeks
to explore the potential nature of a non-terrestrial intelligence with linguistic capa-
bilities. We cannot observe this nature, which moves us into a different realm of
inquiry than the standard scientific approach. Thus, rather than being a speculative
volume about the potential nature of non-Earth intelligence/linguistic systems, this
volume largely explores what we know about communication systems, languages,
and other cognitive systems on Earth through the lens of what we might observe
beyond it. In doing so, we hope to help define the limits of what we presently know
about such systems and provide the foundation for future explorations into broader
questions about the nature of language and communication.
This volume draws together a range of experts from various fields and divergent
perspectives. Our goal is not to provide any definitive answers. Rather, we seek to
show the division points across these fields and perspectives. The volume may be
roughly divided into three major themes: animal communication systems, human
language and linguistics, and communication systems more generally. As each of
these topics closely intersect, there is not a neat division from one topic to the next –
rather, each contribution connects with others to try to define the question more
broadly.

DOI: 10.4324/9781003352174-1
2 Jeffrey Punske

It is our hope that in trying to define what we know and where the limitations
of our knowledge lie, we can work towards an integrated science of language and
communication that takes seriously the fundamental and universal natures of both.
By utilizing our fascination of with stars and the unknown, we can perhaps enhance
our own knowledge of ourselves – and we may continue to explore the nature of
the unknown with imagination but also with a core knowledge of what might be
out there.
2
MANY WAYS TO SAY THINGS
What the Diversity of Animal Communication on
Earth Can Tell Us About the Likely Nature of Alien
Language

Arik Kershenbaum

One of the most perplexing questions in modern evolutionary biology is this: Why
has only one species on Earth evolved a true, rich language? In almost four billion
years of evolution, humans stand out among all the species that have ever evolved
as the only one able to communicate a limitless number of concepts between con-
specifics. From an anthropocentric perspective, it seems obvious that language
brings tremendous adaptive advantages, particularly in terms of the cooperation
it enables, and humanity’s huge technological and intellectual advances testify to
the power of such cooperation. But the question may not be a straightforward as it
seems, because we need to examine the claim that humans are the only species with
language. Indeed, our definition of language may be biased in such a way that only
humans fit the definition – maintaining the qualitative difference between humans
and other animals that has been a particular quest of philosophers and natural sci-
entists for generations.
Animals on Earth have exploited a staggering range of strategies to increase
their fitness by communicating both with con- and hetero-specifics, and exam-
ples of these evolutionary strategies can aid us in understanding what extra-
terrestrial communication may be like (Herzing, this volume [Chapter 4];
Pepperberg, this volume [Chapter 6]; Slobodchikoff, this volume [Chapter 9]).
Almost every physical modality has been explored, from the familiar acoustic
and visual channels to the seemingly unlikely electromagnetic – and within each
modality, different encoding strategies have been used. Indeed, it is hard to find
a potential form of physical communication that has not been explored some-
where in evolutionary history on Earth. It therefore seems quite likely that life
on other planets will make use of some of the underlying communicative tech-
niques exploited on this planet, albeit with particular adaptations suited to the
specific conditions of each habitat. For this reason, it is instructive to examine

DOI: 10.4324/9781003352174-2
4 Arik Kershenbaum

the breadth of communicative strategies employed by animals on Earth, and to

ask what is it about each strategy that would, or would not, favour it as a carrier
for linguistic behaviour.
Without doubt human language arose through natural selection. Whether lan-
guage evolved from a precursor in the form of a simple system of communica-
tion in proto-humans, or arose independently following rapid social and cognitive
development of our ancestors (Jackendoff 2011), evolutionary forces certainly
shaped the uptake and development of the linguistic trait, as with any phenotypic
trait. Therefore, certain selective pressures favoured the evolution of language in
one lineage of hominins but not others, or in any other taxa across the animal
kingdom. Nonetheless, most – if not all – animals communicate, and sometimes
with phenomenally complex and precise signals that appear able to carry linguistic
content – although they are not used in that way. What is it that caused one evo-
lutionary trajectory to accelerate in the direction of language while other potential
trajectories were discarded?
Sensory systems are energetically expensive (Niven and Laughlin 2008), and
every sensory system we know of makes use of a considerable part of an animal’s
energy budget. For example, in the blowfly Calliphora vicina, energy consumption
by the retina alone is as much as 8% of the resting metabolic rate (Howard et al.
1987). On Earth, such sensory systems, and the processing mechanisms attached
to them, are energetically expensive largely due to the pumping of ions in and out
of cells in neural tissue – a parochial mechanism that may not be used by crea-
tures that have evolved on other worlds. However, whatever its implementation,
communication is fundamentally energetic because it acts to reduce information
entropy in a system. Therefore, as long as energy is a limiting resource in a particu-
lar ecosystem, we expect that communication systems will evolve on other planets
only to address important selective pressures that justify investing a considerable
proportion of an animal’s energy budget into producing, detecting, and processing
communication signals. In a hypothetical habitat where energy is not, in fact, a
limiting resource, the consequential reduction in competition and selective pres-
sure may mean that communication – and many other things – would be unlikely
to evolve at all!
Because of the energetic cost of sensory systems, most such systems should
be broadly tuned (Stevens 2013: 62), i.e., should respond in a general fashion to a
wide range of signals rather than be dedicated to detecting signals of a particular
type. However, communication requires a degree of sensor specificity because dis-
crimination between signals with different meanings is a fundamental requirement
of communication. Therefore, sensory systems should evolve in the face of strong
selective pressure, and communicative systems should evolve only when selective
pressures are even stronger, such that precious sensory resources can be dedicated
to transmitting and receiving specific signals. This will also allow broadly tuned
sensory systems (such as hearing) to respond to narrow signals with specific mean-
ing, such as alarm calls.
 Many Ways to Say Things 5

Communicative Modalities

Sound and Vision

We are most familiar with the particular communication modality that has come to
dominate animal communication on this planet and which forms the basis of spo-
ken human language. Acoustic communication is not the most ancient form of com-
munication but it is undoubtedly the most studied. Despite the dangerous tendency
to attach more significance than it deserves to our own preference for a particular
modality, it is nonetheless the case that acoustic communication appears to be so
widespread because it provides a number of objective and significant advantages.
Acoustic communication is fast: sound travels at 340 m/s in air, and 1500 m/s in
water, far faster than the movement of any animal on this planet. Speed also means
that acoustic signals can be temporally discriminated, which greatly increases the
amount of information they can carry compared to slowly varying modalities such
as olfactory cues. Critically, acoustic signals also diffract around objects commonly
found in both the terrestrial and marine environments, which means that the signal
can be detected even if the receiver is not in the line of sight of the sender – a key
disadvantage of the visual modality.
Visual signals can contain large amounts of information, particularly by manip-
ulating the geometrical configuration of objects. This is most obviously demon-
strated by the shapes of the letters on this page, which provide almost unlimited
flexibility for information encoding. Visual signalling is widespread across the ani-
mal kingdom, quite likely because the production of complex geometrical configu-
rations does not require specialist production mechanisms beyond those normally
possessed by animals. The bristling of hair on an angry cat, or the thumbs-down
signal of a Roman emperor, make use of anatomical features of the animals that are
already evolved for manipulating the environment. Therefore, it seems likely that
in any non-opaque alien environment, some form of visual signalling will exist.
However, although visual signalling almost certainly preceded acoustic commu-
nication in ancient hominins, it is possible that the very fact that visual signal-
ling relies on geometric configuration may limit its utility in many environments.
Although sound and light both dissipate similarly with distance, decoding visual
information also depends on angular resolution, so visual signals are more difficult
to distinguish at long ranges. Finally, acoustic signals can be generated with vary-
ing frequency, which adds another layer of complexity to the channel and allows
even more information to be encoded in signals. Although light also can be com-
bined in different frequencies, the extremely short wavelength means that animals
on this planet do not directly measure light frequency as we do sound frequency;
rather, colours are distinguished by arbitrary filters. Human language and bird
song make use of both amplitude and frequency modulation as well as formant
manipulation (Fitch 2000), so that subtle differences in the acoustic structure of a
signal can be detected, perceived, and interpreted by the receiver as having distinct
6 Arik Kershenbaum

meanings. However, organisms of our physical scale are unlikely to be able to per-
form a similar analysis of visual signals; an observation made by J.B.S. Haldane,
who pointed out that, “It is a curious fact that we men can place musical notes in
their natural order by intuition, while it required the genius of Newton to do the
same for colour” (Haldane 1927: 282).
The disadvantages of both acoustic communication (dependence on the density
of a planet’s atmosphere, difficulty in localising the sound source, etc.) and visual
communication (it can be obscured by solid objects or scattered by atmospheric
particles) has led many animals to rely on multimodal communication, whereby dif-
ferent modalities are used either to provide a more robust communication channel,
or to add extra and richer information. Indeed, both Herzing (this volume [Chapter
4]), and Ortner (this volume [Chapter 5]) argue that for animals on Earth and for
aliens on another planet, respectively, multimodal channels are fundamental to the
development of complex communication and the development of language.
Acoustic communication is very powerful, and it may be that animals that
evolved on other planets and live in a dense gaseous or rarefied liquid environment
like Earth’s may have evolved a language based on an acoustic channel. Nixon
and Tomaschek (this volume [Chapter 17]) explore the possible effects of differ-
ent planetary environments on the suitability or otherwise of acoustic and visual
communication – both the opacity of the atmosphere and the fluid density can have
dramatic effects on whether particular signals propagate well or not. However,
what other possibilities in other physical modalities exist for conveying sufficient
information of sufficient complexity to constitute a language? Earth’s animals have
exploited a number of other communications modalities, and the extent to which
these do – or do not – support complex communication helps us consider the pos-
sible nature of language on other planets. Here I consider three candidates for lin-
guistic communication that appear less likely on Earth but whose characteristics
may provide diverse opportunities on other planets: olfactory or chemical signals,
electrical signals, and magnetism.

Chemical Sensing and Olfaction

Although chemical sensing is likely the most primitive communications channel,
olfaction (which, for simplicity, I loosely define as chemical sensing at a distance)
seems a poor candidate for linguistic communication. In typical environments on
Earth (aqueous, terrestrial), chemicals diffuse slowly. As a result, although phero-
mones can be used to convey urgent signals, the complexity of such signals must
be very low – barely more than indicating presence or absence. As Ross (this vol-
ume [Chapter 12]) points out, the linguistic potential in chemical signalling is very
limited, as it fails to provide what Ross considers a crucial design feature, “rapid
fading” (Hockett 1960). Although an alarm signal with low temporal specificity is
easily conveyed (“a predator is near”), it is much harder to imagine olfactory cues
giving information on fast-changing situations (“he’s coming up on your left!”).
 Many Ways to Say Things 7

Olfactory signals are also difficult to localise if they diffuse throughout a medium
(although they can be deposited at specific points), particularly in any environment
with turbulent fluid flow that tends to homogenise the signal within its transport
medium (air or water).
Despite this, several animal species have made widespread and rather complex
use of chemical cues that never evolved to the complexity of a language on Earth
but may give us clues about the kind of world where olfactory language may be
possible. Stereo-olfaction, for instance, allows animals such as dogs (Craven et al.
2010), rats (Rajan et al. 2006), and sharks (Gardiner and Atema 2010) to localise
chemical cues in space by measuring concentration differences between multiple
detectors (in the case of Bilateria, usually two nostrils or antennae). Even in a
complex and turbulent marine environment, animals such as lobsters can perceive
remarkably complex “olfactory landscapes”, including deriving information on
temporal and spatial variation in chemical cues (Atema 1995). As our understand-
ing of animal olfactory perception increases, it becomes clear that the integration of
multiple detectors can generate complex signals based on suites of chemical cues
(Su et al. 2009). Nonetheless, the communicative potential of olfaction appears
slight. What kind of alien environments might favour the evolution of more com-
plex and possibly linguistic smells? Clearly, one requirement is a medium through
which chemicals can be transported – probably by diffusion, but mass transfer cur-
rents could provide an alternative. Laminar flow of the medium would be helpful,
as turbulence will tend to reduce information content and mask important cues such
as location and time of signalling. Second, the slow nature of diffusion (compared
to sound and light) may mean that olfactory language is realistic only in small
niches, whereby communication over long distances is not required. Unfortunately,
a world in which animals compose sonnets through the carefully controlled release
of strong-smelling gases still seems very unlikely – and alien.
Perhaps most significantly, the slow speed of diffusion – and the unidirectional
nature of the passive transport of olfactory cues by mass flow – limit the potential
for bidirectional chemical “conversations”. Hobaiter et al. (this volume [Chapter 3])
provide a convincing argument that cooperative, two-way signalling is essential for
the evolution of complex communication that could eventually become something
we would label “language”. Higher levels of communicative complexity are often
linked to time-critical decisions that animals need to make, and it is difficult to see
how the Gricean paradigm that Hobaiter et al. discuss would evolve from olfactory
signals that are in essence slow and unidirectional.

Electrical Signals
All life on Earth makes use of electrically charged ions to store and mobilise
energy. When ions move, they create electric fields, so electrical signals are inher-
ent to life on Earth. It is not clear that mobilisation of energy in the form of the
movement of ions is necessarily a universal feature of life – and on other planets,
8 Arik Kershenbaum

storing and using energy could be accomplished in other force fields – but all life
on Earth makes use of electrical potential to store and release energy, and animals
in particular are a source of copious electrical fields, particularly from muscles and
nerve cells. Given that all animals produce electrical signals, it is almost surpris-
ing that the use of passive electroreception, i.e., the detection of the fields created
by other animals, is not more widespread both among predators detecting prey
and prey avoiding predators. In fact, many animals do have weak electrosensing,
although our understanding of the distribution of this ability in the animal kingdom
is still expanding. The ability both to sense and generate electric fields as an active
signal is, however, most developed among two groups of teleost fish, the Mormyri-
formes, a group of African electric fish, and the Gymnotiformes, electric fish from
Central and South America (Hopkins 2009). Unlike olfaction, the electric sense has
the complexity and fidelity needed to convey large amounts of information reliably
and quickly, and could serve as a possible model for linguistic communication.
Nonetheless, no electric fish possess language. What features of electrosensing
make it amenable to complex communication and limit is usefulness?
Although electric fields are pervasive among animal tissues, the nature of these
indirectly generated fields is not particularly conducive to being co-opted for com-
municative uses. Detecting your prey (or your predator) by passive sensing of
the electric field of their muscles requires detection of a low-frequency or direct
current (DC) electromagnetic field, and the information content of such a field
is necessarily low. Electric fish also have electric organs that actively generate
an electrostatic field around them, and receptors on the fish’s body detect varia-
tions in that field caused by the presence of nearby animate and inanimate objects
(Meyer 1982). Under the selective pressure of predatory animals searching for
low-frequency electrical fields of prey animals, electric fish evolved active search
using electrical fields at a higher frequency, which could not be as easily detected
by predators (Stoddard 1999). In contrast to slowly varying or DC electric fields,
high-frequency communication by electric fish holds the potential to encode large
amounts of information and transmit it quickly and reliably. Both Mormyriformes
and Gymnotiformes have diverse and complex electrical communication signals
that have been shown to encode information such as species identity (Sullivan et al.
2000), individual identity (Scheffel and Kramer 2006), sex (Lorenzo et al. 2006),
and social status (Hagedorn and Zelick 1989).
There are, however, drawbacks to the use of an electrical channel for complex
communication. Signal range in most media is limited, especially because the sig-
nal strength of a dipole falls as the cube of the distance, rather than the square
as with most dissipative signals. Electrocommunication is therefore a short-range
modality on Earth, with typical distances generally being less than 1 metre (Squire
and Moller 1982). However, just as an oscillating electrical field such as those pro-
ducing radio signals can propagate long distances through an insulating medium
(e.g., air) from a dipole antenna, it is possible to theorise that certain environmental
conditions could favour the evolution of high-frequency electrical communication
 Many Ways to Say Things 9

that is not as range-restrictive as the aquatic solutions found on Earth. Other con-
siderations, however, cause us to be sceptical of such a possibility. Decoding such
signals would require a mechanism for frequency analysis, analogous to the basilar
membrane in the ear, in which multiple detectors are tuned to different acoustic fre-
quencies. Such an array of electrical receptors, each tuned to a different frequency,
is not impossible, and electric fish do have a limited range of sensory cells tuned
to different frequencies (Kawasaki 2005). However, the number and specificity of
frequency detectors needed greatly exceeds those observed, and the complexity
of such a frequency analyser would imply very strong selective pressure. Even
the evolution of frequency-dependent acoustic communication on Earth appears to
have been in response to strong selective pressures, because even the most diverse
animal class, insects, have very little ability to discriminate acoustic frequency
(Stebbins 1983; Michelsen 1973).

Magnetic Communication
Unlike communication using electric fields, no animals on Earth – as far as we
know – communicate using magnetism. This may seem surprising, given the
closely related physical properties of magnetic and electric fields, and the fact
that many animals – including possibly humans – can detect weak variations in
magnetic field strength. However, it seems that no species has adopted this abil-
ity for communications purposes, probably because magnetic signals are difficult
to generate and manipulate. Magnetic sensing in organisms on Earth is achieved
by measuring the force exerted by external magnetic fields on small pieces of
magnetised materials located in special detector cells in an animal’s body, or by
measuring changes in the magnetic properties of photoreceptor chemicals in the
retina, as they absorb photons (Gould 2008; Wiltschko and Wiltschko 2010). These
subtle detector mechanisms also hint at another possible reason for not commu-
nicating via magnetic signals: the strength of any such signal is likely to be very
small compared to Earth’s magnetic field, and thus difficult to detect and, more
importantly, to quantify. But even if small variations in magnetic field strength
are unlikely to be detectable, it is still possible that communication could occur
via reversals in magnetic polarity – repeated changes in the north-south alignment
of a magnetic field generating organ in a hypothetical magnetic signalling animal.
Thus, we can speculate that in the presence of a weak planetary magnetic field,
animals could communicate using a digital encoding of information as a series of
north-south inversions (analogous to the 1–0 binary representation), perhaps by
moving appendages containing permanent magnets of different orientations. How-
ever, despite the fact that we may wonder, “why didn’t evolution invent . . . ?”,
such a mechanism remains unlikely. On a planet with a weak magnetic field, the
prerequisite magnetosensing ability is unlikely to arise in the first place, because it
may not provide a primary adaptive advantage. Similarly, digital communication is
extremely rare in the natural world, and it too is likely to arise only as an adaptation
10 Arik Kershenbaum

on top of pre-existing analogue signalling, which is simpler to generate and inter-

pret. Therefore, given that there is no plausible mechanism for analogue magnetic
signalling, our hypothetical magnetic communicator seems increasingly unlikely
in any planetary environment.

Information Encoding
It seems clear that any effective communication requires that both parties agree
on the nature of the communicative units that make up an interaction. Indeed, Ort-
ner (this volume [Chapter 5]) argues that a two-way protocol is fundamental to
the development of complex communication on any planet. Humans, with a well-
developed language, are prone to look for similar information-encoding strategies
in every species we examine. We all understand the semantic relationship between
abstract symbols – words, or more strictly, morphemes – and their intended mean-
ing. Indeed, this semantic relationship is often considered a fundamental feature of
language, known as duality of patterning (Hockett 1960), and is often used as the
baseline for analysing the nature of language not just as we understand it, but also
how it must be (Fitch 2005; Hauser et al. 2002; Samuels and Punske, this volume
[Chapter 16]). However, I believe that evidence from examining the communica-
tion of other species requires us to re-examine this most basic assumption, espe-
cially when considering the possibility of alien languages.
In fact, many species do not appear to communicate using discrete encoding of
information into distinct and distinguishable packages. Many species, including
jumping spiders (Elias et al. 2005) and kangaroo rats (Randall and Lewis 1997),
communicate using a series of seismic pulses, whereby the inter-pulse interval, as
well as the frequency characteristics of the pulses themselves, contain information
on individual identity and social status (Randall 1989). A unidimensional trait such
as inter-pulse interval seems to be insufficient for linguistic complexity, but other
species use more complex graded signals. Bottlenose dolphins (Tursiops trunca-
tus), for example, use highly frequency modulated narrowband acoustic signals
(whistles) as their main communication channel, and experimental evidence shows
that the precise modulation pattern contains information such as individual identity
(Sayigh et al. 2017; Sayigh et al. 1999). However, even whistles that ostensibly
carry identical information show considerable variance in their modulation pat-
tern, and the total repertoire of an animal could be anywhere between very small
(if similar whistles are considered identical) to almost infinite (if minor differences
between whistles are considered significant). Clearly, the perception of the receiv-
ing animal (and the cognitive implications of received signals are the true arbi-
ters of information content) remain at least for now hidden to us. Nonetheless,
it appears increasingly likely that these whistles do not fit the morpheme para-
digm of traditional linguistics. Similarly, wolves (Canis lupus) communicate using
a long-range signal (howling) that shares many acoustic properties with dolphin
whistles. Although wolf howling is believed to contain many levels of information
 Many Ways to Say Things 11

(Kershenbaum et al. 2016), no two wolf howls are the same, and it is not clear
whether subtle variations in howl modulation indicate subtle variations in meaning
(Déaux and Clarke 2013).
Notwithstanding our lack of knowledge of the actual information content of
dolphin and wolf communication, the use of non-discrete, continuously variable
signals in animals on Earth raises important questions when we consider alien lan-
guage. Is it possible that an entire language could be constructed on the basis of a
signal that does not consist of morphemes, each attached to a particular meaning?
The idea that dolphins may be the most sophisticated animal communicators after
humans (Gregg 2013) leads us to consider this possibility very seriously. At first, it
seems that the primary constraint acting on a non-discrete communication channel
is that of the resolution between semantically distinct concepts. Two similar whis-
tles may have different meanings, but as their acoustic patterns may be arbitrar-
ily similar, a receiving animal may not acquire the signal-to-noise ratio needed to
distinguish between different meanings. However, we must seriously consider the
possibility that the cognition of the animals may operate in a non-discrete way, and
rather than considering the use of continuous signals to convey discrete concepts,
continuous signals may convey continuous concepts by which subtle variation in
whistle modulation implies only subtle (but relevant) variation in semantic mean-
ing. Experimental studies have shown that dolphins clearly possess the ability to
make discrete semantic judgements (Herman et al. 1994; Harley 2008), and a more
detailed understanding of the abilities of these and other animals to perceive non-
discrete information is essential to understanding the evolution of communication
systems quite different from our own.

Conclusion
Of the vast range of communicative strategies that we observe on Earth, only
one very specific form of information encoding led to the evolution of lan-
guage; namely discrete sematic symbols (words) transmitted through an acoustic
medium. Animals communicate using many different modalities (visual, olfac-
tory, vibratory), and many different encodings (discrete, continuous, pulsatile).
While all of these other approaches are effective at communication, in that they
evolved to provide an adaptive advantage, why did none of them result in lan-
guage? Even the most complex communication seen in the animal world – the
acoustic communication of dolphins and other cetaceans – does not appear to
possess the properties of a true language. I propose that one possible explanation
for this seemingly unlikely situation is the shared nature of their communica-
tion channels. Considering that passive sensory systems first evolved to gather
information on the environment, and given the costly nature of the sensing and
processing apparatus, sensory systems will have become more complex together
with the need for more complex information. However, once a system became suf-
ficiently complex to allow sophisticated communication, that system was already
12 Arik Kershenbaum

tied to a vital adaptive need of the animal to survive in its niche. For example,
the complex biosonar system of bats requires such sophisticated processing that
the animals have difficulty processing more than one stream of acoustic informa-
tion at a time (Barber et al. 2003). Similarly, in electric fish, separate neural and
sensory apparatus have evolved to process information originating from different
sources (Pothmann et al. 2012). Thus, in bats, dolphins (which also use biosonar),
and other animals, communicative content may be limited by the bandwidth of
a sensory and cognitive channel that is essential for gathering information about
their environments. Humans, of course, sense their environment primarily through
the visual channel, and so the development of complex acoustic communication
does not impede critical information about their surroundings. Naturally, other
constraints on the evolution of language exist, but this constraint of a dedicated
sensing channel remains a useful test for the degree of communicative complexity
that we may expect in a particular modality.
Given that one species that did evolve language, to what extent did the specific
characteristics of the communication channel of our ancestors constrain or enable
the rapid onset of such exceptional complexity? For the consideration of alien lan-
guages, the first question we must answer is whether the particular conditions that
led to a tipping point that made the evolution of language inevitable were rare or
unique, and whether those conditions might be expected to occur at some point on
any inhabited planet. In particular, in our discussion, to what extent was the nature
of the communication channel a factor in the selective pressure for language? From
our review, it seems almost inevitable that a linguistic communication system on
our planet would exist in the acoustic modality. Further, other modalities, perhaps
with the exception of electrical fields, seem unlikely candidates for language on
any planet. Nonetheless, perhaps the most important message from reviewing com-
munication strategies on Earth is that virtually every communication modality that
we can imagine evolving has in fact evolved, and as information on environmental
niches on exoplanets becomes available, we can use – cautiously – analogy with
adaptive strategies on Earth to speculate about what communicative solutions have
been found elsewhere.

References
Atema, J. 1995. “Chemical signals in the marine environment: Dispersal, detection, and
temporal signal analysis.” Proceedings of the National Academy of Sciences of the
United States of America 92 (1): 62–66.
Barber, J.R., K.A. Razak, and Z.M. Fuzessery. 2003. “Can two streams of auditory informa-
tion be processed simultaneously? Evidence from the gleaning bat antrozous pallidus.”
Journal of Comparative Physiology A 189 (11): 843–855.
Craven, B.A., E.G. Paterson, and G.S. Settles. 2010. “The fluid dynamics of canine ol-
faction: Unique nasal airflow patterns as an explanation of macrosmia.” Journal of the
Royal Society, Interface 7 (47): 933–943.
Déaux, Éloïse C., and Jennifer A. Clarke. 2013. “Dingo (Canis Lupus Dingo) acoustic rep-
ertoire: Form and contexts.” Behaviour 150 (1): 75–101.
 Many Ways to Say Things 13

Elias, Damian O., Eileen A. Hebets, Ronald R. Hoy, and Andrew C. Mason. 2005. “Seismic
signals are crucial for male mating success in a visual specialist jumping spider (Ara-
neae: Salticidae).” Animal Behaviour 69 (4): 931–938.
Fitch, W. 2000. “The evolution of speech: A comparative review.” Trends in Cognitive Sci-
ences 4 (7): 258–267.
Fitch, W. Tecumseh. 2005. “The evolution of language: A comparative review.” Biology and
Philosophy 20 (2–3): 193–203.
Gardiner, Jayne M., and Jelle Atema. 2010. “The function of bilateral odor arrival time dif-
ferences in olfactory orientation of sharks.” Current Biology 20 (13): 1187–1191.
Gould, James L. 2008. “Animal navigation: The evolution of magnetic orientation.” Current
Biology 18 (11): R482–R484.
Gregg, Justin. 2013. Are Dolphins Really Smart?: The Mammal Behind the Myth. Oxford:
Oxford University Press.
Hagedorn, Mary, and Randy Zelick. 1989. “Relative dominance among males is expressed
in the electric organ discharge characteristics of a weakly electric fish.” Animal Behav-
iour 38 (3): 520–525.
Haldane, J.B.S. 1927. Possible Worlds. London: Chato and Windus.
Harley, H.E. 2008. “Whistle discrimination and categorization by the atlantic bottlenose dol-
phin (Tursiops Truncatus): A review of the signature whistle framework and a perceptual
test.” Behavioural Processes 77 (2): 243–268.
Hauser, M.D., N. Chomsky, and W. Fitch. 2002. “The faculty of language: What is it, who
has it, and how did it evolve?” Science 298 (5598): 1569–1579.
Herman, Louis M., Adam A. Pack, and Amy M. Wood. 1994. “Bottlenose dolphins can
generalize rules and develop abstract concepts.” Marine Mammal Science 10 (1): 70–80.
Hockett, Charles F. 1960. “The origin of speech.” Scientific American 203 (3): 88–97.
Hopkins, C.D. 2009. “Electrical perception and communication.” Encyclopedia of Neuro-
science 3: 813–831.
Howard, J., B. Blakeslee, and S.B. Laughlin. 1987. “The intracellular pupil mechanism and
photoreceptor signal: Noise ratios in the fly Lucilia Cuprina.” Proceedings of the Royal
Society of London. Series B, Biological Sciences 231 (1265): 415–435.
Jackendoff, Ray. 2011. “What is the human language faculty?: Two views.” Language 87
(3): 586–624.
Kawasaki, Masashi. 2005. “Physiology of tuberous electrosensory systems.” In Electrore-
ception, edited by T. Bullock, C. Hopkins, A. Popper, and R. Fay, 154–194. Dordrecht:
Springer.
Kershenbaum, Arik, Holly Root-Gutteridge, Bilal Habib, Jan Koler-Matznick, Brian Mitch-
ell, Vicente Palacios, and Sara Waller. 2016. “Disentangling canid howls across multiple
species and subspecies: Structure in a complex communication channel.” Behavioural
Processes 124: 149–157.
Lorenzo, Daniel, Ana Silva, and Omar Macadar. 2006. “Electrocommunication in gymno-
tiformes: Jamming avoidance and social signals during courtship.” Communication in
Fishes 2: 753–779.
Meyer, J. Harlan. 1982. “Behavioral responses of weakly electric fish to complex imped-
ances.” Journal of Comparative Physiology 145 (4): 459–470.
Michelsen, Axel. 1973. “The mechanics of the locust ear an invertebrate frequency ana-
lyzer.” In Mechanisms in Hearing, edited by Aage R. Møller, 911–934. Dordrecht:
Springer.
Niven, J.E., and S.B. Laughlin. 2008. “Energy limitation as a selective pressure on the evolu-
tion of sensory systems.” The Journal of Experimental Biology 211 (Pt 11): 1792–1804.
14 Arik Kershenbaum

Pothmann, Leonie, Lon A. Wilkens, and Michael H. Hofmann. 2012. “Two modes of infor-
mation processing in the electrosensory system of the paddlefish (Polyodon Spathula).”
Journal of Comparative Physiology A 198 (1): 1–10.
Rajan, Raghav, James P. Clement, and Upinder S. Bhalla. 2006. “Rats smell in stereo.” Sci-
ence 311 (5761): 666–670.
Randall, J.A. 1989. “Individual footdrumming signatures in banner-tailed kangaroo rats Di-
podomys Spectabilis.” Animal Behaviour 38 (4): 620–630.
Randall, J.A., and E.R. Lewis. 1997. “Seismic communication between the burrows of
kangaroo rats, Dipodomys Spectabilis.” Journal of Comparative Physiology A 181 (5):
525–531.
Sayigh, L.S., P.L. Tyack, R.S. Wells, A.R. Solow, M.D. Scott, and A.B. Irvine. 1999. “Indi-
vidual recognition in wild bottlenose dolphins: A field test using playback experiments.”
Animal Behaviour 57 (1): 41–50.
Sayigh, Laela S., Randall S. Wells, and Vincent M. Janik. 2017. “What’s in a voice? Dolphins
do not use voice cues for individual recognition.” Animal Cognition 20 (6): 1067–1079.
Scheffel, Andreas, and Bernd Kramer. 2006. “Intra-and interspecific electrocommunication
among sympatric mormyrids in the Upper Zambezi River.” In Communication in Fishes,
edited by F. Ladich, S.A. Collins, P. Moller, and B.G. Kapoor. Enfield, New Hampshire:
Science Publishers.
Squire, Ann, and Peter Moller. 1982. “Effects of water conductivity on electrocommunica-
tion in the weak-electric fish brienomyrus niger (Mormyriformes).” Animal Behaviour
30 (2): 375–382.
Stebbins, William C. 1983. The Acoustic Sense of Animals. Cambridge, MA: Harvard Uni-
versity Press.
Stevens, Martin. 2013. Sensory Ecology, Behaviour, and Evolution. Oxford: Oxford Uni-
versity Press.
Stoddard, Philip K. 1999. “Predation enhances complexity in the evolution of electric fish
signals.” Nature 400 (6741): 254.
Su, Chih-Ying, Karen Menuz, and John R. Carlson. 2009. “Olfactory perception: Receptors,
cells, and circuits.” Cell 139 (1): 45–59.
Sullivan, John P., Sébastien Lavoue, and Carl D. Hopkins. 2000. “Molecular systematics
of the African electric fishes (Mormyroidea: Teleostei) and a model for the evolution of
their electric organs.” Journal of Experimental Biology 203 (4): 665–683.
Wiltschko, Roswitha, and Wolfgang Wiltschko. 2010. “Avian magnetic compass: Its func-
tional properties and physical basis.” Current Zoology 56 (3).
3
RECOGNIZING INTENTIONAL SIGNALS
AND THEIR MEANINGS IN NON-
HUMAN COMMUNICATION
Catherine Hobaiter, Adriano R. Lameira, and Derek Ball

Introduction
How can we tell if someone is trying to talk to us – especially when that someone
is not a member of our species (or even from our solar system)? And once we have
detected a signal, how can we tell what it means?
Every organism leaves evidence of its activity, and in some cases these cues can
be exploited to gain rich information. But in the study of communication, signals
sent with the intent to communicate are of special interest because they indicate
the existence of human-like forms of intelligence. The study of pre-verbal children
and non-human animal communication has provided toolkits for the diagnosis of
intentional communication. At present, these typically employ a Gricean approach,
which explores evidence that the signaler directs their communication to a specific
recipient, in order to achieve a particular goal. We will discuss the potential to adapt
the Gricean strategy to the interstellar context, where our knowledge of the envi-
ronment, interests, ideas, and purposes of our possible interlocutors is extremely
limited, and explore the idea that signal simplicity may play a crucial role.

Grice on Communication and Meaning

Imagine driving along a country road and noticing the driver of a passing car flash-
ing her headlights at you. You would probably assume that she was trying to tell
you something. But what? Figuring this out requires a bit of thought. The other
driver intends to help you – to give you information that you can use. Is there some-
thing wrong with your car? Suppose that a quick check makes this unlikely – your
lights are on, and there is no sign of smoke. Most likely, you might conclude, the
passing driver knows something about the road ahead; the driver wants to inform

DOI: 10.4324/9781003352174-3
16 Catherine Hobaiter, Adriano R. Lameira, and Derek Ball

you that there is reason to be cautious, which means you should reduce your speed
and be careful.
Similar reasoning is often involved in understanding linguistic communication.
To adapt a famous example (Grice 1991: 33), fully understanding a letter of recom-
mendation for an academic position that reads only “The candidate has excellent
handwriting and is always very punctual” requires us to reason along the follow-
ing lines: the author of the recommendation intends to give the reader information
relevant to the issue of whether the candidate is suitable for the position, and is in
possession of such information; information about the candidate’s handwriting and
punctuality is not relevant; therefore, there must be further information that the
author is unwilling (perhaps for reasons of politeness) to share; and this can only
be information to the effect that the candidate is grossly unsuitable.
The philosopher Paul Grice developed a view of meaning that provides several
insights into how communication works in this kind of case. For our purposes, two
of Grice’s themes will be particularly important: (i) the idea that communication
is a cooperative endeavor, and (ii) the idea that meaning (in one interesting sense)
is a matter of acting with particular intentions. To get an idea of the motivation for
these themes, let’s unpack the very beginning of our reasoning about the passing
driver: “The other driver intends to help you – to give you information that you can
use.” Note first that we assume that the other driver is trying to help: we have cer-
tain desires (to drive safely, avoid accidents, etc.), she knows this, and she intends
to help us satisfy those desires. This assumption is crucial to the reasoning that fol-
lows: without it, even if we could somehow come to know that the passing driver is
trying to communicate, we would have no way of making even a reasonable guess
about what she is trying to say.
Grice formulates the idea that communication is cooperative as an injunction to
communicators:

The Cooperative Principle: “Make your conversational contribution such as is

required, at the stage at which it occurs, by the accepted purpose or direction of
the talk exchange in which you are engaged.”
(1991: 26)

Grice identifies several “maxims,” observing which (he claims) “will, in gen-
eral, yield results in accordance with the Cooperative Principle” (1991: 26); these
include principles such as “Try to make your contribution one that is true” and
“Be perspicuous.” (We will return to the status of these maxims in the context of
interstellar communication.)
Note also what we took the passing driver’s cooperation to consist of: she
intends to give us certain information, to make us form a certain belief about the
conditions of the road ahead. Grice took this to be crucial to a certain kind of
meaning that he called “non-natural meaning” (to contrast with cases of “natural
 Intentional Signals, Their Meanings in Non-Human Communication 17

meaning” or cues – roughly, reliable correlation – ascribed in claims like “smoke

means fire”) (1991: 213–214). But not every action performed with the intention
that others form a belief is meaningful in Grice’s sense. I may plant misleading
evidence at a crime scene, intending that investigators will form the belief that
my rival is the culprit; but Grice would deny that the misleading evidence (or my
planting it) is meaningful in the relevant sense. What is crucial to communica-
tion that the non-naturally meaningful is that the communicator intends that her
audience form a certain belief on the basis of recognizing that very intention.
Non-natural meaning is thus very much connected to interaction, an engagement
between entities. For example, when we saw the passing driver flash her lights,
we recognized that she was trying to tell us something – in other words, that she
intended us to take up some bit of information or form some belief – and the
driver intended us to recognize this. In Grice’s view, this is what distinguishes
her light flashing from other signals that might (inadvertently) alert us to the
danger ahead (such as the screech of her brakes in the distance, or her concerned
expression).
As the passing driver example shows, communication involving non-natural
meaning in the Gricean sense can take place without language. But typical uses of
language are paradigmatic cases of intentional communication, and Grice maintains
that the notion of non-natural meaning is crucial to understanding the meaningful
use of language. Grice (1991: 89–90) distinguishes between occasion meaning –
roughly, what a particular communicator meant by a use of a particular word or
phrase in a particular set of circumstances – and timeless meaning – roughly, the
meaning that every use of that word or phrase has in common (sometimes called
the literal or semantic meaning). For example, the occasion meaning of the recom-
mendation letter is (roughly) that the candidate is unsuitable; but that is no part of
the timeless meaning of the sentence “The candidate has excellent handwriting and
is always very punctual,” because many possible uses of that sentence would not
mean that the candidate is unsuitable.
In Grice’s view, occasion meaning is non-natural meaning; one occasion means
“the candidate is unsuitable” just in case one acts with the intention that one’s
audience forms the belief that the candidate is unsuitable on the basis of recogniz-
ing that very intention. And Grice argues further that timeless meaning should be
accounted for in terms of occasion meaning: abstracting away from the details, his
idea is that a sentence has a particular timeless meaning for a group just in case
the members of that group have what he calls a “procedure in their repertoire” –
roughly, a disposition – to use that expression with that occasion meaning, and it
is common knowledge that this is so. Thus, for example, the sentence “The can-
didate has excellent handwriting” timelessly means (for English speakers) that the
candidate has excellent handwriting because we are disposed in appropriate cir-
cumstances to use the sentence to occasion-mean that the candidate has excellent
handwriting, and we all know this fact about each other.
18 Catherine Hobaiter, Adriano R. Lameira, and Derek Ball

Intentional Communication in Non-Human Animals:

The Gricean Toolkit

Communication with other members of our own species can be very challeng-
ing. Even when we share the same language, we must infer the occasion meaning
a signaler intends to communicate. Even small differences in our life experience
and day-to-day events (an argument with a friend, our team winning the game)
affect the layers of meaning in any phrase. Our ideas shape our language, which
may in turn shape our ideas (Lupyan and Dale 2010). Communication across more
substantial cultural divides presents an even greater challenge. What hope do we
have of decoding speaker meaning (in the Gricean sense) in another species? One
approach has been to try to teach other species to use human systems (Gardner and
Gardner 1969; Pepperberg, this volume [Chapter 6]) but doing so has required not
only that these individuals not only share a human environment, but that they are
also raised in one from infancy. Not only is this situation improbable in the context
of communication with extraterrestrial species (Kryptonians aside), it does not pro-
vide us with a means to communicate with others of their species. An enculturated
chimpanzee raised as a ‘human child’ may be almost as poorly equipped to under-
stand the ideas and purposes of a wild chimpanzee as we are.
An alternative is to decode the information in a species’own communication, and
while doing so try to eliminate (or at least recognize the limitations of) our human-
centric perceptual biases (see also Slobodchikoff, this volume [Chapter 9]; Herzing,
this volume [Chapter 4]; Pepperberg, this volume [Chapter 6] Grund et al. 2023). The
first problem any signaler or recipient faces is how to distinguish a signal from the
noise. How do we ensure that our signal is recognized as such? Many other authors
in this volume focus on this part of the puzzle. There is increasingly widespread
evidence for ‘universal’ patterning of information (see Berea, this volume [Chapter
7]); for example, Zipf’s Law of brevity and Menzerath’s law – universal to human
languages – are also present in other systems of information from gibbon calls to
genomic structure (Zipf 1949; Gustison et al. 2016; Ferrer-i-Cancho and Lusseau
2009; Huang et al. 2020; Heesen et al. 2019; Safryghin et al. 2022). Humans, like
many species, acquire the fundamentals of communication easily, intuitively, and
without a rich input – suggesting that there are biologically inherited components
to our language acquisition (Grice 1991; Senghas et al. 2004; Hauser 2002; Nixon
and Tomaschek, this volume [Chapter 17]; Roberts et al., this volume [Chapter 15];
Samuels and Punske, this volume [Chapter 16]) or language production (Harbour,
this volume [Chapter 18]). From continuous streams of sound or movement, human
and non-human animals parse out discreet signals (e.g., Barutchu et al. 2008; Kuhl
1979) in both our own and other species’ communication (e.g., Kuhl and Miller
1975; Kamiloğlu et al. 2020). Our species is able to recombine these signals into
hierarchically structured, recursive sequences (syllables, words, sentences) that,
from a finite set of – for example – sounds, allow us to communicate any idea we
can imagine. While humans have a bias toward auditory and visual communication
 Intentional Signals, Their Meanings in Non-Human Communication 19

(Ortner, this volume [Chapter 5]), across species on Earth, information from cues
and signals is decoded from a far more diverse range of modalities, from touch to
magnetism (see Kershenbaum, this volume [Chapter 2]). However, once we have
passed the step of discriminating a signal from background noise, as generated by
someone, successful communication of ideas requires we also recognize whether it
was intended as a message.
Many things signal information: the color of a tree’s leaves tells us something
about season; the color, smell, and feel of a berry tell us whether we should eat it.
This information is broadcast to the world irrespective of whether or not an audi-
ence is there or can see it: a tree does not change the color of its leaves depend-
ing on who is looking at it, and a ripe berry tastes just as good at night when its
color is hidden. Most animal signals are similar; the bright pink genital swelling
of a female chimpanzee is on display throughout her fertile period, irrespective of
which males are around (or if there are any there at all). It is easy to make a dis-
tinction between these fixed signals and human language – but what about those
that seem superficially more similar? There may be rich nuanced information in a
bird’s song, a wolf’s howl, or a monkey’s alarm call (Templeton et al. 2005; Ker-
shenbaum et al. 2016; Pepperberg, this volume [Chapter 6]; Ouattara et al. 2009).
But when a vervet monkey gives an eagle alarm call, does he intend for the other
members of his group to understand that there is an eagle there and that they need
to flee? These calls reliably elicit this behavior from other vervet monkeys (Cheney
and Seyfarth 1992), and careful experiments have shown that similar alarm calls
across primate species generate a mental representation of the threat in the recipi-
ents (Zuberbühler et al. 1999). They are called functionally referential – they func-
tion as if they refer to an eagle. There is a reason for the caveat: as yet, we have
no evidence to suggest that the signaler intends to refer the recipients’ attention to
the eagle. Absence of evidence for intention cannot confirm that these signals are
produced without intention, and they are certainly produced with greater flexibility
than we find in fall leaf colors or female sexual swelling. Signalers may call more
loudly or for longer if an audience is present (Schel, Machanda et al. 2013; Cheney
and Seyfarth 1992; Wich and de Vries 2006), or if specific types of individuals are
present (Schel, Machanda et al. 2013; Crockford et al. 2012; Slocombe and Zuber-
bühler 2007). However, these effects can typically be explained without invoking
the communicative intentions that appear to be fundamental to language use – for
example, variation in whether and how a call is produced may be modulated by
similar variation in physiological arousal, generated by the presence of (relevant)
social peers. Humans show the same behavioral flexibility in our broadcast signals.
In a context whereby you would smile when alone, you smile more – or laugh
longer – if others are present (Gervais and Wilson 2005; Devereux and Ginsburg
2001; Wild et al. 2003; Provine 1992). In our species, language did not replace this
system of broadcast signals – we continue to broadcast information, whether an
involuntary blush, laugh, or yelp (see also: Ortner, this volume [Chapter 5], for how
non-linguistic signals can be co-opted when language’s original modality becomes
20 Catherine Hobaiter, Adriano R. Lameira, and Derek Ball

unavailable). Instead, language provides us with a different system of signaling,

one in which we can choose whether and how to share information depending on
our audience, and on what we know about them, including what they already know.
Grice’s view of meaning requires that, in meaningful communication, the signal
be produced with the intent that the audience form a belief on the basis of recog-
nizing that intention. But even among humans, not all interesting communicative
behaviors need be accompanied by intentions with such complex structure. Follow-
ing Dennett (1983), we may distinguish different levels of complexity in the inten-
tions associated with a signal. In the simplest case, a signal may be produced with
no intentions at all. Signals like this could be produced even by what Dennett calls
a zero-order intentional system: a creature with no understanding of other minds,
no attitudes at all about what their audience believes or does. Fully Gricean commu-
nication would require what he describes as a second-order (or higher) intentional
system, which is capable of thinking about the mental states of other creatures: the
signaler intends that the recipient (at least) recognize their intention. But it is pos-
sible that signals be produced with some intention, but not with a full-blown Gricean
intention. For example, a signaler may intend that her audience change her behavior
without having any intentions (or any attitudes at all) about the audience’s mental
state – about what the audience recognizes or believes. To date, the exploration of
intentional communication in non-human species has largely focused on whether
they meet the criteria for this broader, first-order use of intentional communication;
and we will go on to speak of the meanings of intentional communication of this
kind, while recognizing that these may fall short of Gricean non-natural meaning.
On first observing her, we have no way of interrogating a primate signaler (or
an extraterrestrial life form, or a pre-verbal human infant for that matter) to find
out what she means when she signals; but we can employ behavioral cues that
together indicate intentional use. Typically, these include: checking whether or not
the audience can see visual signal and, if not, adjusting signal selection to other
modalities or changing signaler position to bring the signal into view; waiting’;
waiting for a response after an initial signal; and, if this signal fails, persisting or
perhaps elaborating with further signals. Consider a human example: if I want you
to pass me the coffee, I should not point to it if you are facing in the wrong direc-
tion; I would either use an audible or tactile signal (speaking or tapping you on the
shoulder), or move into a position where you can see me (Tanner and Byrne 1996;
Pika et al. 2005; Tomasello et al. 1985). Similarly, and despite the fact that my
desire for coffee is, frequently, a continuous stimulus, I would be unlikely to repeat
an endless string of “give-me-coffee-give-me-coffee-give-me-coffee.” I would ask
once and wait to see what you did next. If you did not pass it to me, I might then
ask again and then, if I have failed a couple of times, elaborate by adjusting the
signals I use (for example: adding a point to the coffee pot). I would also stop com-
municating once you pass me the coffee (my intended goal), despite the fact that
my desire for coffee would not be satisfied until after I start drinking it. The same
behavior accompanies ape gestural communication. They adjust their selection of
 Intentional Signals, Their Meanings in Non-Human Communication 21

signal types to the audiences’ visual attention and wait for a response after signaling
(Tanner and Byrne 1996; Pika et al. 2003; Tomasello et al. 1985; Genty et al. 2009;
Hobaiter and Byrne 2011a; McCarthy et al. 2013); and when signals fail, they per-
sist (Hobaiter and Byrne 2011b; Liebal et al. 2004; McCarthy et al. 2013; Leavens
et al. 2005). They also elaborate (Hobaiter and Byrne 2011b; Leavens et al. 2005),
distinguishing between likely reasons for failure – whether the recipient partially
or completely misunderstood the initial signal (Cartmill and Byrne 2007). While
there is no single panacea with which we can diagnose a signaler’s intent, taken
together, these steps provide a Gricean toolkit for the recognition of intentional
communication.
The intention to communicate to a particular audience is what takes us beyond
broadcasting information and to a point where we can ask: what does the signaler
mean to say? Early work exploring intentional communication looked at how lan-
guage develops in young children. Bates and colleagues distinguished illocutory
acts, in which an infant employed a conventionalized signal toward a recogniz-
able goal, from perlocutory acts, in which a signal changed a recipient’s behav-
ior, but without any evidence that this effect was intended by the signaler (Bates
et al. 1975). At around the same time, the first field studies of wild chimpanzees
were providing descriptions of their vocal and gestural behavior. In early investiga-
tions, vocalizations appeared relatively fixed in terms of their content, and strongly
linked to signaler affect (Goodall 1986). In contrast, gestures were combined and
used across behavioral contexts in a way that suggested “openness . . . one of the
most characteristic design features of human language” (Plooij 1978: 127). Today,
there is a more nuanced understanding of the different types of signals produced by
non-human apes (hereafter apes), with increasing flexibility demonstrated in their

BOX 3.1 DEFINITION OF TERMS USED IN COMMUNICATION

Cue Passive trait that provides information to a recipient, but on which

there was no selection for it to serve this function. For example:
the color of fall leaves acts as a cue that provides information on
the season.
Signal Behavior or characteristic that has evolved to convey information
to a potential recipient. For example: the color of a berry, or the
pink of a female primate’s genital swelling, have both evolved to
provide information on ripeness, or female ovulation.
Intentional Signals produced by a signaler toward a specific recipient with
signal the intent to change their behavior or state of knowledge. For
example: the signaler intends for that the recipient to understand
she should go away.
22 Catherine Hobaiter, Adriano R. Lameira, and Derek Ball

vocal repertoire (Slocombe and Zuberbühler 2007; Schel, Townsend, et al. 2013;
Crockford et al. 2012; Wich et al. 2012). Nevertheless, despite extensive research
effort across species, it remained challenging to demonstrate intentional signal use in
non-human communication (Rendall et al. 2009; Seyfarth and Cheney 2003). Sup-
porting evidence found was typically limited to single signals used in evolutionarily
important contexts; for example, a chimpanzee snake-alarm call (Schel, Townsend,
et al. 2013; Crockford et al. 2012). One consistent exception is the large repertoires
of great ape gestures – in which we find abundant evidence for intentional use across
eighty or more gesture types used in everyday communication (Tomasello et al.
1985; Pika et al. 2003, 2005; Cartmill and Byrne 2007, 2010; Pollick and de Waal
2007; Genty et al. 2009; Hobaiter and Byrne 2011a; Graham et al. 2017; Bard et al.
2014; Bard 1992; Leavens et al. 1996). From this body of research, we have been
able to explore not only the context in which apes gesture, but the intended meanings
for which they use individual gestures in their repertoires (Hobaiter and Byrne 2014;
Graham et al. 2018 Hobaiter et al. 2022). Today, the distinction between gestural and
vocal signals is blurring; for example, orangutan signalers use their hands to modify
the sound of their calls (Hardus et al. 2009), and all apes regularly combine different
signal types in a single message (Wilke et al. 2017; Hobaiter et al. 2017; Genty et al.
2014). Given that the capacity for intentional communication has been established
in their gestures and at least one vocalization, it is likely that future research will
extend the range of signals and channels (and species) in which we are able to investi-
gate ape signalers’ intended meaning. Fundamentally, language-like communication
can occur independently of modality. While vocal communication is concentrated in
the acoustic channel (although, see McGurk and MacDonald 1976), gestural com-
munication employs visual, acoustic, and tactile information, and many other chan-
nels (for example olfactory, electrical, and magnetic: see Kershenbaum, this volume
[Chapter 2]; Ortner, this volume [Chapter 5]) can be employed.
A recent study found intriguing evidence of an intentional hunting gesture in
grouper fish (Vail et al. 2013). This finding could be interpreted in two ways. On the
one hand, it might be that intentional communication is much more widespread in
non-human communication than is currently recognized – a possibility that increases
the likelihood that intentionality is also present in living systems beyond Earth. If this
is true, then with time and tinkering, the current Gricean toolkit may enable the detec-
tion of these species through intergalactic signals. On the other hand, this evidence
could also indicate that while representing a particularly demanding test of an animal’s
cognition, the Gricean toolkit remains hackable in specific communicative circum-
stances by species that meet our current criteria – but nevertheless lack the human-like
cognition our own attempts at interstellar communication are aiming to reach.

Implications for Interstellar Communication

Importantly, the detection of intentionality is independent of signal complexity.
While some complex features may be at the core of human language, such as
 Intentional Signals, Their Meanings in Non-Human Communication 23

hierarchical structural architecture (Hauser 2002; Senghas et al. 2004; Roberts

et al., this volume [Chapter 15]), these tell us little about whether the signaler is
capable of intentional language-like communication. Take bird song, for exam-
ple, or the intricate sand patterns generated by courting pufferfish; these com-
plex, structured signals provide rich information for a passing recipient about
mate fitness, motivation, and vigor, but there is no evidence, to date, that these
are produced with an intent to communicate to a specific recipient. Similarly, a
computer algorithm may generate complex signal patterns in, again to date, the
pure absence of intention.

BOX 3.2 THE PROBLEM WITH GRICE

Gricean communication appears to require the ability to make complex infer-

ences about others’ mental states for even simple requests (e.g., “I want you
to recognize that I intend that you pass me the coffee”; Scott-Phillips 2015). If
true, we are faced with the conundrum that a 2- or 3-year old child, capable of
using language, may not be – explicitly – able to attribute these mental states
(Wimmer and Perner 1983; Liddle and Nettle 2006). How do children learn to
use language before they have the cognitive capacity to understand it?
One possibility is that the standard interpretation of Gricean intent has over-
stated the cognitive skills needed to acquire intentional language (Gómez 1994;
Moore 2016, 2017; Townsend et al. 2017). Alternatively, infants and great apes
may employ non-Gricean forms of intentional communication for which we
do not have an adequate toolkit (Leavens et al. 2017; Bar-On 2013). Alterna-
tive defnitions for intentional signal use tend to suffer from the same issue as
the original toolkits for detecting intentional use developed from the Gricean
defnition: a bias toward recording visual attention. In doing so, they limit our
ability to explore intentional use across a wide range of signal types. If the infor-
mation in the signal is primarily audible, as is the case in ape vocalizations and
long-distance gestures (such as buttress-drumming, for example), how do we
measure this? The signaler has no need to check the recipient’s auditory state of
attention; as long as they are aware of the recipient’s location, they are aware
of whether the signal can be received. As a result, we likely underestimate the
frequency with which non-humans employ intentional communication. The
absence of any evidence to date in areas such as cetacean or elephant commu-
nication, or in a wider range of vertebrate communication, may have more to
do with the restrictive nature of the Gricean toolkit than with the true absence
of intentional communication in these species.
24 Catherine Hobaiter, Adriano R. Lameira, and Derek Ball

Would a Gricean framework allow us to decode intentional communication from another

world? As Grice illustrates with his discrimination of occasion meaning from timeless
meaning, decoding intent can depend to a very large extent on context. Our (still lim-
ited) knowledge of ape behavior, body plans and ecology provides us with a foundation
from which to explore signaler intent and meaning on Earth. However, across interstel-
lar space, two interlocutors would share virtually no common ground. Nevertheless,
there is hope – the recognition of intentional communication depends on the behavioral
interaction of the interlocutors rather than decoding the signal itself. We should there-
fore distinguish two steps in the process of understanding intentional communication –
detecting the presence of an intentional signal, and coming to understand the meaning of
that signal – and the Gricean account can give us some purchase on each step.

Step 1: Detecting Intention

When we consider the question of detecting intention in interstellar signals, there
are two importantly different cases to consider. First, it may be that extraterrestri-
als are sending a signal to us, or just broadcasting a signal in an attempt to make
their presence known. We will call this case extraterrestrial–human communi-
cation. Alternatively, we may inadvertently pick up on a signal that is intended
for other extraterrestrials. We will call this case extraterrestrial–extraterrestrial
communication.
Most empirical work on non-human animals in the Gricean paradigm is focused
on communication between animals of the same species: for example, chimpanzees
gesturing to other chimpanzees. It might therefore be thought that the most promis-
ing possible application of Gricean tools is to the case of extraterrestrial–extrater-
restrial communication. However, the extent to which the Gricean toolkit can be
applied depends on exactly what we have detected. We are able to detect intention
in ape communication because we can see the two-way interaction between com-
municator and audience. Information from both sides of the interaction is required
in order to pick up on the characteristic features of intentional communication:
making sure the audience can detect the signal, waiting for a response, and per-
sisting or elaborating. With a single signal, the opportunities for investigating its
intentional nature would be limited; it is the interaction between two or more par-
ties that allows us to employ our Gricean toolkit to investigate whether the signaler
intended to communicate.
Ideally, what is required to build up a richer evidence base of the kind that might
help us decide whether a signal is intentional is an exchange of signals. If we could
somehow pick up on an entire exchange of signals between extraterrestrials, we
would have a greater chance of gathering evidence (such as repetition and elabora-
tion) that would bear on whether the signals were intentional. But the case of extra-
terrestrial–human communication would provide even better opportunities, since
in that case we could generate our own exchange. Signal repetition and elaboration
are strategies that can be easily explored in the exchange of signals in which you
 Intentional Signals, Their Meanings in Non-Human Communication 25

are not yet able to decode meaning, but can nevertheless recognize intent (Moore
2014). In investigating the intentional nature of the communication rather than
attempting to understand what it means, signal simplicity may be key. Through
the exchange of simple signals, we are able to explore the structure of the commu-
nication without the risk of error incorporated by signal variation or flexibility. A
common solution found in nature to establish communicative contact is to bounce
back an exact copy of the original signal several times (before starting to adjust it).
Here, signal simplicity prevents information leakage before the nature of the social
relationship is established. The likelihood of even a simple signal being repeated
exactly back and forth by chance is low. If the form of one signal is further ran-
domly altered with each iteration being sent individually before an exact response
is received, then an agent’s intentional action would be a (or perhaps the only)
plausible explanation of any such repetition. Here it is not the signal content that
is language-like; it is the way in which the signal is used. The repeated exchange
of a single simple signal allows you to indicate cooperation as a partner, positive
engagement in the communication, and your intent to respond. (Thus, the Gricean
approach justifies conclusions reached on other grounds by Granger et al., this
volume [Chapter 8], and Ortner, this volume [Chapter 5], among others.) With this
step, we may be able to establish our intent to communicate; however, we cannot
establish meaning by repeating back to someone only what they have already said.

Step 2: Decoding Meaning

In the study of non-human ape communication, researchers initially identify the
meaning of specific signals by exploring patterns in repeated signal use across
individuals and contexts. As in the study of many animal signals, including primate
vocalizations, initial exploration of ape gestures used behavioral context as a proxy
for meaning: feeding gestures or traveling gestures (Goodall 1986; Crockford and
Boesch 2003; Ouattara et al. 2009; Cheney and Seyfarth 1992; Genty et al. 2009;
Hobaiter and Byrne 2011a; Pika et al. 2005; Pollick and de Waal 2007; Tomasello
et al. 1985). Using established behavioral ethograms, these meanings can then be
reliably established by the direct observation of behavior that we easily recognize:
you can see that your signaler is feeding, or resting. However, not only is this
approach not possible in interstellar communication, it may also not be well suited
to exploring meaning on Earth. Take the example when a chimpanzee intends to
communicate the meaning: !Stop. We observe her signaling this with a gesture pro-
duced when she is feeding (and someone tries to take her food), when she is resting
(and a boisterous juvenile tries to play with her), and when she is traveling (and her
young infant tries to run off). By recording the context of use, her gesture’s meaning
appears to be ambiguous or flexible – while in fact it is both clear and consistent.
Instead, to establish her meaning in the Gricean sense of what she intends, we must
employ both signaler behavior and recipient response, specifically: the behavioral
change in the recipient that stops the signaler from signaling (Cartmill and Byrne
26 Catherine Hobaiter, Adriano R. Lameira, and Derek Ball

2010; Hobaiter and Byrne 2014; Graham et al. 2018). A recipient’s response may
be a refusal, or indicate misunderstanding, in which case we would expect a sig-
naler to persist toward her goal. Take our earlier example: I ask you to pass me the
coffee; if you refuse, or pass me the tea, I will ask you again. I may even – after
several attempts – elaborate. Occasionally, I may give up. But the one thing that
will consistently and across different attempts and partners stop me from asking
for the coffee, is when I am passed the coffee. In this way, we have been able
to establish that great apes employ their gestures toward at least nineteen mean-
ings (Hobaiter and Byrne 2014; Graham et al. 2018; Byrne et al. 2017). To date,
this means of exploring signaler meaning is itself limited to particular types of
meaning – imperative demands that require a consistent behavioral response from
the recipient, but other meanings may be expressed (Hobaiter et al. 2022).
In the case of extraterrestrial–extraterrestrial communication, our very limited
knowledge makes it hard to see how we could apply Gricean resources to attempt
to interpret or decode a signal. Even in the case of extraterrestrial–human com-
munication, in which a two-way exchange of signals has been established, very
significant challenges remain. First, the standard applications of the Gricean toolkit
do not straightforwardly extend to declarative statements that do not require a con-
sistent behavioral response. (Because a declarative message such as “what a lovely
planet you have” does not call for any particular response, it is unlikely to be fol-
lowed by repetition or elaboration, or indeed any observable change in behavior at
all) Second, decoding meaning in this way requires hundreds – if not thousands –
of exemplars (Hobaiter and Byrne 2014). Obtaining such a rich base of signals to
work with will be difficult in the interstellar case, wherein signal return time would
likely be measured in years – if not decades, centuries, or even longer.
These considerations suggest that, while we may be able to establish intentional
communication, taking the step to understanding occasion meaning may be more
challenging. Without both signaler and recipient having access to information on the
physical and social context in which a communication system evolved, we appear
to be stuck. Occasion meaning only emerges in context. We can draw, for example,
a comparison to the decoding of ancient languages. In order to decode such a lan-
guage, we first needed to establish the repertoire of possible signals and symbols
used. In the case of human language, we could assume some basic shared syntac-
tic presuppositions, which emerge across human languages even without explicit
instruction (e.g., Senghas et al. 2004; Roberts et al., this volume [Chapter 15]).
However, the transition to decoding the meaning of a specific script was often based
on some understanding of communicative context and behavior. We were aware of
not only the tools available to communicate with, but also of what was likely to be
communicated in that context. A Rosetta stone for other animal species (see Slo-
bodchikoff, this volume [Chapter 9]) can be imagined in part because we recognize
how particular behavioral responses or ideas – fear, arousal, ‘predator,’ etc. – might
be expressed in species in which we share some context in common; for example,
physiology or socio-ecological environment (Graham et al. 2022). Cracking scripts
 Intentional Signals, Their Meanings in Non-Human Communication 27

of extinct languages or unknown or secret codes typically starts with breaking a

single item within the code in relation its possible meaning given away by context.
Once the first item is decoded, it opens the door to further decoding of other items
that appear in combination with the first. Simplicity in both signal and meaning
of the initial information received may be key to taking these first steps. Even if
one accepts the contentions of Samuels and Punske, this volume [Chapter 16], and
Roberts et al., this volume [Chapter 15], that many features of human language
may be shared with extraterrestrials, our attempts at parsing alien messages are
more likely to meet with success if we can begin with simple messages and work
our way up.
Even with the advantage of knowing the context of a signal, there remain
challenges. Systems of communication are adapted to the needs of the species
employing them (Bradbury and Vehrencamp 1998). For example, a change in the
environment (such as increased traffic noise) leads to a change in the signaling
(e.g., in bird song; Slabbekoorn and Peet 2003; Nemeth and Brumm 2010; Gross et
al. 2010). Communication is also bound to the socio-ecological niche of a particu-
lar cultural group, or species, whether humans, primates, mammals, or Earthlings.
Thus, interpreting meaning requires that both parties in the exchange can make a
series of presuppositions about the context in which the signal is given. Without
these, the signal is at best ambiguous, at worst meaningless. We can see the impor-
tance of this in cross-cultural exchanges within our species, when even the basic
positive or negative valence of an item varies: is a thumbs-up gesture supportive
or obscene? Grice tells us that communication is cooperative, but we can cooper-
ate only when we have a sense of what our audience wants; and we can exploit the
fact that someone is cooperating in order to interpret her, only if we know what
she thinks we want. If we humans typically obey Grice’s maxims to speak truth-
fully and be perspicuous, that is because doing so typically serves what we want
and what we take others to want. The presupposition that other species share these
desires may be problematic; extraterrestrials may greatly prefer praise to truth, or
politeness (as they see it) to perspicuity. (Thus, the Gricean account casts some
doubt on the distinction between understanding the message and understanding
from a cultural context suggested by Ross, this volume [Chapter 12]; understand-
ing the message may require much information about cultural context.) In short,
understanding what someone means does not require sharing their same commu-
nicative context; but we do need sufficient information about our interlocutors to
either select or interpret even simple signals appropriately.
Gaining such information is extremely problematic when our interlocutors are
light years away (see Slobodchikoff, this volume [Chapter 9], for discussion of
related problems). But perhaps the task is not hopeless. We should not underesti-
mate the common ground with which we begin: an exchange of signals in which
there is evidence of the intent to communicate already establishes some presup-
positions about the recipient: for example, that they are capable of communication
in which we can potentially recognize their intentions. And perhaps we can gain
28 Catherine Hobaiter, Adriano R. Lameira, and Derek Ball

leverage on the problem by exploring the full breadth of human and non-human
animal signal systems on Earth. In doing so, we may be able to establish patterns of
consistency (natural probabilistic universals) across species that provide a starting
place for exploring the context of communication with an interstellar species; for
example, an increase in relative speed or pitch of signaling typically indicates an
increase in arousal, with the reverse being also true. A mirroring or synchronizing
of behavior typically signals the intention to strengthen and extend a positive social
interaction in humans and non-humans. In contrast, an increase in the distance
between signalers indicates neutral or negative affect, while elaboration may signal
both positive interest and social challenge. If we have strong reasons to expect that
an alien language is culturally inherited and transmitted socially across individuals,
then some of the same features that are known to universally enable the learning of
languages could also be expected to have emerged across different planets. Such a
systematic comparison across species signaling may also provide us with possible
methods to expand or evolve our Gricean toolkit. Similarly, new techniques from
the study of related areas of cognition, such as theory of mind, may allow us to
recognize a mind capable of intentional communication as indicated in other areas
of behavior outside of signal exchange.

Conclusion
Grice’s idea that communication is a cooperative activity involving intentional
action has been productively utilized in research on animal communication, and
the resulting Gricean toolkit has promise for recognizing attempted interstellar
communication, as well. But coming to understand what someone is telling us
requires understanding a great deal, not only about their communicative faculties,
but also about their aims and desires, and potentially also how they think about
our aims and desires; and it hardly needs to be said that understanding the aims
and desires of extraterrestrials prior to establishing an effective means of com-
munication is an extremely difficult (though, as we have suggested, perhaps not
impossible) task. The Gricean toolkit does not provide a simple recipe for under-
standing xenolanguage – but perhaps it can point us towards useful directions for
further inquiry.

References
Bard, Kim A. 1992. “Intentional behavior and intentional communication in young free-
ranging orangutans.” Child Development 63 (5): 1186–1197.
Bard, Kim A., Sophie Dunbar, Vanessa Maguire-Herring, Yvette Veira, Kathryn G. Hayes,
and Kelly McDonald. 2014. “Gestures and social-emotional communicative develop-
ment in chimpanzee infants: Gestural development in chimpanzees.” American Journal
of Primatology 76 (1): 14–29.
Bar‐On, Dorit. 2013. “Origins of meaning: Must we ‘go gricean’?” Mind & Language 28
(3): 342–375.
 Intentional Signals, Their Meanings in Non-Human Communication 29

Barutchu, Ayla, Sheila G. Crewther, Patricia Kiely, Melanie J. Murphy, and David P.
Crewther. 2008. “When /b/ill with /g/ill becomes /d/ill: Evidence for a lexical effect in
audiovisual speech perception.” European Journal of Cognitive Psychology 20 (1): 1–11.
Bates, Elizabeth, Luigia Camaioni, and Virginia Volterra. 1975. “The acquisition of per-
formatives prior to speech.” Merrill-Palmer Quarterly of Behavior and Development
21 (3).
Bradbury, Jack W., and Sandra Lee Vehrencamp. 1998. Principles of Animal Communica-
tion. Sunderland, MA: Sinauer Associates.
Byrne, Richard W., Erica Cartmill, Emilie Genty, Kirsty E. Graham, Cathrine Hobaiter, and
Joanne Tanner. 2017. “Great ape gestures: Intentional communication with a rich set of
innate signals.” Animal Cognition 20 (4): 755–769.
Cartmill, Erica A., and Richard W. Byrne. 2007. “Orangutans modify their gestural signal-
ing according to their audience’s comprehension.” Current Biology 17 (15): 1345–1348.
Cartmill, Erica A., and Richard W. Byrne. 2010. “Semantics of primate gestures: Intentional
meanings of orangutan gestures.” Animal Cognition 13 (6): 793–804.
Cheney, Dorothy L, and Robert M Seyfarth. 1992. How Monkeys See the World: Inside the
Mind of Another Species. Chicago, IL: University of Chicago Press.
Crockford, Catherine, and Christophe Boesch. 2003. “Context-specific calls in wild chim-
panzees, pan troglodytes verus: Analysis of barks.” Animal Behaviour 66 (1): 115–125.
Crockford, Catherine, Roman M. Wittig, Roger Mundry, and Klaus Zuberbühler. 2012.
“Wild chimpanzees inform ignorant group members of danger.” Current Biology 22 (2):
142–146.
Dennett, Daniel C. 1983. “Intentional systems in the cognitive ethology: The ‘Panglossian
paradigm’ defended.” The Behavioral and Brain Sciences 6: 343–390.
Devereux, Paul G., and Gerald P. Ginsburg. 2001. “Sociality effects on the production of
laughter.” The Journal of General Psychology 128 (2): 227–240.
Ferrer-i-Cancho, Ramon, and David Lusseau. 2009. “Efficient coding in dolphin surface
behavioral patterns.” Complexity 14: 23–25.
Gardner, R. Allen, and Beatrix T. Gardner. 1969. “Teaching sign language to a chimpanzee.”
Science 165: 664–672.
Genty, Emilie, Thomas Breuer, Catherine Hobaiter, and Richard W. Byrne. 2009. “Gestural
communication of the Gorilla (Gorilla Gorilla): Repertoire, intentionality and possible
origins.” Animal Cognition 12 (3): 527–546.
Genty, Emilie, Zanna Clay, Catherine Hobaiter, and Klaus Zuberbühler. 2014. “Multi-modal
use of a socially directed call in bonobos.” PLoS One 9 (1): e84738.
Gervais, Matthew, and David Sloan Wilson. 2005. “The evolution and functions of laughter
and humor: A synthetic approach.” The Quarterly Review of Biology 80 (4): 395–430.
Gómez, Juan Carlos. 1994. “Mutual awareness in primate communication: A gricean ap-
proach.” In Self-Awareness in Animals and Humans: Developmental Perspectives,
61–80. New York, NY, US: Cambridge University Press.
Goodall, Jane. 1986. The Chimpanzees of Gombe: Patterns of Behavior. Cambridge, IL:
Cambridge University Press.
Graham, Kirsty E., Takeshi Furuichi, and Richard W. Byrne. 2017. “The gestural repertoire
of the wild bonobo (Pan Paniscus): A mutually understood communication system.” Ani-
mal Cognition 20 (2): 171–177.
Graham, Kirsty E., Catherine Hobaiter, James Ounsley, Takeshi Furuichi, and Richard W.
Byrne. 2018. “Bonobo and chimpanzee gestures overlap extensively in meaning.” PLoS
Biology 16 (2): e2004825.
30 Catherine Hobaiter, Adriano R. Lameira, and Derek Ball

Graham, Kirsty E., Gal Badihi, Alexandra Safryghin, Charlotte Grund, and Catherine Ho-
baiter. 2022. “A socio-ecological perspective on the gestural communication of great ape
species, individuals, and social units.” Ethology, Ecology, & Evoltuion 34 (3): 235–259.
Grice, H. Paul. 1991. Studies in the Way of Words. Cambridge, MA: Harvard University Press.
Gross, Karin, Gilberto Pasinelli, and Hansjoerg P. Kunc. 2010. “Behavioral plasticity allows
short‐term adjustment to a novel environment.” The American Naturalist 176 (4): 456–464.
Grund, Charlotte, Gal Badihi, Kirsty E. Graham, Alexandra Safryghin, and Catherine
Hobaiter. 2023. “GesturalOrigins: A bottom-up framework for establishing system-
atic gesture data across ape species.” Behavior Research Methods doi:10.3758/
s13428-023-02082-9.
Gustison, Morgan L., Stuart Semple, Ramon Ferrer-i-Cancho, and Thore J. Bergman. 2016.
Gelada vocal sequences follow Menzerath’s linguistic law.” Proceedings of the National
Academy of Sciences 113: e2750–2758.
Hardus, Madeleine E., Adriano R. Lameira, Carel P. Van Schaik, and Serge A. Wich. 2009.
“Tool use in wild orang-utans modifies sound production: A functionally deceptive inno-
vation?” Proceedings of the Royal Society B: Biological Sciences 276 (1673): 3689–3694.
Hauser, Marc D. 2002. “The faculty of language: What is it, who has it, and how did it
evolve?” Science 298 (5598): 1569–1579.
Heesen, Raphaela, Catherine Hobaiter, Ramon Ferrer-i-Cancho, and Stuart Semple. 2019.
“Linguistic laws in chimpanzee gestural communication.” Proceedings of the Royal So-
ciety B 286: e20182900.
Hobaiter, Catherine, and Richard W. Byrne. 2011a. “The gestural repertoire of the wild
chimpanzee.” Animal Cognition 14 (5): 745–767.
Hobaiter, Catherine, and Richard W. Byrne. 2011b. “Serial gesturing by wild chimpanzees:
Its nature and function for communication.” Animal Cognition 14 (6): 827–838.
Hobaiter, Catherine, and Richard W. Byrne. 2014. “The meanings of chimpanzee gestures.”
Current Biology 24 (14): 1596–1600.
Hobaiter, Catherine, Richard W. Byrne, and Klaus Zuberbühler. 2017. “Wild chimpanzees’
use of single and combined vocal and gestural signals.” Behavioral Ecology and Socio-
biology 71 (6): 96.
Hobaiter, Catherine, Kirsty E. Graham, and Richard W. Byrne. 2022. “Are ape gestures like
words? Outstanding issues in detecting similarities and differences between human lan-
guage and ape gesture.” Philosophical Transactions of the Royal Society B 377: 20210301.
Huang, Mingpan, Haigang Ma, Changyong Ma, Paul A. Garber, and Pengfei Fan. 2020.
“Male gibbon loud morning calls conform to Zipf’s law of brevity and Menzerath’s law:
Insights into the origin of human language.” Animal Behaviour 160: 145–155.
Kamiloğlu, Roza G., Katie E. Slocombe, Daniel, B.M. Haun, and Disa A. Sauter. 2020. “Hu-
man listeners’ perception of behavioural context and core affect dimensions in chimpan-
zee vocalizations.” Proceedings of the Royal Society: Biological Sciences. doi:10.1098/
rspb.2020.1148.
Kershenbaum, Arik, Holly Root-Gutteridge, Bilal Habib, Janice Koler-Matznick, Brian
Mitchell, Vicente Palacios, and Sara Waller. 2016. “Disentangling canid howls across
multiple speices and subspecies: Structure in a complex communication channel.” Be-
havioural Processes 124: 149–157.
Kuhl, Patricia K. 1979. “The perception of speech in early infancy.” Speech and Language
1: 1–47.
Kuhl, Patricia K., and James D. Miller. 1975. “Speech perception by the chinchilla: Voiced-
voiceless distinction in alveolar plosive consonants.” Science 190 (4209): 69–72.
 Intentional Signals, Their Meanings in Non-Human Communication 31

Leavens, David A., Kim A. Bard, and William D. Hopkins. 2017. “The mismeasure of ape
social cognition.” Animal Cognition. doi:10.1007/s10071-017-1119-1.
Leavens, David A., William D. Hopkins, and Kim A. Bard. 1996. “Indexical and referential
pointing in chimpanzees (Pan troglodytes).” Journal of Comparative Psychology 110
(4): 346–353.
Leavens, David A., Jamie L. Russell, and William D. Hopkins. 2005. “Intentionality as
measured in the persistence and elaboration of communication by chimpanzees (Pan
troglodytes).” Child Development 76 (1): 291–306.
Liddle, Bethany, and Daniel Nettle. 2006. “Higher-order theory of mind and social competence
in school-age children.” Journal of Cultural and Evolutionary Psychology 4 (3–4): 231–244.
Liebal, Katja, Josep Call, and Michael Tomasello. 2004. “Use of gesture sequences in chim-
panzees.” American Journal of Primatology 64 (4): 377–396.
Lupyan, Gary, and Rick Dale. 2010. “Language structure is partly determined by social
structure.” pLoS One 5 (1): e38559.
McCarthy, Maureen S., Mary Lee Abshire Jensvold, and Deborah H. Fouts. 2013. “Use of
gesture sequences in captive chimpanzee (pan troglodytes) play.” Animal Cognition 16
(3): 471–481.
McGurk, Harry, and John MacDonald. 1976. “Hearing lips and seeing voices.” Science
5588: 746–748.
Moore, Richard. 2014. “Ape gestures: Interpreting chimpanzee and bonobo minds.” Current
Biology 24 (14): R645–R647.
Moore, Richard. 2016. “Meaning and ostension in great ape gestural communication.” Ani-
mal Cognition 19 (1): 223–231.
Moore, Richard. 2017. “Social cognition, stag hunts, and the evolution of language.” Biol-
ogy & Philosophy 32 (6): 797–818.
Nemeth, Erwin, and Henrik Brumm. 2010. “Birds and anthropogenic noise: Are urban songs
adaptive?” The American Naturalist 176 (4): 465–475.
Ouattara, Karim, Alban Lemasson, and Klaus Zuberbühler. 2009. “Campbell’s monkeys use
affixation to alter call meaning.” pLoS One 4 (11): e7808.
Pika, Simone, Katja Liebal, and Michael Tomasello. 2003. “Gestural communication in
young gorillas (Gorilla gorilla): Gestural repertoire, learning, and use.” American Jour-
nal of Primatology 60 (3): 95–111.
Pika, Simone, Katja Liebal, and Michael Tomasello. 2005. “Gestural communication in subadult
bonobos (Pan paniscus): Repertoire and use.” American Journal of Primatology 65 (1): 39–61.
Plooij, Frans X. 1978. “Some basic traits of language in wild chimpanzees?” In Action,
Gestures, and Symbol: The Emergence of Language, edited by Andrew Lock, 111–132.
Pollick, Amy S., and Frans B.M. de Waal. 2007. “Ape gestures and language evolution.”
Proceedings of the National Academy of Sciences 104 (19): 8184–8189.
Provine, Robert R. 1992. “Contagious laughter: Laughter is a sufficient stimulus for laughs
and smiles.” Bulletin of the Psychonomic Society 30 (1): 1–4.
Rendall, Drew, Michael J. Owren, and Michael J. Ryan. 2009. “What do animal signals
mean?” Animal Behaviour 78 (2): 233–240.
Safryghin, Alexandra, Catharine Cross, Brittany Fallon, Raphaela Heesen, Ramon Ferrer-i-Can-
cho, and Catherine Hobaiter. 2022. “Variable expression of linguistic laws in ape gesture:
A case study from chimpanzee sexual solicitation.” Royal Society Open Science 9: Online.
Schel, Anne Marijke, Zarin Machanda, Simon W. Townsend, Klaus Zuberbühler, and Katie
E. Slocombe. 2013. “Chimpanzee food calls are directed at specific individuals.” Animal
Behaviour 86 (5): 955–965.
32 Catherine Hobaiter, Adriano R. Lameira, and Derek Ball

Schel, Anne Marijke, Simon W. Townsend, Zarin Machanda, Klaus Zuberbühler, and Katie
E. Slocombe. 2013. “Chimpanzee alarm call production meets key criteria for intention-
ality.” PLoS One 8 (10): e76674.
Scott-Phillips, Thomas C. 2015. “Meaning in animal and human communication.” Animal
Cognition 18 (3): 801–805.
Senghas, Ann, Sotaro Kita, and Asli Özyürek. 2004. “Children creating core properties of language:
Evidence from an emerging sign language in nicaragua.” Science 305 (5691): 1779–1782.
Seyfarth, Robert M., and Dorothy L. Cheney. 2003. “Signalers and receivers in animal com-
munication.” Annual Review of Psychology 54 (1): 145–173.
Slabbekoorn, Hans, and Margriet Peet. 2003. “Ecology: Birds sing at a higher pitch in urban
noise.” Nature 424 (6946): 267.
Slocombe, Katie E., and Klaus Zuberbühler. 2007. “Chimpanzees modify recruitment
screams as a function of audience composition.” Proceedings of the National Academy
of Sciences 104 (43): 17228–17233.
Tanner, Joanne E., and Richard W. Byrne. 1996. “Representation of action through iconic
gesture in a captive lowland gorilla.” Current Anthropology 37 (1): 162–173.
Templeton, Christopher, N., Erick Greene, and Kate Davis. 2005. “Allometry of alarm calls:
Black-capped chickadees encode information about predator size.” Science 308: 1934–1937.
Tomasello, Michael, Barbara L. George, Ann Cale Kruger, and Andrea Evans Farrar. 1985.
“The development of gestural communication in young chimpanzees.” Journal of Hu-
man Evolution 14: 175–186.
Townsend, Simon W., Sonja E. Koski, Richard W. Byrne, Katie E. Slocombe, Balthasar
Bickel, Markus Boeckle, Ines Braga Goncalves, et al. 2017. “Exorcising Grice’s ghost:
An empirical approach to studying intentional communication in animals: Intentional
communication in animals.” Biological Reviews 92 (3): 1427–1433.
Vail, Alexander L., Andrea Manica, and Redouan Bshary. 2013. “Referential gestures in fish
collaborative hunting.” Nature Communications 4 (1).
Wich, Serge A., Michael Krützen, Adriano R. Lameira, Alexander Nater, Natasha Arora,
Meredith L. Bastian, Ellen Meulman, et al. 2012. “Call cultures in orang-utans?” pLoS
One 7 (5): e36180.
Wich, Serge A., and Han de Vries. 2006. “Male monkeys remember which group members
have given alarm calls.” Proceedings of the Royal Society of London B: Biological Sci-
ences 273 (1587): 735–740.
Wild, Barbara, Michael Erb, Michael Eyb, Mathias Bartels, and Wolfgang Grodd. 2003.
“Why are smiles contagious? An FMRI study of the interaction between perception of
facial affect and facial movements.” Psychiatry Research: Neuroimaging 123 (1): 17–36.
Wilke, Claudia, Eithne Kavanagh, Ed Donnellan, Bridget M. Waller, Zarin P. Machanda, and
Katie E. Slocombe. 2017. “Production of and responses to unimodal and multimodal signals
in wild chimpanzees, Pan Troglodytes Schweinfurthii.” Animal Behaviour 123: 305–316.
Wimmer, Heinz, and Josef Perner. 1983. “Beliefs about beliefs: Representation and con-
straining function of wrong beliefs in young children’s understanding of deception.”
Cognition 13 (1): 103–128.
Zipf, George Kingsley. 1949. Human Behaviour and the Principle of Least-Effort. Cam-
bridge, MA: Addison Wesley.
Zuberbühler, Klaus, Dorothy L. Cheney, and Robert M. Seyfarth. 1999. “Conceptual se-
mantics in a nonhuman primate.” Journal of Comparative Psychology 113 (1): 33–42.
4
GETTING OUT OF OUR SKIN
What Decoding Interspecies Communication
and Nonhuman Intelligence Can Tell Us About
Deciphering Alien Languages

Denise L. Herzing

Introduction
Although researchers have studied intraspecific communication signals in many
nonhuman taxa and across different sensory systems, interspecies communication
remains largely uncharted territory. Organisms across taxa often live in close prox-
imity to each other and may learn to hear and interpret signals from their allospe-
cific neighbors. In some cases, decoding signals can aid in survival. Many species
use the calls and signals of nearby species to their survival advantage (Munn 1986).
Some species take advantage of their neighbors’ monitoring abilities to learn the
meaning of appropriate alarm calls (birds, primates). In other cases, it can help one
species interact socially with another species. Dolphins in the wild have a history
of interaction, both competitive and cooperative, with other cetaceans. Atlantic
spotted dolphins in the Bahamas interact with bottlenose dolphins (Tursiops trun-
catus) on a regular and intimate basis (Herzing and Johnson 1997).
Such interactions have also occurred throughout the nonhuman world, and also
between humans and nonhumans. Domestic animals, in some cases our closest
working partners, have tuned into human signals for decades. For example, McCo-
nnell (1990) describes specific cross-cultural signals work between domestic dogs
and their owners. Some universal rules across species have been described for
mammals and birds (Morton 1977) but have yet to be explored adequately. Even
wild animals, such as dolphins, have a history interaction with humans in the wild
(Pryor and Lindberg 1990).
One place to begin exploring the mechanisms for interpreting alien signals
may reside in this area of research. Here, many species have either incorporated
allospecific signals into their own interpretative systems or have developed
new communication repertoires as a way to interpret and transmit information.

DOI: 10.4324/9781003352174-4
34 Denise L. Herzing

Passive decoding or active interactions can both be present during interspe-

cific communication. How far taxa can be nonrelated and physically differ-
ent can vary, and their success in cross-species communication may depend
on similarities between both their signals and their sensory systems. Complex
communication can include one or more sensory modalities, including visuals,
tactile, chemical, and others. Humans have focused on acoustic communication
primarily because of our understanding of the complexity of our own acous-
tic linguistic language. However, it may be worth some investigation to look
for other encoded information in these nonhuman interspecific communication
systems.

Deciphering Nonhuman Communication Signals

How do nonrelated species understand each other? Some examples from the wild
suggest that although listening and deciphering may occur with signals such as
acoustic alarm calls, studies may also suggest that it is more efficient to create a
new communication system with another species than to learn the intricacies of the
others. For example, two examples in the dolphin world illustrate the strategy of
creating shared/mutual calls during interaction. Resident killer whale pods in the
Pacific Northwest have unique pod dialects. When pods interact, they use a small
repertoire of shared calls (Ford 1991). There have been recent reports of complex
dynamics of vocalization used between two species of sympatric dolphins in Costa
Rica, suggesting that they alter aspects of their calls when together and revert back
to their own species’ calls when apart (May-Collado 2010). Creating a shared sys-
tem of information has also proven successful in experimental settings for bonobo
chimpanzees (Savage-Rumbaugh et al. 1986), African grey parrots (Pepperberg
1986) and, to a limited degree, dolphins (Delfour and Marten 2005; Herzing et al.
2012; Xitco et al. 2001).
Past studies have focused on acoustic communication, but even in the analysis
of complex human language, context and information are distributed and interac-
tive (Johnson 2001, Forster 2002), and the interplay of multi-modal signals and
social dynamics is the essence of complex information. As we can see, acoustic
channels of communication have been the focus of most linguistic studies. This
is primarily due to our bias, as humans, to both the acoustic and visual chan-
nels of communication. Arik Kershenbaum (in this volume [Chapter 2]) takes
a look at less likely sensory modalities that we have thought might be available
for linguistic communication. Although chemical, electric, and magnetic senses
are not usually discussion in the depth that acoustic channels are, these alterna-
tive sensory systems remain a viable option for exploring alien communication.
The ability to identify appropriate sensory channels for communication, along
with frameworks that include species-specific and within-species signals and
rules, will be critical to assess alien communication signals, whether on Earth
or elsewhere.
 Getting Out of Our Skin 35

Categorizing Signals
Classifying communication signals to determine the natural boundaries of signal
units has been for a variety of taxa (e.g., Ehret 1992; Slobodchikoff et al. 1991;
Marler 1982; May et al. 1989; Seyfarth et al. 1980). Whether nonhuman animals
use a referential or graded systems of communication is still unknown. Although it
is likely that most nonhuman animals use a graded system, referential signals are
known only to a few taxa; dolphin signature whistles (e.g., Caldwell et al. 1990;
Smolker et al. 1993), vervet monkey alarm calls (Seyfarth et al. 1980, Seyfarth and
Cheney 1993), ground squirrels (Robinson 1981), and prairie dogs (Slobodchikoff
et al. 1991).
Cutting-edge computer techniques (e.g., Kohlsdorf et al. 2014; Kershenbaum
et al. 2013) have recently been applied to nonhuman acoustic signals. Keeping
in mind that referential and graded communication are not necessarily mutually
exclusive, such techniques may be valuable for categorization. Other possible out-
comes of machine learning/artificial intelligence (AI) might include the discovery
of smallest or fundamental units of an animal’s repertoire, unhindered by human
biases.

The Importance of Metadata for Interpretation

What are metadata, and how are they important? In most animal studies, metadata
may include things like age, sex, relationships, life history, and other parameters.
They are often used as dependent variables during analysis. However, such meta-
data plays a large role in the interpretation and function of signal use in many
studies.
In my own work with dolphins, such metadata is overlaid during different devel-
opmental periods, in relation to quantifiable personality traits, etc.
Another important component of interpretation and meta-analysis lies in the
prosodic features of communication. These features of communication (frequency,
duration, amplitude, relative signals, rhythm, synchrony, etc.) should not be over-
looked for their importance (Herzing 2015). If there are any universals across
species, it may be found in these types of modulation, and ascribing meaning to
communication signals is critical to any meaningful study of nonhuman signals.

Observational and Experimental Verifcation of Meaning

For a complete understanding of communication signals and their meaning, we
may have to turn to other techniques based in the cognitive realm. Distributed
cognition, which suggests that cognition occurs not just within an individual mind,
but also between individuals (Johnson 2001; Forster 2002), is one example. In this
framework, measurable behaviors are considered the “media” that are exchanged
between individuals. Because such interactions can be recorded (e.g., behavior),
36 Denise L. Herzing

they become measurable phenomena, unlike mental states and concepts like “inten-
tion” that are difficult to assess. Such techniques applied to both nonhuman species
and our interaction with them may further illuminate the communication process,
critical for any alien signal interpretation.
Researchers have also used experimental tests to study both behavioral and cog-
nitive flexibility. Experiments in laboratories with a variety of taxa have shown that
many animals – including dolphins – can understand word order (syntax), word
meaning (semantics), and abstract thought, and show self-awareness (Herman et al.
1990; Pack and Herman 1995; Marino et al. 1994; Delfour and Marten 2001), com-
mon and pygmy chimpanzees (Savage-Rumbaugh et al. 1986) bottlenose dolphins
(Herman et al. 1984), and African grey parrots (Pepperberg 1986). Some of these
experimental techniques have also involved the creation of a technological bridge
that allows and enhances communication between species (Herzing 2016).

Conclusions
Interspecific communication analysis is a viable avenue to explore and exercise
our abilities to categorize and interpret potential future alien communications.
Scientists have new machine learning techniques that can expedite data mining.
However, metadata still remain an important element in the interpretation of com-
munication signals. Understanding a species sensory system and it species-specific
cultures and social systems will enhance these interpretations. However, such
information is likely to be unavailable – or unrecognizable – in the reception of
alien communications, and experimental and cognitive-based tests will not be an
option for signals received remotely. Imagining different sensory, perceptual, and
social systems when interpreting alien communication signals will be critical for
our understanding. Utilizing Earth’s great repertoire of other species communica-
tion signals in an exercise to extract and decode information would be a worthy
endeavor in preparation for non-terrestrial signal interpretation in the future.

References
Caldwell, Melba C., David K. Caldwell, and Peter L. Tyack. 1990. “Review of the signa-
ture whistle hypothesis for the Atlantic bottlenose dolphin.” In The Bottlenose Dolphin,
edited by Stephen Leatherwood and Randall R. Reeves, 199–233. San Diego: Academic
Press.
Delfour, Fabienne, and Ken Marten. 2001. “Mirror image processing in three marine mam-
mal species: killer whales (Orca orcinus), false killer whales (Pseudorca crassidens)
and California sea lions (Zalophus californianus).” Behavioral Processes 53: 181–190.
Delfour, Fabienne, and Ken Marten 2005. “Inter-modal learning task in bottlenosed dol-
phins (Tursiops truncatus): A preliminary study showed that social factors might influ-
ence learning strategies.” Acta Ethologica 8: 57–64.
Ehret, Günter. 1992. “Categorical perception of mouse-pup ultrasounds in the temporal do-
main.” Animal Behaviour 43: 409–416.
 Getting Out of Our Skin 37

Ford, John K.B. 1991. “Vocal traditions among resident killer whales (Orcinus orca) in
coastal waters of British Columbia.” Canadian Journal of Zoology 69: 1454–1483.
Forster, Debbie. 2002. “Consort turnovers as distributed cognition in olive baboons: A sys-
tems approach to mind.” In The Cognitive Animal: Empirical and Theoretical Perspec-
tives on Animal Cognition, edited by Marc Bekoff, Collin Allen, and Gordon Burghardt,
163–172. Cambridge: MIT Press.
Herman, Louis M., Palmer Morrel-Samuels, and Adam A. Pack. 1990. “Bottlenosed dolphin
and human recognition of veridical and degraded video displays of an artificial gestural
language.” Journal of Experimental Psychology 119: 215–230.
Herman, Louis M., Douglas G. Richards, and James P. Wolz 1984. “Comprehension of sen-
tences by bottlenosed dolphins.” Cognition 16: 129–219.
Herzing, Denise L. 2015. “Synchronous and rhythmic vocalizations and correlated under-
water behavior of free-ranging Atlantic spotted dolphins (Stenella frontalis) and Bot-
tlenose dolphins (Tursiops truncatus) in the Bahamas.” Animal Behavior and Cognition
2 (1): 14–30.
Herzing, Denise L. 2016. “Interfaces & keyboards for human/dolphin communication: What
have we learned?” Animal Behavior and Cognition 3 (4): 243–254.
truncatu Herzing, Denise L., Fabienne Delfour, and Adam A. Pack. 2012. “Responses of
human-habituated wild Atlantic spotted dolphins to play behaviors using a two-way hu-
man/dolphin interface.” International Journal of Comparative Psychology 25: 137–165.
Herzing, Denise L., and Christine M. Johnson. 1997. “Interspecific interactions between At-
lantic spotted dolphins (Stenella frontalis) and bottlenose dolphins (Tursiops truncatus)
in the Bahamas, 1985–1995.” Aquatic Mammals 23: 85–99.
Johnson, Christine M. 2001. “Distributed primate cognition: A review.” Animal Cognition
4: 167–183.
Kershenbaum, Ari, Layla S. Sayigh, and Vincent M. Janik. 2013. “The encoding of indi-
vidual identity in dolphin signature whistles: How much information is needed?” PLoS
One. doi:10.1371/journal.pone.0077671.
Kohlsdorf, Daniel, Celeste Mason, Denise Herzing, and Thad Starner. 2014. “Probabilistic
extraction and discovery of fundamental units in dolphin whistles.” Acoustics, Speech
and Signal Processing (ICASSP), 2014 IEEE International Conference, 8242–8246.
IEEE, Piscataway, New Jersey.
Marino, Lori, Diana Reiss, and Gordon G. Gallup, Jr. 1994. “Mirror self-recognition in bot-
tlenose dolphins: Implications for comparative investigations of highly dissimilar spe-
cies.” In Self-Awareness in Animals and Humans: Developmental Perspectives, edited
by S. Taylor Parker, Robert W. Mitchell, and Maria L. Boccia, 380–391. New York:
Cambridge University Press.
Marler, Peter R. 1982. “Avian and primate communication: the problem of natural catego-
ries.” Neuroscience and Biobehavior 6: 87–92.
May, Brad, David B. Moody, and William C. Stebbins. 1989. “Categorical perception of
conspecific communication sounds by Japanese macaques, Macaca fuscata.” Journal of
the Acoustical Society of America 85: 837–847.
May-Collado, Laura. 2010. “Changes in whistle structure of two dolphin species during
interspecific associations.” Ethology 116: 1065–1074.
McConnell, Patricia B. 1990. “Acoustic structure and receiver response in domestic dogs,
Canis familiaris.” Animal Behaviour 39: 887–904.
Morton, Eugene S. 1977. “On the occurrence and significance of motivation: Structural rules
in some bird and mammal sounds.” American Naturalist 111: 855–869.
38 Denise L. Herzing

Munn, Charles A. 1986. “The deceptive use of alarm calls by sentinel species in mixed-
species flocks of neotropical birds.” In Deception: Perspectives on Human and Nonhu-
man Deceit, edited by Robert W. Mitchell and Nicholas S. Thompson, 169–176. Albany:
New York State University Press.
Pack, Adam A., and Louis M. Herman. 1995. “Sensory integration in the bottlenose dolphin:
Immediate recognition of complex shapes across the sense of echolocation and vision.”
Journal of the Acoustical Society of America 98: 722–733.
Pepperberg, Irene M. 1986. “Acquisition of anomalous communicatory systems: Implica-
tions for studies on interspecies communication.” In Dolphin Cognition and Behavior: A
Comparative Approach, edited by Ron. J. Schusterman, Jeanette A. Thomas, and Forrest
G. Wood, 289–302. Hillsdale: Lawrence Erlbaum Associates.
Pryor, Karen, and Jon Lindberg. 1990. “A dolphin-human fishing cooperative in Brazil.”
Marine Mammal Science 6: 77–82.
Robinson, Scott R. 1981. “Alarm communication in Belding’s ground squirrels.” Zier
Tierpsychologie 56: 150–168.
Savage-Rumbaugh, Sue, Kelly McDonald, Rose A. Sevcik, William D. Hopkins, and Eliza-
beth Rupert. 1986. “Spontaneous symbol acquisition and communicative use by pygmy
chimpanzees [Pan paniscus].” Journal of Experimental Psychology 112: 211–235.
Seyfarth, Robert M., and Dorothy L. Cheney. 1993. “Meaning, reference, and intentionality
in the natural vocalizations of monkeys.” In Language and Communication: Compara-
tive Perspectives, edited by Herbert R. Roitblat, Louis M. Herman, and Paul Nachtigall,
195–219. Hillside: Erlbaum.
Seyfarth, Robert M., Dorothy L. Cheney, and Peter Marler. 1980. “Vervet monkey alarm
calls: Semantic communication in a free-ranging primate.” Animal Behaviour 28:
1070–1094.
Slobodchikoff, Con N., Judith Kiriazis, C. Fischer, and E. Creef. 1991. “Semantic informa-
tion distinguishing individual predators in the alarm calls of Gunnison’s prairie dog.”
Animal Behaviour 42: 713–719.
Smolker, Rachel A., Janet Mann, and Barbara B. Smuts. 1993. “Use of signature whistles
during separations and reunions by wild bottlenose dolphin mothers and infants.” Behav-
ioural and Ecological Sociobiology 33: 393–402.
Xitco, Jr. Mark J., John D. Gory, and Stan A. Kuczaj II. 2001. “Spontaneous pointing by
bottlenose dolphins (Tursiops truncatus).” Animal Cognition 4: 115–123.
5
COMMUNICATIVE RESOURCES
BEYOND THE VERBAL TIER
A View on Xenolinguistics From Interactional
Linguistics

Heike Ortner

Introduction
When we try to anticipate contact with extraterrestrial intelligence (ETI), we
speculate about communication systems other than human language. We develop
assumptions concerning the grammatical structures of extraterrestrial languages
(cf. Roberts et al., this volume [Chapter 15]), the types of signals that might be used
for transmission (cf. Herzing, this volume [Chapter 4], Granger et al., this volume
[Chapter 8]), and the impact of what is almost certain to be different cognitive
abilities (cf. Samuels and Punske, this volume [Chapter 16]). In this chapter, I will
reflect on the corporality of our terrestrial existence and its impact on our everyday
interactions, drawing from the theoretical background of interactional linguistics.
This conception was prompted by two experiences I had while I was preparing a
project about multimodal instructions in physical therapy at a clinic for neuroreha-
bilitation. There were two patients who I remember vividly.
The first was an elderly woman with global aphasia, which means that she
could neither fully understand the therapist’s utterances nor produce intelligible
speech. However, she was fairly well able to communicate with the therapist by
attentively watching him, mainly interpreting his facial expressions, gestures,
and the pitch of his voice (Goodwin 2000, 2003 for detailed analyses of interac-
tions with aphasic patients). With sufficient communicative success, she replied
by gaze, body posture, and gestural hand and head movements. She relied on
these clues so much that the therapist deliberately toned down his para-verbal and
non-verbal behavior,1 forcing her to train her remaining productive and receptive
linguistic competence.
The second patient was a young man with locked-in syndrome, which left him
unable to move or talk, while his mental abilities were probably undamaged. In the

DOI: 10.4324/9781003352174-5
40 Heike Ortner

session that I observed, the speech therapist concentrated on teaching the patient
how to swallow food on cue. For this patient, communicating via eye movement
was not the solution. His interactions and the shared meaning-making with his fel-
low human beings were diminished not only by his lack of a voice but also by his
almost fully paralyzed body. The therapist was nonetheless able to read the current
mood and compliance of the patient and got him to work with her on the task at
hand.
These two extraordinary interactions lead me to the topic of this chapter: the
multimodality of human interaction and its implications for xenolinguistics. Some
current trends in humanities and social studies indicate a heightened interest in
the human body and its functions in meaning-making – so much so that there is
talk of an “embodied turn” (Nevile 2015). Interactional linguistics is a relatively
new research program dedicated to the practices of shared meaning-making with
all available communicative resources in face-to-face interaction: verbal utter-
ances, voice, gaze, body postures and body orientation, positioning in space, facial
expression, object manipulation, gesture and other body movements, or generally
any kind of action (Couper-Kuhlen and Selting 2018: 9, 12).
If we ever get to establish immediate contact or at least interpret video signals
of visible bodily behavior, this aspect of the interaction is going to be of utmost
importance, yet also of utmost extraneousness. This chapter gives an overview of
some key terms and topics in interactional linguistics and their significance for
speculation on extraterrestrial languages, xenolinguistics, and messages to extra-
terrestrial intelligence (METI). Some helpful insights from studies on human–ani-
mal interaction will be included, as well.

Multimodality in Interaction
Interactional linguistics is rooted in functional theories (Bühler 1978 [1934]; Hal-
liday 1973), constructionist and ethnomethodologist traditions (Garfinkel 1967),
sociology (Goffman 1983), linguistic anthropology (Goodwin 1981), and conver-
sation analysis (Sacks et al. 1974; Heath 1986) (see Couper-Kuhlen and Selt-
ing 2018; Streeck et al. 2011 for a historical overview). Conversation analysis
(CA) was originally mainly concerned with verbal processes such as turn-taking,
repairs, and the specifics of linguistic constructions in speech (Sacks et al. 1974;
Goodwin 1981). While facial expressions and gestures were recognized as inte-
gral to conversation early on, there was still a strong focus on the “verbal tier”2 of
conversation. Catalyzed by the advancements in recording and transcribing visual
data, interactional linguistics became a popular research program. By these tech-
nological means, it is possible to analyze all facets of interaction in more detail:
sequential and collaborative processes in several modes at once and their coordi-
nation in social interaction, and technology-mediated interaction or resources in
the environment, such as the use of tools (Deppermann and Schmitt 2007; Had-
dington et al. 2013).
 Communicative Resources Beyond the Verbal Tier 41

Social interaction is not chaotic: By the strictly empirical and qualitative

sequential micro-level analysis of natural data, it is possible to find systematic pat-
terns in the use of “verbal, vocal and visible resources in the service of performing
actions and activities” (Couper-Kuhlen and Selting 2018: 9). One of the key terms
is multimodality; it expands “the social interpretation of language and its meanings
to the whole range of representational and communicational modes or semiotic
resources for making meaning with employed in culture – such as image, writing,
gesture, gaze, speech, posture” (Jewitt 2014: 1). I would like to stress the aspect of
culture in this definition to make clear that multimodal signs may have iconic and
indexical properties, yet their forms (e.g., appropriate amplitude) and pragmatic
functions in interaction are learned and culturally diverse. For CETI (communica-
tion with extraterrestrial intelligence) and METI, this means that we should have
little hope of easily understanding bodily expressions like gesture or facial expres-
sions of ETI, even if their outer morphology were similar to ours. After all, the
sensorimotor system of our body is an integral part of the Chomskyan “faculty of
language” (cf. Samuels and Punske, this volume [Chapter 16]). Growing research
in interactional linguistics aims at formulating a grammar of interactional displays
and their coordination.
Beyond linear sequentiality and systematicity, another guiding principle of inter-
actional research is the notion of joint meaning-making: Participants in interaction
constantly negotiate and co-construct meaning in cooperation; for example, in pro-
cesses of multimodal identity construction (Norris 2011), in narratives (Quasthoff
and Becker 2005), and in cross-cultural settings. Summing up the interactional
linguistics program, Couper-Kuhlen and Selting (2018: 551) name several design
features of language that are very different from a structuralist approach:

[Language is] a dynamic process rather than a static entity; it is physically

embodied and publicly displayed; it delivers actions; its resources are organized
as paradigmatic alternatives for specific positions in the syntagma of turns and
sequences; it is temporally emergent, allowing for projection, extension, and
repair; it is shared and continuously provides for co-participation.

Preconditions of Interaction With ETI

First, I would like to distinguish three types of interaction that have very different
interactional consequences: 1) direct face-to-face interaction, or being at the same
place at the same time; 2) technology-mediated communication, e.g., through a
screen, but still in a way “face-to-face” by partly visible displays of communica-
tive resources; and 3) using only abstract symbols without any direct perception
of “the other.” Henceforth, I will deal mainly with the first notion (Granger et al.,
this volume [Chapter 8], thoroughly discuss the case of “hyper-asynchronicity” of
contact).
42 Heike Ortner

What are the preconditions of interacting with ETI in a meaningful, coopera-

tive way? The human visual and auditory systems and our cognitive abilities such
as attention, memory, and judgment are highly developed (Evans 2001). To us,
perceiving an object as moving is the basis of perceiving it as living and animated.
The perception of a “dynamic congruency of affectivity and movement” – moving
toward something or moving away from it as an index of emotion – is the basis
of any relationship we have with other life forms (Sheets-Johnstone 2009: 376,
emphasis in original). However, a working visual sensory apparatus is not neces-
sary for bodily interaction (cf. Herzing, this volume [Chapter 4]). Children who
were born blind, for example, produce gestures at the same rate as their sighted
counterparts, but use more body-focused types and rather refer to objects and per-
sons that are near them. All in all, they use more auditory gestures and cues and
other modalities (Avital and Streeck 2011: 170, 180; cf. also Wells-Jensen, this
volume [Chapter 13], on the abilities of blind persons and their implications for
ETI). There are many resources that could be central to the interaction of ETI:
touch, colors, scents, and so on.
To give an example of interactional argumentation, let us look at an element that
is thought to be universal in all human languages, according to Dingemanse et al.
(2013): the interjection “huh?” The universality of this element, according to them,
is not due to an innate mechanism or a genetic given but to a process of “conver-
gent cultural evolution”: The “huh?” simply has the perfect articulatory syllabic
structure in any language to account for a swift initiation of a repair. Of course, it is
highly unlikely that ETI will repeat their message because we say “huh?” to them,
just as it is unlikely that our gestures and facial expressions will be meaningful to
them. When thinking about future applications of CETI and METI, we ought to
develop strategies to complete communicative tasks such as initiating repair crea-
tively, starting from what we know so far.
This is because there are further important preconditions for mutual understand-
ing (additionally to those discussed in depth by Ross, this volume [Chapter 12]).
There has to be some common ground or a starting point for intersubjectivity (see
what follows). Even more basically, we need to recognize relevant displays in
the different modes of interaction (Deppermann and Schmitt 2007: 35) – in other
words, the ability and opportunity to isolate functional units. There has to be some
kind of systematicity of cues. Fuchs and De Jaegher (2009: 471, emphasis in origi-
nal following Lyons-Ruth et al. 1998) speak of “implicit relational knowing” – in
CETI and METI, the implicit aspects are not clear-cut from the start. Perceivable
changes in the different modes of interaction (e.g., gaze, gesture, facial expression)
are the basic unit of interactional analysis. The proposed co-construction of mean-
ing implies that a communicative signal is “what the participant treats as a signal”
(Norris 2011: 49) – but this depends a lot on our communicative experiences and is
therefore deeply rooted in cultural and social structures (Ross, this volume [Chap-
ter 12], also highlights the difficulty of cultural understanding, especially with little
context given).
 Communicative Resources Beyond the Verbal Tier 43

For thinking about CETI and METI, we need to ask the following questions.
What constitutes a change in a communicative resource – to us and to ETI? How
will we recognize the units of interactional displays, and how will we give cues
on how to disassemble our signals within one single mode or in different modes?
Human–robot interaction faces a similar dilemma that is solved by orientation
toward human habits and practices. For example, a robot in a museum can be pro-
grammed to pause and restart at a certain rate to guide the gaze of visitors, acting
on the assumption that gazing at something equals attention (Kuzuoka et al. 2008).
In general, the idea of “Gestalt” is very important to interaction (Mondada 2014:
37), meaning that our perception of a functional display depends on whether we see
a figure that becomes eye-catching before the ground.
Incentives from human–animal interaction (HAI) can help us better understand
the challenges of CETI. I will not engage in a discussion of the linguistic compe-
tence of animals (see Pearson 2013 for a comprehensive overview). It is problem-
atic to talk about animal signals as conveying meaning, transmitting information,
or being representational. Horisk and Cocroft (2013), for example, argue against
any representational notion to get away from any speculation on the existence of
a theory of mind in animals. They propose to discuss signals of animals from the
viewpoint of their influence on other animals and humans and the responses. Actu-
ally, this is a very interactional linguistics’ type of reasoning: The aboutness of a
signal is not representational, but co-constructed in interaction. Whatever we inter-
pret as a communicative sign of an animal conveying meaning is shared meaning
in interaction, whether or not the interpretation is “accurate.” In this sense, inten-
tionality is less important than interpretation (but cf. Hobaiter et al., this volume
[Chapter 3], for more profound thoughts on the interplay between intentionality,
cooperativity, meaning, behavior, and context from a Gricean perspective). There
is a lot of research on multimodal displays of animals. While smell and taste are
central to animal communication (chemical signaling), many animals can also
fine-tune their multimodal displays – for example, birds coordinating their dancing
choreography and their singing voices (Miles and Fuxjager 2018), gibbons adjust-
ing their facial expressions depending on social context (Scheider et al. 2016),
and cephalopods attuning their body postures and patterns of colors and spots to
express specific meanings, not to mention the astounding abilities of dolphins (Bal-
lesteros 2010: 114; cf. Pepperberg, this volume [Chapter 6] and Kershenbaum, this
volume [Chapter 2], for more detailed discussions on HAI, the interpretation of
different types of signals and the likeliness of ETI using specific modes other than
the acoustic one). There are even attempts to apply multimodal CA to the study of
animal communication, e.g., to determine vocal and gestural turn-taking of birds
and great apes (Fröhlich 2017). To me, studies of HAI show how we interpret ani-
mal signals according to our experiences with humans. In general, the behavior of
animals is interpreted as emotional, based on the notion that the affective neuronal
and hormonal mechanisms – as well as the social functions of affective displays –
can be compared (Watanabe and Kuczaj 2013; Kotrschal 2013; cf. also Pepperberg,
44 Heike Ortner

this volume [Chapter 6], on the functions of social interaction). On the other side,
our closest allies in the animal kingdom, dogs, can use gaze as a communicative
tool and comprehend many human bodily expressions, having an overall “sensi-
tivity to human stimuli” (Wynne et al. 2011: 106). We can only speculate on our
sensitivity for biologically more different life forms.

Intersubjectivity Is Key
Intersubjectivity is a key term in interactional linguistics. I will turn only to one of
its many meanings, described by Deppermann (2014: 65) as “intersubjectivity as
practical accomplishment.” It has a certain sequential organization: first, a partici-
pant in the interaction conducts an “action-be-understood”; second, the recipient
displays his or her understanding of the action; and third, the producer acknowl-
edges this understanding (Deppermann 2015: 66) – in CETI, of course, we might
need to find a different “flow” of intersubjective understanding.
Another interesting branch of research deals with processes of accommodation
and mimicry in interaction. The communication accommodation theory (CAT)
states that “Accommodation – as a process – refers to how interactants adjust their
communication so as to either diminish or enhance social and communicative dif-
ferences between them” (Giles 2009: 278); one function is “reducing uncertainties
about the other” (Giles 2008: 163). This can be done by verbal means such as word
choices, but also by interactive displays – for example smiling, nodding, and ges-
turing. According to Giles 2008), the main strategies are the following.

• Convergence: upward or downward adjustment to the other’s language or dia-

lect, rhythm, posture, and their interpretative abilities.
• Maintenance or non-accommodativeness: not shifting one’s language use.
• Divergence: accentuating one’s social identity by emphasizing differences.

Related concepts are recipient design, i.e., orientation toward co-participants

(Sacks et al. 1974), audience design; i.e., linguistic choices to seek approval of
the audience (Bell 2009); and behavioral mimicry, i.e., “the automatic imitation
of gestures, postures, mannerisms, and other motor movements” (Chartrand and
Lakin 2013: 285). Verbal aspects such as syntax may be subject to mimicry, as
well. For any of these concepts, the connection to interactional linguistics lies in
the notion that meaning “unfolds only in the context of an embodied contextual
configuration,” building “on prior interactional histories and on both shared and
private experiences” (Deppermann 2015: 96), so we are led by assumptions and
expectations regarding other participants at first, but we can quickly establish com-
mon ground (Clark 1996) molded by our interactive negotiations.
When trying to apply these concepts to communication with ETI, behavioral
mimicry seems be a good starting point for building rapport. Only after the estab-
lishment of a common “language” in any meaning of the world may we think
 Communicative Resources Beyond the Verbal Tier 45

about more complex strategies such as convergence, maintenance, and divergence

to communicate social meanings toward ETI.

The Intricate Temporality and Spatiality of the Human Condition

Multimodal resources are grounded in time and space (Mondada 2014: 37).
Regarding time, there are both successive and simultaneous temporal relationships
of multimodal displays in human–human interaction (Deppermann and Schmitt
2007); for example, we can combine a bodily demonstration with a verbal explana-
tion (Mondada 2014; Haddington et al. 2014). Regarding space, there are specific
practices on how to refer to physical as well as social positions of interactants
(Stukenbrock 2014; Haddington et al. 2013: 20). A very basic human action is
pointing to objects in the immediate spatio-temporal surrounding (Bühler 1978
[1934] calls this “demonstratio ad oculos”) but also to objects that are not present
(Bühler’s “deixis am phantasma”) (Stukenbrock 2014: 89). I would not call point-
ing a universal gesture, even though it is one of the earliest communicative means
infants make use of. Dogs that are socialized in a human environment and have
continuous experience with it can follow pointing gestures (Wynne et al. 2011).
We may assume that pointing at something can easily guide attention and inter-
relate objects or persons with terms and names, facilitating language learning or
more general understanding. The cognitive foundations and perceptions of time
and space may be very different in ETI, however. While we perceive interactions
as a sequence of utterances, the different layers of multimodal interaction are pro-
duced and processed simultaneously. How may we get a grasp on ETI’s “flow”?

Coordination and Projection

Multimodal resources are not separate layers but interwoven, flexible, and dynamic.
Their coordination is so complex that we only start to understand the processes of
human mutual calibration in interaction: on the one hand, intrapersonal coordina-
tion, the coordination of modes by one person; on the other hand, interpersonal
coordination, i.e., the coordination with multiple participants or the multiplicity of
modes of expression while interacting (Deppermann and Schmitt 2007: 35). Still,
we are very capable of coordinating our own bodily expressions and verbal utter-
ances, as well as ourselves with other participants in the interactional ensemble.
This is in part due to our ability to predict (in linguistic terms project) “what is likely
to come next” (Auer 2015: 28) – for example, there are syntactic restrictions in
verbal utterances that make it easy for us to complete a sentence before it is uttered.
Therefore, we can also co-construct utterances by taking over a construction.
It is more difficult to imagine how this process could work in CETI. Comple-
tions might not only turn out completely wrong because predictions of things to
come are more difficult, but what is absolutely normal in human interaction could
be interpreted as unfriendly interruption. Communication with ETI would probably
46 Heike Ortner

stay very explicit and careful for a long time. To cite an example from human–ani-
mal interactions, there seems to be a gradual adjustment between cattle and their
herders after long-lasting contact (Ellingsen et al. 2014). With this in mind, coordi-
nation between humans and ETI may come about slowly, but steadily.

The Body as a Resource of Meaning-Making

In interactional linguistics, “embodiment” refers to the bodily resources of com-
munication, the embodied experiences of making sense in real time (Nevile 2015;
Streeck et al. 2011). Sheets-Johnstone (2009) calls for a closer look at kinesthesia,
i.e., a sensory-kinetic understanding of the body, rather than a senorimotor defini-
tion of the body as the “place” of cognition.
The notion of intercorporeality was introduced by Merleau-Ponty (2005 [origi-
nally published in 1945]: 413–422) referring to the coordination of human bodies
in interaction as a “together-ness,” some sort of merge of object and subject. Fuchs
and De Jaegher (2009: 470) use the terms “mutual incorporation” and “coupling:”
Social understanding is achieved dynamically by the coupling of “two embod-
ied subjects” and their interactive coordination through voice, facial expressions,
touch, gesture, and other features. The other person’s body that we experience
through perception becomes incorporated into one’s own, so to say as an extension
of ourselves. This incorporation facilitates “participatory sense-making” (Fuchs
and De Jaegher 2009: 477), which means providing the participants with knowl-
edge, experience, and domains not available to the individual. This also implies
mutual affection; similarly, Trevarthen (1987: 194) defines meaning as “meaning
in relationships” and “interpersonal symbiosis.”
From research on infant–caregiver interactions, Fuchs and De Jaegher (2009:
482) draw the conclusion that mutual incorporation also allows us to get a hold
of another person’s experiences, even if they are not capable of verbal expression.
Whether this not only applies to infants, but also to ETI, is very questionable. A
pessimistic consequence of mutual incorporation might be that we are not able to
understand someone else’s body if this body and its perceptions are fundamentally
different from ours. Meyer and von Wedelstaedt (2017: 13) emphasize the impor-
tance of “tacit bodily coordination.” Unfortunately, anything that is deemed “tacit”
in human interaction is tricky for communication with ETI. All of our long-stand-
ing, acquired presumptions on interaction may mislead us when we have to inter-
pret a different mindset and its bodily expressions – or, to apply a more accurately
interactional attitude, when we strive toward understanding the relations between
thinking and body. There is a high risk of “sending the wrong message” or subtext
without bodies and voices (Harbour, this volume [Chapter 18]). However, as Berea
(this volume [Chapter 7]) points out, communication pathologies have the potential
to be transformative in the most helpful way when trying to find common ground
with ETI. I also second the remarks by Granger et al. (this volume [Chapter 8]) in
favor of a systematic, creative, and empathic approach devoid of judgmental good/
bad evaluations of ETI and premature assumptions.
 Communicative Resources Beyond the Verbal Tier 47

So far, my remarks mainly concerned a dyadic participation framework – yet

the term “multiactivity” gained a lot of attention in the last couple of years, i.e.,
paying reference to more complex participation frameworks, the coordination of
multiple bodies and activities, and temporal and sequential aspects of concurrent
activities (Haddington et al. 2014; Deppermann 2014; Mondada 2014). I can point
only to the increasing complexity of an interactional ensemble consisting of several
human beings and representatives of ETI that would probably be the more probable
case anyway. So, the participation framework does not have only two fundamen-
tally dissimilar parties (human–ETI) but calls for efficient communication between
interlocutors of the “same” communicative universe, as well – a feat not as trivial
as it may sound, given the vast cultural differences on Earth and the pitfalls of co-
constructing meaning in everyday interaction.

Conclusion
“Combine and repeat” is a central principle of language mentioned in several
contributions to this volume (cf. Roberts et al. [Chapter 15], Samuels and Punske
[Chapter 16], and Sperlich [Chapter 14]). It can also be a helpful strategy in find-
ing patterns in interactional displays of any mode, communicative resource, signal
type, and code. Summing up, I would like to repeat the following main challenges
we would face in IETI (interactions with extraterrestrial intelligence).

• Recognizing relevant displays and even systems of displays as codes and cues
for understanding – given the fact that corporeal displays are not iconic, it would
be extremely hard to translate them without some sort of Rosetta stone (cf. Slo-
bodchikoff, this volume [Chapter 9], on thoughts on the irreplaceability of a tool
like this).
• Getting a grasp on the temporal “flow” of ETI, taking into account the option
that spatial and temporal concepts and therefore coordination and projective
forces might be completely different.
• Making ourselves comprehensible, not taking our corporal design for granted.
• Trying to think with a different body as probably the biggest challenge.

Given these fascinating questions, it is a pity that direct “face-to-face” inter-

action with extraterrestrial intelligence is even less likely than other forms of
contact. I still think that it is stimulating to make this endeavor. Interactional
linguistics refrains from the thought of “body language” as being more honest,
an exhibition of the subconscious, or mainly constituted by iconic signs. The
most helpful implications of applying the methodology and terminology of inter-
actional linguistics on CETI is a warning: we will need to be careful to draw
deterministic conclusions on the basis of ETI’s cognitive make-up because of
their bodily experiences. After all, meaning is meaning in relationships, meaning
in interaction – we as humans, embedded in our own corporeality (!), are part of
the equation.
48 Heike Ortner

Notes
1 In interactional linguistics, the terms “para-verbal” and “non-verbal” are avoided be-
cause they imply the separation of codes that interactionist studies wish to overcome
(Streeck et al. 2011: 9). In the same vein, there is no mention of “body language”: the
body is fundamentally involved in the human language facility.
2 “Tier” refers to linguistic transcripts that dissect interactions into different tiers or lines,
e.g., speech (word-by-word transcription of utterances) and gesture.

References
Auer, Peter. 2015. “The temporality of language in interaction: Projection and latency.” In
Temporality in Interaction, edited by Arnulf Deppermann and Susanne Günthner, 27–56.
Amsterdam, Philadelphia: Benjamins.
Avital, Sharon, and Jürgen Streeck. 2011. “Terra incognita: Social interaction among blind
children.” In Embodied Interaction: Language and Body in the Material World, edited
by Jürgen Streeck, Charles Goodwin, and Curtis LeBaron, 169–181. Cambridge: Cam-
bridge University Press.
Ballesteros, Fernando J. 2010. E.T. Talk: How Will We Communicate with Intelligent Life on
Other Worlds? Astronomer’s Universe. New York: Springer.
Bell, Allan. 2009. “Language style as audience design.” In The New Sociolinguistics Reader,
edited by Nikolas Coupland and Adam Jaworski, 265–275. Basingstoke, New York: Pal-
grave Macmillan.
Bühler, Karl. 1978 [1934]. Sprachtheorie: Die Darstellungsfunktion der Sprache [The The-
ory of Lanuage: The Representational Function of Language], edited by Ungekürzte
Ausgabe. Ullstein-Buch. Frankfurt: Ullstein.
Chartrand, Tanya L., and Jessica L. Lakin. 2013. “The antecedents and consequences of hu-
man behavioral mimicry.” Annual Review of Psychology 64: 285–308.
Clark, Herbert H. 1996. Using Language. Cambridge: Cambridge University Press.
Couper-Kuhlen, Elizabeth, and Margret Selting. 2018. Interactional Linguistics: Studying
Language in Social Interaction. Cambridge: Cambridge University Press.
Deppermann, Arnulf. 2014. “Multimodal participation in simultaneous joint projects: Inter-
personal and intrapersonal coordination in paramedic emergency drills.” In Multiactivity
in Social Interaction: Beyond Multitasking, edited by Pentti Haddington, Tiina Keisanen,
Lorenza Mondada, and Maurice Nevile, 247–281. Amsterdam, Philadelphia: Benjamins.
Deppermann, Arnulf. 2015. “When recipient design fails: Egocentric turn-design of instruc-
tions in driving school lessons leading to breakdowns of intersubjectivity.” Gesprächs-
forschung. Online-Zeitschrift zur verbalen Interaktion 16: 63–101.
Deppermann, Arnulf, and Reinhold Schmitt. 2007. “Koordination. Zur Begründung eines
neuen Forschungsgegenstandes.” In Koordination. Analysen zur multimodalen Interak-
tion, edited by Reinhold Schmitt, 15–54. Tübingen: Narr.
Dingemanse, Mark, Francisco Torreira, and Nick J. Enfield. 2013. “Is huh? a universal
word? Conversational infrastructure and the convergent evolution of linguistic items.”
PLoS One 8 (11): Online.
Ellingsen, Kristian, Grahame J. Coleman, Vonne Lund, and Cecilie M. Mejdell. 2014. “Us-
ing qualitative behaviour assessment to explore the link between stockperson behaviour
and dairy calf behaviour.” Applied Animal Behaviour Science 153: 10–17.
Evans, Dylan. 2001. Emotion: The Science of Sentiment. Oxford: Oxford University Press.
Fröhlich, Marlen. 2017. “Takng turns across channels: Conversation-analytic tools in animal
communication.” Neuroscience and Biobehavioral Reviews 80: 201–209.
 Communicative Resources Beyond the Verbal Tier 49

Fuchs, Thomas, and Hanne De Jaegher. 2009. “Enactive intersubjectivity: Participatory sense-
making and mutual incorporation.” Phenomenology and the Cognitive Sciences 8: 465–486.
Garfinkel, Harold. 1967. Studies in Ethnomethodology. Englewood Cliffs: Prentice Hall.
Giles, Howard. 2008. “Communication accommodation theory.” In Engaging Theories in
Interpersonal Communication. Multiple Perspectives, edited by Leslie A. Baxter and
Dawn O. Braithwaite, 161–173. Los Angeles: Sage.
Giles, Howard. 2009. “The process of communication accommodation.” In The New Socio-
linguistics Reader, edited by Nikolas Coupland and Adam Jaworski, 276–286. Basing-
stoke, New York: Palgrave Macmillan.
Goffman, Erving. 1983. “The interaction order.” American Sociological Review 48: 1–17.
Goodwin, Charles. 1981. Conversational Organization, Language, thought, and Culture:
Advances in the Study of Cognition. New York: Academic Press.
Goodwin, Charles. 2000. “Pointing and the collaborative construction of meaning in apha-
sia.” Proceedings of the seventh annual Symposium About Language and Society, Austin.
Goodwin, Charles, ed. 2003. Conversation and Brain Damage. Oxford: Oxford University
Press.
Haddington, Pentti, Tiina Keisanen, Lorenza Mondada, and Maurice Nevile. 2014. “To-
wards multiactivity as a social and interactional phenomenon.” In Multiactivity in Social
Interaction: Beyond Multitasking, edited by Pentti Haddington, Tiina Keisanen, Lorenza
Mondada, and Maurice Nevile, 3–32. Amsterdam, Philadelphia: Benjamins.
Haddington, Pentti, Lorenza Mondada, and Maurice Nevile. 2013. “Being mobile: Interac-
tion on the move.” In Interaction and Mobility: Language and the Body in Motion, edited
by Pentti Haddington, Lorenza Mondada, and Maurice Nevile, 3–61. Berlin, Boston: de
Gruyter.
Halliday, Michael A.K. 1973. Explorations in the Functions of Language, Explorations in
Language Study. London: Arnold.
Heath, Christian. 1986. Body Movement and Speech in Medical Interaction. Cambridge:
Cambridge University Press (Studies in Emotion and Social Interaction).
Horisk, Claira, and Reginald B. Cocroft. 2013. “Animal signals: Always influence, some-
times information.” In Animal Communication Theory: Information and Influence, ed-
ited by Ulrich E. Stegmann, 259–280. Cambridge: Cambridge University Press.
Jewitt, Carey. 2014. “An introduction to multimodality.” In The Routledge Handbook of
Multimodal Analysis, edited by Carey Jewitt, 15–30. London, New York: Routledge.
Kotrschal, Kurt. 2013. “Emotions are at the core of individual social performance.” In Emo-
tions of Animals and Humans: Comparative Perspectives, edited by Shigeru Watanabe
and Stan A. Kuczaj, 3–21. Tokyo: Springer.
Kuzuoka, Hideaki, Karola Pitsch, Yuya Suzuki, Ikkaku Kawaguchi, Keiichi Yamazaki,
Akiko Yamazaki, Yoshinori Kuno, Paul Luff, and Christian Heath. 2008. “Effect of re-
start and pauses on achieving a state of mutual orientation between a human and a robot.”
Proceedings of the 2008 ACM Conference on Computer Supported Cooperative Work,
San Diego.
Lyons-Ruth, Karlen, Nadia Bruschweiler-Stern, Alexandra Harrison, Alexander C. Morgan,
Jeremy P. Nahum, Louis Sander, Daniel N. Stern, and Edward Z. Tronick. 1998. “Im-
plicit relational knowing: Its role in development and psychoanalytic treatment.” Infant
Mental Health Journal 19 (3): 282–289.
Merleau-Ponty, Maurice. 2005 [1945]. Phenomenology of Perception. Translated by Colin
Smith. London, New York: Routledge.
Meyer, Christian, and Ulrich von Wedelstaedt. 2017. “Intercorporeality, interkinesthesia,
and enaction: New perspectives on moving bodies in interaction.” In Moving Bodies
50 Heike Ortner

in Interaction – Interacting Bodies in Motion: Intercorporeality, Interkinesthesia, and

Enaction in Sports, edited by Christian Meyer and Ulrich von Wedelstaedt, 1–23. Am-
sterdam, Philadelphia: Benjamins.
Miles, Meredith C., and Matthew Fuxjager. 2018. “Animal choreography of song and dance:
A case study in the Montezuma oropendola, Psarocolius montezuma.” Animal Behaviour
140: 99–107.
Mondada, Lorenza. 2014. “The temporal orders of multiactivity: Operating and demonstrat-
ing in the surgical theatre.” In Multiactivity in Social Interaction: Beyond Multitasking,
edited by Pentti Haddington, Tiina Keisanen, Lorenza Mondada, and Maurice Nevile,
33–75. Amsterdam, Philadelphia: Benjamins.
Nevile, Maurice. 2015. “The embodied turn in research on language and social interaction.”
Research on Language and Social Interaction 48 (2): 121–151.
Norris, Sigrid. 2011. Identity in (Inter)action: Introducing Multimodal (Inter)action analy-
sis, Trends in Applied Linguistics. Berlin, New York: de Gruyter Mouton.
Pearson, Susan. 2013. “Speaking bodies, speaking minds: Animals, language history.” His-
tory and Theory Theme Issue 52: 91–108.
Quasthoff, Uta M., and Tabea Becker, eds. 2005. Narrative Interaction, Studies in Narrative.
Amsterdam, Philadelphia: Benjamins.
Sacks, Harvey, Emanuel A. Schegloff, and Gail Jefferson. 1974. “A simplest systematics for
the organization of turn-taking for conversation.” Language 50: 696–735.
Scheider, Linda, Bridget M. Waller, Leonardo Oña, Anne M. Burrows, and Katja Liebal.
2016. “Social use of facial expressions in hylobatids.” PLoS One 11 (3): 1–11.
Sheets-Johnstone, Maxine. 2009. “Animation: The fundamental, essential, and properly de-
scriptive concept.” Continental Philosophy Review 42 (3): 375–400.
Streeck, Jürgen, Charles Goodwin, and Curtis LeBaron. 2011. “Embodied interaction in the
material world: An introduction.” In Embodied Interaction: Language and Body in the
Material World, edited by Jürgen Streeck, Charles Goodwin, and Curtis LeBaron, 1–26.
Cambridge: Cambridge University Press.
Stukenbrock, Anja. 2014. “Pointing to an ‘empty’ space. Deixis am Phantasma in face-to-
face interaction.” Journal of Pragmatics 74: 70–93.
Trevarthen, Colwyn. 1987. “Sharing making sense: Intersubjectivity and the making of an
infant’s meaning.” In Language Topics: Essays in Honour of Michael Halliday, edited
by Ross Steele and Terry Threadgold, 177–199. Amsterdam, Philadelphia: Benjamins.
Watanabe, Shigeru, and Stan A. Kuczaj, eds. 2013. Emotions of Animals and Humans: Com-
parative Perspectives: The Science of the Mind. Tokyo: Springer.
Wynne, Clive D.L., Nicole R. Dorey, and Monique A.R. Udell. 2011. “The other side of the
bond: Domestic dogs’ human-like behaviors.” In How Animals Affect Us: Examining
the Influence of Human-Animal Interaction on Child Development and Human Health,
edited by Peggy McCardle, Sandra McCune, James A. Griffin, and Valerie Maholmes,
101–115. Washington: American Psychological Association.
6
HOW STUDIES OF COMMUNICATION
AMONG NONHUMANS AND BETWEEN
HUMANS AND NONHUMANS CAN
INFORM SETI
Irene M. Pepperberg

Introduction
Humans have long wished to “talk” with alien beings – first with creatures with whom
they shared their world and later with those in worlds beyond. Legends and stories,
from those involving King Solomon (who purportedly had a ring that enabled him to
communicate at will with all the birds and beasts in his realm; Lorenz 1952), to many
children’s books (e.g., Lofting 1948; King-Smith 1984), to those of Native Ameri-
cans (who supposedly could change into various animals and thus share their lives;
Rasmussen 1972), to various Hollywood blockbusters (e.g., Arrival, Contact, Close
Encounters of the Third Kind), demonstrate humans’ deep interests in these possibili-
ties. For most of human history, understanding nonhuman codes and communicating
with animals were indeed reserved for individuals like the fictional Dr. Doolittle –
though people such as trackers could predict some nonhuman behavior based on, for
example, spoor and probabilistic reasoning. True two-way communication, however,
was impossible. Nevertheless, starting in the twentieth century, humans began to crack
these codes (described in what follows), and some even trained nonhumans to use lim-
ited forms of human systems (e.g., Pepperberg 1999). Although these human endeav-
ors have not been completely successful, some progress has been made, and I review
what is known – or more accurately, what we are beginning to understand – about
some Earthly nonhuman communication systems, and I propose how this knowledge
might be used to understand possible extraterrestrial communicative systems and how
we might establish two-way communication systems with other beings.

Decoding the World Around Us

Around the middle of the twentieth century, researchers began using rigorous sci-
entific protocols to decipher the meaning of nonhuman communication systems.
DOI: 10.4324/9781003352174-6
52 Irene M. Pepperberg

Decades of work led to some knowledge of information conveyed in systems such

as birdsong (Kroodsma and Miller 1996; Catchpole and Slater 2008), vervet and
prairie dog alarm calls (respectively, Cheney and Seyfarth 1990; Slobodchikoff,
this volume [Chapter 9]), chimpanzee vocalizations (Goodall 1986; Laporte and
Zuberbühler 2010, Hobaiter et al., this volume [Chapter 3]), and dolphin whis-
tles (McCowan and Reiss 1997; Ferrer-i-Cancho and McCowan 2009; Herzing,
this volume [Chapter 4]). Initially, experiments simply involved engaging these
nonhumans in playback experiments in which researchers would record various
vocalizations, then play them back to individuals of the same species as those they
recorded and document what ensued. Based on behavioral reactions of numerous
subjects, scientists could then classify a recording as, for example, aggressive or
affiliative – for example, as being used, respectively, to defend territory or attract
a mate. Clearly these findings were but a first step, somewhat akin to a visitor to
Earth playing two recordings outside a theater, one stating “Free beer!” and the
other “FIRE!”, and seeing how humans reacted. Current research has, however,
provided a better understanding of nonhuman systems, in part because of a num-
ber of technological breakthroughs and the availability of extensive computational
power.
Of particular interest were findings that, of all creatures studied, birds were
likely to have some of the most complicated systems because, unlike nonhuman
primates, they were vocal learners (Marler 1967). That is, for avian species known
as ocsines, communication systems were not innately specified, but are acquired,
much as for humans, via learning from older, competent models (e.g., Okanoya
2017; Todt et al. 1979). Depending upon the species, the numbers of songs in a
repertoire can vary from one to thousands, and the ways in which such songs con-
vey meaning usually differ across species. Even songbirds that do not learn their
vocalizations (the suboscines) vary how they used their songs in order to convey
numerous different meanings. Given that our expectation is that the extraterrestrial
signals we receive will likely be converted to auditory frequencies for human bene-
fit (e.g., as were the Laser Interferometer Gravitational-Wave Observatory [LIGO]
signals of two black holes colliding), and that such transposition will – if performed
properly – maintain the basic characteristics of the original signal, birdsong might
indeed be an exceptionally good model for study. Notably, after more than sixty
years of research, we are still learning about birdsong. Having found so many dif-
ferences among species in a single taxa on Earth, we must be open to an immense
number of possibilities when examining signals from beyond Earth. For starters, let
us examine possibilities presented by the avian mode of communication.
The first concerns basic attributes of the signal itself. For example, some bird-
song (e.g., hummingbird vocalizations: Olson et al. 2018; Pytte et al. 2004) occurs
at the very limits of human auditory thresholds – and, if we go beyond birds, to
bats, cetaceans, and elephants (see, respectively, Prat et al. 2016; Finneran et al.
2008; McComb and Reby 2009) – we thus must examine frequencies well above
and below those that humans hear. Avian signals may also be visual (e.g., the flash
 How Studies of Communication Can Inform SETI 53

of a redwing blackbird’s epaulets, Peek 1972), so we must be open to any sort of

incoming energetically interesting signals.
We also might expect that the signal (in whatever modality it is delivered) will
be repetitive – the equivalent of “Hello . . . are you there?” sent over and over
(see, however, Harbour, this volume [Chapter 18], for many other intriguing pos-
sibilities). We would have little problem deciphering a signal so semantically and
syntactically similar to human languages; however, aliens’ ideas about everything
will likely be quite different from those of humans, and thus their communication
systems for conveying those ideas will also likely differ from those of humans.
We must, for example, be open to the possibility that the number of repeti-
tions of a signal (or possible combinations of different types of signals) rather than
the individual units are meaningful: Templeton et al. (2005) demonstrated that
the number of “dee” syllables in black-capped chickadee alarm calls indicate the
extent and severity of the threat; Suzuki et al. (2016) show that for the Japanese
great tit, “dee” calls when produced alone are used to attract a caller. Researchers
like Smith (1988) found that combinations of flights and song repetitions indicate
different meaning such as various levels of aggression for flycatchers. Thus, unlike
human words that themselves have meaning, might specific patterns of similar sig-
nals have different meanings in extraterrestrial broadcasts? Interestingly, Balaban
(1988) found that although swamp sparrow song consists primarily of repetitions
of three-note syllables wherever the singer lives, the order of the syllables specifies
the geographic location of the singer – that is, constitutes a dialect, not the song
of a different species. Clearly such possibilities must be taken into account in our
interpretation of incoming information.
Humans might also miss meaning contained in the fine structure of signals. Pos-
sibly what initially appears to be a pure tone or a simple sine wave might have addi-
tional levels of complexity. Several studies showed that birdsongs that appeared to
consist of just a few notes, when played back at much reduced speeds, contained
a huge wealth of additional information. For example, Dooling and Prior (2017)
demonstrated that zebra finches are extremely sensitive to temporal fine structure,
although these authors do not examine what types of information may be encoded
in this fine-structure variation. Vélez et al. (2015) revealed seasonal plasticity in
the fine structure of the songs of three different songbirds; would such variation
exist unless it designates something important? What might we miss if we failed to
actively search for such information in incoming signals?
Although sometimes different songs, somewhat like different human sentences,
are used to convey different meanings (e.g., warblers’ use of one song type for ter-
ritorial defense and another for mate attraction, Demko et al. 2013), it is possible
that a particular vocal pattern is not important, but rather its contextual use. Song
sparrows, for example, have repertoires of about five to seven songs, used pre-
dominantly for territorial defense. Different songs themselves do not have separate
meanings. The meaning of a bird’s choice of song, however, depends on the one
that its neighbor has sung, and the physical origin of the heard song – that is, the
54 Irene M. Pepperberg

spot from which the neighbor sings in their respective territories. Stoddard (1996)
demonstrated that the same song sung on the edge of a territory will evoke a com-
pletely different response – and neighbor’s song choice – from one sung during an
intrusion just over the territorial border. Beecher et al. (2000) refined this further,
showing that a sparrow will exactly match the song of a neighbor that is singing at
or past the territorial boundary, using the song to present a strong vocal challenge;
in contrast, the sparrow will choose to sing a different song – one merely from
the repertoire of a neighbor that is singing deep within its own territory, thereby
acknowledging the presence of the singer as a familiar, non-threatening entity. For
some species such as nightingales, it is not song choice but the temporal pattern
of overlapping that acts as a threat (Hultsch and Todt 1982). Marsh wrens seem to
play a kind of poker, in which bird A not only matches the song of bird B whom
he is trying to dominate, but then acts to “raise him one” by singing the next song
in bird B’s repertoire, which is not the next song in his own (Kroodsma 1979).
For other species, repertoire size might indicate fitness, so that the point is to pro-
duce an extremely large number of different signals (reviewed in Todt and Naguib
2000). Would we be smart enough to recognize such patterns in extraterrestrial
signals in our attempts to respond appropriately?
The preceding discussions do not examine issues of any possible direct paral-
lels between what could be syntactic arrangements in avian song elements and
human language, which is currently a matter of debate (e.g., see Buckingham and
Christman 2010; Lipkind et al. 2013). My point is not to emphasize the similarities
between birdsong and human language (of which there are many) but some of the
numerous differences, so that we are aware of the many various possible forms that
communicative structures can take. Such knowledge would be crucial if we are to
make sense of alien signaling systems.

Learning to Communicate
Although the likelihood of an actual physical encounter between humans and alien
species is far less than that of our having to decode their broadcast signals, the pos-
sibility is non-zero, and if such is the case, we would need to attempt some form
of direct two-way communication. Such beings, having figured out how to get
to us, would likely be far more advanced than humans, and thus would probably
have already developed protocols to try to learn our communication code, or teach
us theirs. Having our own procedures in place to instruct such visitors would be
prudent, but lacking any information as to the most appropriate modality (auditory,
visual, tactile, olfactory, or some form we cannot yet even imagine) puts us at a dis-
advantage. A likely basis for such procedures might, however, be methods devel-
oped by researchers who have already taught nonhumans some elements of human
communication systems – those of us who have worked with nonhuman primates,
cetaceans, and parrots. Given that communication systems are by definition social
in nature, the expectation is that some form of social learning system would be
 How Studies of Communication Can Inform SETI 55

most appropriate. Researchers actually tasked to establish human-ET communica-

tion will likely develop a system based on some combination of forms proposed by
me and other authors in this volume. To make such material available, I describe in
detail procedures used with my grey parrots.
The techniques are based on social modeling theory, which has its origins in
the work of researchers such as Piaget (1952), Vygotsky (1962), Todt (1975), and
Bandura (1977). Space does not permit a full discussion of their findings, but the
principles of the theory involve providing input that is optimal – one that is tai-
lored to the learners’ level of competence and adjusted as competence increases
(for those interested, see summary in Pepperberg 1999). Social modeling theory
provides a framework for devising training procedures, but this framework must be
put into practice. To whit: what factors characterize optimal input? Based on ideas
of Piaget and Vygotsky, and competence issues, optimal input should (1) correlate
well with specific aspects of an individual’s environment (i.e., be “referential”); (2)
have functional meaning relevant to the individual’s environment (also known as
“contextual applicability”); and (3) be socially interactive. I briefly described these
factors, quoted from Pepperberg (1999).

Reference – Reference is, in general, what signals “are about” (Smith 1991). Refer-
ence concerns the direct relationship between a signal and an object or action.
Reference is not always easily determined. For example, “ape” can refer to a
subset of nonhuman primates, but may also mean an action, as to “ape,” or
imitate, a behavior. Similarly, a bird’s alarm call may refer to either the predator
nearby, the action the bird is about to take, or both. Thus, not all information
contained in a signal involves a single referent. The more explicit the signal’s
referent, however, the more easily the signal can be learned.
Functionality – Functionality (contextual applicability) involves pragmatics of sig-
nal use: when a signal is to be used and the effects of using information in the
signal. Explicit demonstration of functionality shows when using a signal is
advantageous and the specific advantage gained by its use. The way a signal is
used and its effect on recipients may depend upon environmental context – the
phrase “My, don’t we look nice today?” has one meaning and effect for a little
girl in a party dress and a different meaning and effect for a hungover friend.
Functionality also helps define reference; that is, context defines “ape” as noun
versus verb. The more explicit a signal’s functionality, the more readily the sig-
nal can be learned.
Social interaction – Social interaction has three major functions that can be clari-
fied by examples. First, social interaction can highlight which environmental
components should be noted; a subject can be directed to an object’s color to
learn color labels (“Look at the blocks. The color of this one is blue; the color
of that one is green”). Second, social interaction can emphasize common attrib-
utes – and thus possible underlying rules – of diverse actions (i.e., “Give me
the ball” versus “Take the block”). Third, social interaction allows input to be
56 Irene M. Pepperberg

continuously adjusted to match the receiver’s level (“Yes, you found a block
among these toys! Now can you find the green block?”). Interaction can pro-
vide a contextual explanation for an action and demonstrate its consequences
(“I don’t know what toy you want . . . do you want the ball or the block? Tell me
what you want, and you can have it.”). Interactive input thus facilitates learning.

In sum, reference and functionality refer to real world use of input, social inter-
action highlights various components of input, and all are necessary for meaning-
ful learning. I thus reasoned that to teach a bird to communicate with humans, my
training procedure needed to take these factors into account. The critical point,
however, was my hypothesis that a parrot’s acquisition of a human-based code
was a form of exceptional learning: I believed that despite these birds’ abilities to
reproduce all sorts of sounds, some strong inhibition existed toward learning to use
allospecific sounds in a functional manner; I further believed that, to overcome this
inhibition, training would have to be carefully adjusted to the parrot’s abilities and
include intense interactions and extremely clear demonstrations of reference and
functionality. I decided that the best approach would be to modify a procedure used
by Todt (1975), called the Model/Rival or M/R technique. He had demonstrated the
effectiveness of social interaction for training parrots to produce human speech in
simple dialogues without requiring them to understand the meaning of these dia-
logues; what if I adapted his method to incorporate referentiality and functionality
in order to establish communication?
My training system, because of its similarity to Todt’s, is also called the M/R
technique. In my procedure, however, an interaction is not only modeled; it also
involves three-way interactions among two human speakers and the avian student.
I provide details of the procedure, although the material is available elsewhere
(e.g., Pepperberg 1981).
During M/R training, humans demonstrate to the bird the types of interactive
responses that are to be learned. In a typical interaction, the bird is on a perch,
cage, or the back of a chair, and observes two humans handling an object in which
it has already demonstrated interest (e.g., as a preening implement). While the bird
watches, one human “trains” the second human. The trainer presents an object, asks
questions about the object (e.g., “What’s here?”, “What color?”, “What shape?”),
and gives praise and the object itself as a reward for a correct answer – thereby
showing the direct connection between the label and the item to which it refers.
Unlike Todt’s procedure, our technique demonstrates referential and contextual use
of labels for observable objects, qualifiers, quantifiers, and, on occasion, actions.
As in Todt’s procedure, the second human is a model for the bird’s responses and
a rival for the trainer’s attention. The model/rival occasionally errs (i.e., produces
garbled utterances, partial identifications, etc., that are similar to mistakes being
made by the bird at the time). Disapproval for an incorrect response is demon-
strated by scolding and temporarily removing the object from sight. Because the
human model/rival is, however, encouraged to try again or talk more clearly (e.g.,
 How Studies of Communication Can Inform SETI 57

“You’re close; say better”), the procedure also allows the bird to observe “correc-
tive feedback,” which also assists acquisition (see Goldstein 1984; Vanayan et al.
1985; Moerk 1994).
As part of the demonstration of functionality and relevance, the M/R protocol
also repeats the interaction while reversing roles of trainer and model/rival, and
occasionally includes the bird in interactions. Thus, unlike Todt’s subjects (see also
Goldstein 1984), our birds do not simply hear stepwise vocal duets, but rather
observe a communicative process that involves reciprocity. We show that interac-
tion is indeed a “two-way street” in that one person is not always the questioner
and the other always the respondent, and how the process can be used to effect
environment change. I surmised, based on Bandura’s (1977) studies of interac-
tive modeling, that Todt’s failure to demonstrate role reversal between trainer and
model/rival explained why his birds could not transfer their responses to anyone
other than the particular human who posed the questions, and why his birds never
learned both parts of the interaction. In contrast, my birds respond to, interact with,
and learn from all their trainers.
Three actions by trainers insure that our birds indeed attend to sessions (Pep-
perberg 1992); these actions are consistent with the principles of social modeling
theory. First, trainers adjust the level of modeling to match a bird’s current capaci-
ties. If, for example, a label being trained (“wool”) resembles one already in the
repertoire (“wood”), trainers praise but do not reward the bird’s likely initial use
of the existent label and clearly demonstrate how the two labels differ with respect
to both sound and referent. Trainers then adjust their rewards as a bird practices its
utterances to challenge it to achieve correct pronunciation. Second, a bird must be
interested in obtaining the items used in training. Trainers working on a numerical
task who choose, for example, corks rather than keys are more likely engage one
bird’s attention, whereas another bird might prefer keys. Third, trainers must act as
though they themselves find the task interesting. A bird is less likely to ignore the
session and begin to preen if the emotional content of the trainers’ interactions sug-
gests there is real relevance to the task, and trainers who actively engage the bird in
the task are more likely to be successful.
The M/R technique is the primary method for introducing new labels and con-
cepts and for shaping correct pronunciation, but another procedure helps clarify
pronunciation (Pepperberg 1981). Because this technique does encourage some imi-
tation, it is used only after a bird begins to attempt a new label in the presence of
a new object (i.e., after a bird makes some connection between sound and object).
We present the new object along with a string of “sentence frames” – phrases like
“Here’s paper!”, “Such a big piece of paper!” The target label, “paper,” is con-
sistently stressed and the one most frequently heard, but not as a single, repetitive
utterance. The target label is also consistently at the end of the phrase; conceivably
parrots, like humans (e.g., Lenneberg 1971), most easily remember ends of word
strings. This combination of nonidentical but consistent vocal repetition and physi-
cal presentation of object resembles parental behavior for introducing labels for new
58 Irene M. Pepperberg

items to very young children (Berko-Gleason 1977; de Villiers and de Villiers 1978),
and appears to have two effects: a bird (1) hears the label employed as it is to be used
in normal, productive speech; and (2) learns to reproduce the emphasized, targeted
label without associating word-for-word imitation of its trainers with reward.
On occasion, birds experiment with labels in their repertoire and produce novel
vocalizations. To encourage such recombinations of – or variants upon – parts of
labels and enlarge the referential repertoire, we reward these utterances (when pos-
sible) with appropriate objects and use a variant of the M/R technique to associate
the novel vocalizations and objects. The technique, called “referential mapping,” is
described in detail in Pepperberg (1990).
Thus, by integrating Bandura’s social modeling theory with the ideas and results
of scientists as diverse as Piaget, Vygotsky, and Todt, I devised training procedures
for teaching a parrot to communicate with humans. Whether such procedures could
be adapted for use with creatures far more “alien” than a grey parrot remains to be
seen. Clearly, extraterrestrials, unlike parrots, likely will not share similarities with us
in auditory and visual systems, may not respond to the same forms of social interaction
(e.g., like nonhuman primates, may see direct gaze as a threat rather than a socially
acceptable form of interaction), and thus “optimal input” for an Earth-dwelling creature
maybe completely suboptimal for ET. However, I propose these procedures as a pos-
sible basis, not as a rigid protocol, for the establishment of interspecies communication.

Conclusions
From the material discussed here and in many other chapters, it is clear that organ-
isms do not need all the elements of human language to establish some level of
inter- or intraspecific communication, and that communication systems may
involve aspects completely foreign to human language. Granted, human language
enables levels of communication that have so far not been found to exist in non-
humans, but one can argue that humans have not been clever enough to unearth all
possible complexities inherent in nonhuman systems. We nevertheless can under-
stand the gist of nonhuman systems; we have taught them elements of ours. Such
knowledge could be used to establish rudimentary interplanetary dialogues.
Another possibility is that extraterrestrial communication systems have layers
of complexity far beyond those of human systems. That is, might our system seem
lacking in many essential components when compared to theirs? Might we find
ourselves struggling to learn a system far more advanced than ours? Will we have
the humility to accept instruction if necessary? Only by being open to all possibili-
ties and eventualities might we have some chance of success.

References
Balaban, Evan. 1988. “Bird song syntax: Learned intraspecific variation is meaningful.”
Proceedings of the National Academy of Sciences, USA 85: 3657–3660.
 How Studies of Communication Can Inform SETI 59

Bandura, Albert. 1977. Social Modeling Theory. Chicago: Aldine-Atherton.

Beecher, Michael D., Elizabeth Campbell, John M. Burt, Christopher E. Hill, and J. Cully
Nordby. 2000. “Song-type matching between neighboring sparrows.” Animal Behaviour
59: 21–27.
Berko-Gleason, Jean. 1977. “Talking to children: Some notes on feedback.” In Talking to
Children; edited by Catherine E. Snow and Charles A. Ferguson, 199-205. Cambridge:
Cambridge University Press.
Buckingham, Hugh W., and Sarah S. Christman. 2010. “Charles Darwin and the evolution
of human grammatical systems.” Journal of the History of Neuroscience 19: 121–139.
Catchpole, Clive K., and Peter J.B. Slater. 2008. Bird Song: Biological Themes and Varia-
tion. Cambridge: Cambridge University Press.
Cheney, Dorothy L., and Robert M. Seyfarth. 1990. How Monkeys See the World: Inside the
Mind of Another Species. Chicago: University of Chicago Press.
Demko, Alana D., Leonard R. Retsma, and Cynthia A. Staicer. 2013. “Two song categories
in the Canada warbler (Cardellina canadensis).” Auk 130: 609–616.
de Villiers, Jill G., and Peter A. de Villiers. 1978. Language Acquisition. Cambridge, MA:
Harvard University Press.
Dooling, Robert J., and Nora H. Prior. 2017. “Do we hear what birds hear in birdsong?”
Animal Behaviour 124: 283–289.
Ferrer-i-Cancho, Ramon, and Brenda McCowan. 2009. “A law of word meaning in dolphin
whistle types” Entropy 11: 688–701.
Finneran, James J., Dorian S. Houser, Dave Blasko, Christie Hicks, Jim Hudson, and Mike
Osborn. 2008. “Estimating bottlenose dolphin (Tursiops truncates) auditory thresholds
from single and multiple evoked potentials.” Journal of the Acoustical Society of Amer-
ica 123: 542–551.
Goldstein, Howard. 1984. “The effects of modeling and corrected practice on generative
language and learning of preschool children.” Journal of Speech and Hearing Disorders
49: 389-398.
Goodall, Jane. 1986. The Chimpanzees of Gombe: Patterns of Behavior. Cambridge, MA:
Harvard.
Hultsch, Henrike, and Dietmar Todt. 1982. “Temporal performance roles during vocal inter-
actions in nightingales (Luscinia megarhynchos).” Behavioural Ecology & Sociobiology
11: 253–260.
King-Smith, Dick. 1984. Harry=s Mad. London: Victor Gollantz Ltd.
Kroodsma, Donald E. 1979. “Vocal dueling among male marsh wrens: Evidence for ritual-
ized expression of dominance/subordinance.” Auk 96: 506–515.
Kroodsma, Donald E., and Edward H. Miller, eds. 1996. Ecology and Evolution of Acoustic
Communication in Birds. Ithaca, NY: Cornell University Press.
Laporte, Marion N.C., and Klaus Zuberbühler. 2010. “Vocal greeting behavior in female
chimpanzees.” Animal Behaviour 80: 467–473.
Lenneberg, Eric H. 1971. “Of language, knowledge, apes, and brains.” Journal of Psycho-
linguistic Research 1: 1-29.
Lipkind, Dina, Gary F. Marcus, Douglas K. Bemis, et al. 2013. “Stepwise acquisition of vo-
cal combinatory capacity in songbirds and human infants.” Nature 498: 104–109.
Lofting, Hugh. 1948. The Voyages of Dr. Doolittle. Philadelphia: Lippincott.
Lorenz, Konrad. 1952. King Solomon’s Ring. Translated by M.K. Wilson. New York: Harper
& Row.
Marler, Peter. 1967. “Comparative study of song development in sparrows.” Proceedings of
the XIVth International Ornithological Congress, 231–244. Blackwell, Oxford.
60 Irene M. Pepperberg

McComb, Karen, and David Reby. 2009. “Communication in terrestrial animals.” In Ency-
clopedia of Neuroscience, Vol. 2, edited by Larry R. Squire, 1167–1171. Oxford: Aca-
demic Press.
McCowan, Brenda, and Diana L. Reiss. 1997. “Vocal learning in captive bottlenose dol-
phins: A comparison with human and nonhuman animals.” In Social Influences on Vocal
Development, edited by Charles T. Snowdon and Martine Hausberger, 178–207. Cam-
bridge: Cambridge University Press.
Moerk, Ernst L. 1994. “Corrections in first language acquisition: Theoretical controversies
and factual evidence.” International Journal of Psycholinguistics 10: 33–58.
Okanoya, Kazuo. 2017. “The biopsychology and development of birdsong.” In APA Hand-
book of Comparative Psychology, edited by Josep Call, Gordon M. Burghardt, Irene M.
Pepperberg, Charles T. Snowdon, and Thomas R. Zentall, 539–555.
Olson, Christopher R., Marcella Fernández-Vargas, Christine V. Portfors, and Claudio V.
Mello. 2018. “Black Jacobin hummingbirds vocalize above the known hearing range of
birds.” Current Biology 28: R187–R207.
Peek, Frank W. 1972. “An experimental study of the territorial function of vocal and visual
displays in the male Red-winged Blackbird (Agelaius phoeniceus).” Animal Behaviour
20: 112-118.
Pepperberg, Irene M. 1981. “Functional vocalizations by an African Grey parrot (Psittacus
erithacus).” Zeitschrift für Tierpsychologie 55: 139-160.
Pepperberg, Irene M. 1990. “Referential mapping: A technique for attaching functional sig-
nificance to the innovative utterances of an African Grey parrot.” Applied Psycholinguis-
tics 11: 23-44.
Pepperberg, Irene M. 1992. “Social interaction as a condition for learning in avian species.”
In The Inevitable Bond, edited by Hank Davis and Dianne Balfour, 178–204. Cambridge:
Cambridge University Press.
Pepperberg, Irene M. 1999. The Alex Studies. Cambridge, MA: Harvard University Press.
Piaget, Jean. 1952. The Origins of Intelligence in Children. Translated by Margaret Cook.
New York: International Universities Press. (Original written in French in 1936).
Prat, Yosef, Mor Taub, and Yossi Yovel. 2016. “Everyday bat vocalizations contain informa-
tion about emitter, addressee, context, and behavior.” Scientific Reports 6: 39419.
Pytte, Carolyn L., Millicent S. Ficken, and Andrew Moiseff. 2004. “Ultrasonic singing by
the blue-throated hummingbird: A comparison between production and perception.”
Journal of Comparative Physiology A 190: 665–673.
Rasmussen, Knud. 1972. The Netsilik Eskimos. In Shaking the Pumpkin, edited by J. Roth-
enberg, 45. Garden City, NY: Doubleday.
Smith, W. John. 1988. “Patterned daytime singing of the eastern wood-pewee (Contopus
virens).” Animal Behaviour 36: 1111-1123.
Smith, W. John. 1991. “Animal communication and the study of cognition.” In Cognitive
Ethology: The Minds of Other Animals, edited by Carolyn A. Ristau, 209-230. Hillsdale,
NJ: Erlbaum.
Stoddard, Philip. 1996. “Vocal recognition of neighbors by territorial passerines.” In Ecol-
ogy and Evolution of Acoustic Communication in Birds, edited by Donald E. Kroodsma
and Edward. L. Miller, 356–374. Ithaca: Cornell University Press.
Suzuki, Toshitaka N., David Wheatcroft, and Michael Griesser. 2016. “Experimental evi-
dence for compositional syntax in bird calls.” Nature Communications 7: 10986.
Templeton, Christopher N., Erick Greene, and Kate Davis. 2005. “Allometry of alarm
calls: Blackcapped chickadees encode information about predator size.” Science 308:
1934–1937.
 How Studies of Communication Can Inform SETI 61

Todt, Dietmar. 1975. “Social learning of vocal patterns and modes of their applications in
Grey parrots.” Zeitschrift für Tierpsychologie 39: 178-188.
Todt, Deitmar, Henrike Hultsch, and Dietmar Heike. 1979. “Conditions affecting song ac-
quisition in nightingales.” Zeitschrift für Tierpsychologie 51: 23-35.
Todt, Dietmar, and Mark Naguib. 2000. “Vocal interactions in birds: The use of song as a
model in communication.” Advances in the Study of Behavior 29: 247–296.
Vanayan, Marina, Heather A. Robertson, and Gerald B. Biederman. 1985. “Observational
learning in pigeons: The effects of model proficiency on observer performance.” Journal
of General Psychology 112: 349-357.
Vélez, Alejandro, Megan D. Gall, and Jeffrey R. Lucas. 2015. “Seasonal plasticity in audi-
tory processing of the envelope and temporal fine structure of sounds in three songbirds.”
Animal Behaviour 103: 53–63.
Vygotsky, Lev S. 1962. Thought and Language. Cambridge, MA: MIT Press.
7
PATTERNS OF COMMUNICATION
OF HUMAN COMPLEX SOCIETIES
AS A BLUEPRINT FOR ALIEN
COMMUNICATION
Anamaria Berea

Introduction
In this chapter, I explore patterns of ubiquitous communication as a sign of suc-
cessful communication evolution that started local and became global. These pat-
terns are the framework within which an alien language can develop and would
bring us to a better understanding of what alien languages might look like.
The common ways in which we communicate today, whether language-based
or symbol-based or sounds-based – or any other forms – are essentially patterns of
communication that have emerged due to the coevolution of our languages, cul-
tures, and technologically inclined civilization. A great example is the growth of
the English language (and not only English; Latin is another historical example)
with words such as “to google” or “meme” in recent years, while in the digital
realm the growth of the menu of emoticon symbols has been quite astonishing
during the past couple of decades, concurrent with a specific technological evolu-
tion that is very recent in our history. But at a larger, global point of view, what
I am arguing for is that the globalization of certain systems of communication
which transcend boundaries and cultures, such as the emergent ones I exempli-
fied or the designed ones, such as those used in transportation (e.g., aviation or
maritime symbols, traffic lights), are – in effect – great representations of how
language coevolves with our complex society. Therefore, this is an invitation to
think about xenolanguages as emergent and coevolving phenomena that can be
thought of only within the context of the social and technological alien civiliza-
tion, as well.
While thinking about xenolinguistics in this framework might complicate things
further, it also helps us consider which coevolutionary patterns have endured the
passing of time, civilization rises and falls, and crossed cultural borders, and

DOI: 10.4324/9781003352174-7
 Patterns of Communication of Human Complex Societies 63

therefore transcended the context of time and space, becoming more and more
generalizable and potentially universal.
Besides this coevolutionary aspect between languages and the complexity of
societies, I am also exploring a methodological reversal: Can the laws that describe
social and collective behavior function as proxies for universal language patterns,
as well? Can we extrapolate from empirical laws and statistical regularities of the
social and collective phenomena features specific to the evolution of language, as
well? These are undoubtedly interesting questions, and answering them will help
us get an even better understanding with respect to what we can expect from an
alien language.

Universals of Human Languages are Coevolving With

Technology and Society
My argument is that intelligent communication is a coevolutionary process in
which innovation in artifactual communication (i.e., writing, the telegraph, the
Internet, etc.) and types of collective behavior that represents a civilization (i.e.,
infrastructure systems) intertwine and produce large patterns of communication,
such as transportation-system signals, brand logos (symbols of global organiza-
tions that transcend specific human languages), universal computer languages,
universal data-storage systems, data encoding and analyses, and many more. At
the same time, our actual languages incorporate these changes and become more
symbolic and more global (Finn 2017). Words related to technology are more likely
to be universal across languages (Chun et al. 2016).
We can prepare to understand the language of another intelligence by using
the best technological tools that our science and socio-technological advancements
provide. Technology is very likely to be a “universal” in the cosmos (Cohen and
Stewart 2002), meaning that there is a high probability that if another intelligence
willing to make contact exists, they will have technology and that their language
is technologically based. Some of these points are also addressed in more depth by
Ross in this volume (Chapter 12).
The current age of data and information in the history of our civilization is the
result of two major forces: the global spread of artifactual means of communica-
tion (i.e., the commonality of human-computer interactions, the globalization of
the English language) and the global spread of physical systems of recording and
storing communication (if we think about the long evolution from Mesopotamian
cuneiform tablets to current data-server farms and supercomputers). But while
information as means of communication is global, its meaning is still local. In case
of alien detection, we would similarly expect their language to have coevolved
with the social and technological evolution of that particular civilization and,
while the meaning of its information can still be highly contextual or parochial, the
means of communication might have universal features that we can connect with
and understand. A great such example is the fact that a German person who speaks
64 Anamaria Berea

only German can use the same computer as an English person who can speak only
English.
On Earth, technology has evolved as an extension of our intelligence (Hubbard
2015), while remaining constrained by our intelligence. There are socio-economic
evolutionary factors that triggered or constrained technological innovation (Arthur
2014), but at the same time technological innovation depends on natural laws as
well as on our discovery and understanding of these laws through science (McClel-
lan and Horn 2015). Therefore, it is likely that the technological advancement of
another civilization will depend both on their understanding of the universe (their
science), as well as on independent constraints that pertain to the evolution of
their own social and collective behavior (for example, a single, non-social entity
would pose an entirely different set of questions – could such an entity even need
language?).
This is why I believe that current technologies – particularly in computing and
information theory (artificial intelligence, cellular automata, quantum computing,
etc.) – are on track to give us a deeper understanding about universals in tech-
nologically based communication, or at least an understanding of the universal
laws of information theory (i.e., Shannon boundary, the physical limits of informa-
tion, etc.). At the same time, a deeper understanding of what is universal in the
coevolution of our collective behavior with the current technological and scien-
tific discoveries (i.e., information-based technologies seem to be common from
the Antikythera mechanism to modern computing) versus what is contextual (i.e.,
technologies that did not survive after a civilization crashed or disappeared, such
as mechanical technologies of ancient agriculture) will help us narrow down the
framework for xenolinguistics. In other words, an understanding of the evolution
of technology in our history, with an emphasis on what transcended cultures and
histories versus what did not, would help us frame the technology advancements in
terms of universals versus contextual/parochials.
We know that language has largely evolved organically and endogenously, and
with few exceptions (scholarly rules of linguistics imposed by the academies in a
few European countries), language is essentially an informational representation
of the society and culture where it was used (Goodenough 1981). Before digital
means of communication were invented, the economies of the world were local-
ized, embedded in the communities and the cultures of their specific societies (salt
in North Africa, silk in China, wine in France, tourism in Italy, etc.) and with large
variations in geography (Castells 1997). The embeddedness of language in culture
is also addressed in depth by Granger et al. in this volume (Chapter 8).
On another hand, we have designed informational systems that are universal to
some degree (e.g., mathematics, computer languages), but devoid of any meaning
or content, and therefore used only in small scientific and high-tech communities
(Berea 2018). One organic language that has become a standout among others due
to trade and the proto-globalization phase in our civilization is English, while in the
digital age, various computer languages are still competing for global hegemony
 Patterns of Communication of Human Complex Societies 65

and this effect may largely be due to the fact that computers cannot choose the lan-
guage they operate with (yet) or that computational challenges are evolving much
more rapidly than human skills for designing such languages (Berea 2018).
Another way to think about the language/civilization coevolution is to look at
the effect of different informational structures (networks vs. symbols vs. images
vs. text vs. probabilistic vs. Boolean vs. sound vs. any other form of information
representation) into the evolution and emergence of communication patterns, par-
ticularly linguistic ones, and to create a taxonomy of languages that includes both
the organically emerged and the designed communication (as well as any mix in
between) as a better representation of our civilization as a whole. In other words,
we can look deeper into how information representations are being used today and
classify them so that they are representative of all the complexities of a language
of an advanced civilization. These classes of information representation are easier
to generalize into patterns of language and communication in the context of a glo-
balized civilization (Raber and Budd 2003).
Globalized languages and large systems of communication are products of
selection and adaptation mechanisms that also emerged from the economic systems
of the growth of a civilization (McCann 2008). This seems intuitive. At the large
scale of our modern civilization on Earth, the planetary systems of interconnec-
tivity for transportation and communication show some of the artifactual, global
patterns of language of the largest magnitude we know to date. But data is global
only as a communication means: its meaning is local, as previously mentioned.
Examples of systems that use truly global means and meanings of communica-
tion are transportation infrastructures, such as aviation, maritime, and financial
systems, or Internet protocols. Therefore, these transportation systems and their
embedded communication are products of selection and adaptation mechanisms
emerged from the economic systems of our civilization. At the large scale of the
civilization, these planetary systems of interconnectivity show artifactual, global
patterns of signals/communication of the largest magnitude known to date. Ever
since cuneiform writing began in Mesopotamia to record bookkeeping for crops
and animals, thus with an economic functionality, the evolution of our civiliza-
tion’s economy and the emergence of artifactual communication have been inter-
twined and influencing each other.
This spurs another question we should address: To what degree can the currently
designed communication such as computer languages evolve organically (i.e., bots
inventing their own language) or become organic, endogenously evolving – and to
what degree does any type of organic communication become standardized (Eng-
lish words imports into other languages, Internet memes, etc.)? Similarly to the
classifications and taxonomies described here, a guiding map of past and current
communication patterns would help us understand the evolution of communication
in the near future. Will we invent a new type of human–computer language that
would change the current software language paradigm, which has standardized for
now the way humans interact digitally?
66 Anamaria Berea

Language (in the broadest sense) has deep effects on the selection and adapta-
tion of a species, whether this species is ants, dolphins, or humans, and the more
social a species is, the more important language is for its survival (Pinker 2003).
Even in non-human species on Earth, communication and “language” between
species has evolved deeply intertwined with the evolution of collective behavior,
networks, and communities of these species (McGregor 2005). But in the case of a
civilization in which the biological evolution standards of selection and adaptation
are more elusive and not yet fully understood (we don’t really know why we create
art or why we are creating technology that makes us more comfortable and less fit
for survival – whether everything we do as a species has a biological evolutionary
component), while the social evolution standards are more acute (we compete not
only for physical resources but also for networks of influence and power, whether
these networks are finance, firms, institutions, or friends), the language patterns
that will give more fitness to networks of influence would be the ones that take
advantage of both the organic and the designed patterns of communication.
More than languages, which deeply depend on localized communities, art and
symbols have evolved organically as efficient ways of global communication
before we had emoticons or social media (emotions and visual human behavior are
perceived as such in any place on the globe). Therefore, the methodological ques-
tion is whether we can integrate organic, global ways to store information (qualita-
tive visual data) with designed global ways to communicate (digital, mathematical
data) in a way that can become a universal standard for storing, understanding,
and communicating data and information. The big question I would like to see
answered is this: How will language evolve in the near future? The applied ques-
tion I would like to see answered is what will be the next unifying standard in digi-
tal communication? – is it a blend of visual symbols and data science techniques?
Answering these questions is valuable for two reasons: to improve our basic sci-
entific understanding (since we have not really bridged the evolution of signaling
in biological systems with the evolution of grammar in human languages and with
the evolution of computer language structures), and to provide economic insights,
because in the long run, it may advance the next innovations in global technologies
or communication while giving us an understanding of the clashes between the
physical economies and the informational economies. Such lags or rifts in coevo-
lution (Cabrol 2016) of large systems are important to understand, as they are also
the fertile ground for the next innovations in either languages or technologies, in a
somewhat similar fashion to biological mutations: a continuous feedback between
the physical and the informational aspects of an organism or a system of organisms.
Therefore, as there is also a lag between the human use of computer languages
and the human use of human languages, and a lag in the physical and informational
economy that has the potential to create both societal crises and opportunities
for large-scale innovations, societies that will continue to have this physical/
informational duality will also have communication systems that blend organically
with the organisms, the environment, and the informational/technological world.
 Patterns of Communication of Human Complex Societies 67

Identifable Patterns of Communication for Xenolinguistics

Life’s laws follow the laws of physics, but they have a higher degree of granular-
ity and variation (more “local” laws that fall into the “general” laws – i.e., cellular
laws do not violate atomic laws), and civilization’s laws follow life’s laws, with
an even higher degree of granularity and variation (even more “local” laws – i.e.,
macroeconomic convergence does not violate evolution). Therefore, communica-
tion – a phenomenon strictly dependent on living organisms – is life-dependent
and intelligent communication is civilization-dependent, which means that our best
way to find general patterns of communication in the universe is to look for gen-
eral patterns of communication in life and civilization on Earth. Additionally, we
need to understand the limits of variations in the laws of communication within
life from those within civilizations. Essentially, this is a data/information science
problem that involves cleaning, cataloging, mapping, and analyzing the evidence
of all patterns (not observations, which most probably would be impractical) of
communication that we know of today, but with an emphasis on “local” versus
“global” transitions in communication patterns.
Communication is as ubiquitous as information, yet while there is an integrated
discipline of information science with theories and applications, there is no inte-
grated discipline of communication science that studies not only “local” patterns
and observations of human communication or the broader phenomenon of commu-
nication in the living world. Therefore, this chapter is also a rationale for curating
and mapping our current but rich collection of communication evidence, scattered
throughout multiple disciplines and sciences into a taxonomy as mentioned previ-
ously, but with a focus on local/global patterns, on means/meaning of communica-
tion, and on natural/artifactual communication.
The vast majority of scientific disciplines study communication, from biology
(species-specific communication) to linguistics (human languages), to commu-
nication theory (context-specific human communication – i.e., parent–children,
political and social communication) to business and economics (by organization
specific communication). This domain-specific research on communication has
given us very rich details about specific, local communication patterns and the
role of communication in other phenomena (i.e., predation, mating, immunother-
apy, organizational functions, social movements, etc.) (Bradbury and Vehrencamp
1998). This means that we now have many pieces of a big puzzle to help us create
a comprehensive map of communication on Earth, based on taxonomy, ontology,
phylogeny, and evolution (Berea 2019). Having a broad idea about how communi-
cation has emerged and evolved in the living systems on our planet, from cells to
firms and societies, would enable us to highlight: 1) the common patterns across
species, ecosystems, and civilization trends, and 2) the differences/variations and
the boundaries/limits where we know that communication does not and cannot
occur. Given the large datasets and big data techniques available now from various
research projects and federal agencies, we can harvest and create a database of such
68 Anamaria Berea

evidence once the protocols for storing and realigning the data and meta-data are
set. Such a large classification endeavor will help create the protocols for aligning
and re-encoding the current evidence in a unified framework for what is possible
and what is not possible for a xenolanguage.
Following are some examples for focusing classification and categorization
efforts.

1 Locality, globalism, and transition phases in communication patterns. A taxon-

omy would place local patterns of communication of the living systems, from
cell-to-cell interaction, to organism-specific interaction, to human languages and
human–computer interactions, paying attention to transition phases or bounda-
ries between these phenomenon’s “localities,” within same paradigm or frame-
work of thinking. This paradigm is outlined by viewing communication through
the following lenses: natural/artifactual, meaning/means, and local/global.
2 Natural versus artifactual communication. While all living systems on Earth
communicate through natural, genetically evolved means (chemical commu-
nication in non-vertebrate species, ultrasound in bats, olfactory, auditory and
visual in most vertebrate species), only humans have developed artifactual
means of communication – writing, drawing, and computational technologies.
Kershenbaum’s contribution to this volume (Chapter 2) is dedicated to the
understanding of natural/animal forms of communication that can inform us
about xenolinguistics. Moreover, humans have developed complex languages
that combine natural (sounds) and artifactual (grammar) means of communi-
cation. The hallmark of an advanced civilization is its ability to combine the
natural and the artifactual means and meaning of communication in a way that
reflects selection and adaptation processes both in natural communication and in
artifactual one. Pepperberg deepens some of these ideas in this volume (Chapter
6), on human–non-human communication.
3 The “wave-particle” theory of communication: communication as both means
and meaning of information exchange. Shannon explored only the “wave”
theory of communication – communication as a means of exchanging infor-
mation, devoid of meaning (Shannon 1949). On another hand, semiotics and
cryptography explore the “particle” theory of communication – the meaning
of communication devoid of the means of communicating. Therefore, mapping
communication from cells to societies in distinct terms of means and meaning
of communication would help understand where the two aspects of the phenom-
enon of communication converge and diverge.

Based on the large, core ideas outlined in the preceding list, chronologically
and logically, I envision this endeavor of establishing a framework of research for
xenolinguistics in the following three parts.

1 Establish the current facts about large-pattern evolution. For example, visual
language started only with species that developed eyes or visual perception,
 Patterns of Communication of Human Complex Societies 69

artistic language emerged only with species that understood symbolism and
interpretation, language was developed due to grammar, and digital communi-
cation was developed only after both math and grammar were used, etc.
2 Create a model for categorizing and classifying language structures. Which lan-
guage patterns are networks of information? How much of language is physical
or informational representation? Are they data-rich structures? Are languages
structures or processes? Are they symbols or rules? After classifications based
on these types of meta-understanding of language in the context of communi-
cation and information, we can determine which would be the best model for
bridging different types of language representations (i.e., visual and artistic with
textual with 0/1 or mathematical representations).
3 Explore the future scenarios. Given such a meta-classifying model as a “good
enough” (never perfect) representation of the communication evolution for
humans and for computers, we can design scenarios that explore the future of
a linguistic standard. Alternatively, given that we have enough data about these
language patterns (computational semiotics), we can simply look at the com-
monalities versus differences in these patterns and assume that what they have
in common will persist (i.e., perhaps metaphors and leitmotifs) while the dif-
ferences will mutate and thus not able to become universal. An inventory of
communication standards in art (as they already exist, we would only need to
quantify them), in infrastructure systems, and in computer languages might be
sufficient to create a dataset of factors and attributes that we can analyze and
process.

Such a xenolinguistic framework has the potential to succeed if in the process

we end up learning something that we did not know before. Alternatively, the pro-
ject will succeed if we can establish a believable scenario about what the next
language our technology and/or civilization will speak. And alternatively again, the
project will still succeed if we end up establishing a clear link between the infor-
mational economy and the physical economy that can be tracked and understood in
terms of crises or growth.

The Role of Communication Pathologies for Xenolinguistics

In any scientific and scholarly endeavor, there is another side of the coin when we
approach the answer to a question or hypothesis. In this case, the other side of the
coin question is: Are there communication pathologies, and what can we learn from
them?
In general, in science, we aim to explain and understand the what, when, how,
who, when, and why of a phenomenon. We are less likely inclined to explain the
what does not, when it does not, how it does not, who does not, when not, and why
not of the phenomenon – in other words, to put the non-occurrence of the phenome-
non in the same epistemological terms of study as the occurrence of it. The payoffs
are higher and the search space is usually smaller for explaining a phenomenon than
70 Anamaria Berea

for explaining the not of same phenomenon. In the case of language, though, the
search space might be similar in size to the search space for non-language explain-
ing when there is no interspecies communication, why we cannot find meaning in
certain messages, why we do not have a unified language, why computers cannot
understand metaphors, etc. The rationale for such an approach in a framework of
patterns as thus described is that miscommunication is known to trigger cellular
pathologies in molecular biology, but also useful mutations, to cause wars in the
human civilization, but also to have languages converge and evolve. By the nature
of information (information negation creates new information; therefore, it is still
information), communication pathologies lead not to the death of communication
but to the evolution and transformation of communication. Linguistic pathologies
lead to a transformation of language and not the death of language. The framework
previously described will essentially help identify the boundaries in interspecies
communication, the life cycles of communication, and the transition phases – and
from these unrelated local communication failures, we can draw global patterns of
communication pathologies as useful as the global patterns of language and com-
munication described previously.

Methodologies for Pattern Identifcation

Another way to draw a useful, guiding framework for xenolinguistics is to explore
a methodological reversal, as I mentioned in the beginning: Can the laws that
describe social and collective behavior function as proxies for universal language
patterns, as well? Can we extrapolate from empirical laws and statistical regu-
larities of the social and collective phenomena into the evolution of language? For
example, we know that Zipf’s Law applies both to languages and to complex sys-
tems, such as cities or companies (Zipf 1949); we know that marginal (cost/benefit)
analysis in economics also applies to metabolism and chemical communication in
other species (Searcy and Nowicki 2005). What other laws from collective and
social behavior of animals, humans, and computers can we infer with respect to
their effect on language? For example, the “edge of chaos” model in complex sys-
tems behavior shows us where noise becomes signal or signal becomes noise, and
that the “edge of chaos” is where both the means and meaning of communication
happen (Kauffman and Johnsen 1991). The “edge of chaos” model is both global
and local for different complex systems phenomena; therefore, an “edge of chaos”
model of communication can potentially give us insights into the transition from
local to global communication patterns.
Another useful methodological approach would be to create a bridge between
semiotics and data science: an epistemological transition from locality to globalism.
In data science and machine learning, natural language processing techniques are
notoriously bad at identifying meaning in clusters of words and semantic distances
(Berea 2018). But they have the power to search for signals through large sets
of noise, as long as we identify these signals a priori – as long as we know what
 Patterns of Communication of Human Complex Societies 71

to search for. At the same time, computational semiotics techniques are better at
identifying signals in small data sets and reconfiguring algorithms based on seman-
tic identifiers. A computational tool that would combine natural language process-
ing with computational semiotics would be able to search through large datasets
of noise and adapt the signal search based on dynamic semantic identifiers, thus
enhancing the probability of identifying signals globally, not only locally.
A third approach involves patterns again. In this case, the patterns of common,
widespread communication function as signals of successful communication that
survived evolution, started locally and became global, as mentioned in the begin-
ning of this chapter. Classification and cataloging the current evidence of communi-
cation from cells to societies into an integrated view can today easily be done using
machine learning techniques, and the results would be referential for understanding
what can be classified as “common communication” in the living universe, what
possible scenarios of communication can be developed with small variations in the
living forms, and – most importantly – to prune out the scenarios or possibilities
of communication that can never lead to intelligent, decipherable communication.
While using patterns, classifications, catalogs and taxonomies are extremely
helpful for us to begin to understand the “space” and boundaries for how a xenol-
anguage could look like, they are also helpful for us to design and explore various
alternative scenarios. Particularly in the case of xenolinguistics and alien commu-
nication, for which we do not yet have evidence and validation protocols, as we do
in other sciences, scenarios and experiments are some of the best suited method-
ologies we can employ given we can use a robust database of patterns, identified
as explained in this chapter. Following are a few examples of such scenarios based
on patterning:

1 Scenarios of “relaxing”/varying the assumptions from the bottom up. Given that
we have a good map or guide about how our civilization actually communi-
cates globally, we can explore potential scenarios whereby we relax one or more
assumptions from this larger picture, such as: What if cells on a different world
do not communicate chemically? Would that imply different forms of organiza-
tion between higher organized intelligent species? What if organisms do not
develop sight or visual communication and cannot be recorded? How would a
civilization without visual records communicate? Does intelligent communica-
tion always depend on social or collective forms of behavior, or can there be
entities that do not connect with peers – that do not have a local, bounded com-
munication, but have only a global framework of communication?
2 Scenarios of nullifying assumptions from the top down. On the null hypothesis
side, does a civilization need an economic system in order to achieve global sys-
tems of commonly widespread, artifactual communication? Does a civilization
need an economic system in order to emerge general – not contextually local-
ized – communication? Humans have developed a diversity of forms of organ-
izing their social or economic or political behavior that cannot be reproduced
72 Anamaria Berea

or recognized in any other species. But civilization as we know it comprises

culture and social behavior that also gives us not only global communication,
but also ultra-specialized (large variation) communication in technology, art, or
economics that is not widespread and understood across the planet. We cannot
find both global and high-variance local communication in other living systems,
at the same time. Are these statistical patterns crucial for the development of a
civilization and its communication?
3 Scenarios of variation in the natural/artifactual coevolution (Simon 1996). The
interplay between the natural and the artifactual worlds through our economic
systems have determined, for better or for worse, the current state of our civili-
zation. Therefore, we can design scenarios whereby we can explore questions
such as the following. What types of economic systems with coevolving arti-
factual versus natural forms of organized behavior are the ones that can lead
to many types of sustainable civilizations? What types of civilizations that are
based on the natural/artifactual or physical/informational duality can develop
long-term, sustainable communication? What is the interplay between commu-
nication of a species as a collective (whole) and the many types of organizing
behavior (natural/artifactual communication topologies)?

These are only a few of the many scenarios and possibilities that can be explored
by formulating and exploring hypotheses, once a baseline or referential guide and
a unified dataset of evidence are in place. But in order to reduce the search space of
exploration with many “ifs” into “probables” and “possibles,” a unified referential
map would help guide these efforts.

Conclusions
In this chapter, I have tried to explore a few ideas with respect to determining a
framework for xenolinguistics or for identifying the possibilities and boundaries
of a xenolanguage. I am exploring the ideas of language in the context of commu-
nication and information that can take many shapes and forms, but that produces
potentially universal patterns only in the context of technological survival, globali-
zation, and transition from local to global and from contextual to universal. We
now have at hand methods we can use for classifying and categorizing or creating
taxonomies of languages and communication with a focus on these large historical-
scale processes of coevolution, adaptation, and selection that could help us better
understand what can be universal in terms of language and what are the boundaries
and contextualities that are specific to our planet and could not be found elsewhere.

References
Arthur, W.B. 2014. Complexity and the Economy. Oxford: Oxford University Press.
Berea, A. 2018. Emergence of Communication in Socio-Biological Networks. Dordrecht:
Springer International Publishing.
 Patterns of Communication of Human Complex Societies 73

Berea, A. 2019. A Complex Systems Perspective of Communication from Cells to Societies.

Norderstedt, Germany: BoD – Books on Demand.
Bradbury, J.W., and S.L. Vehrencamp. 1998. Principles of Animal Communication. Sunder-
land, MA: Sinauer Associates.
Cabrol, N.A. 2016. “Alien mindscapes—A perspective on the Search for Extraterrestrial
Intelligence.” Astrobiology 16 (9): 661–676.
Castells, M. 1997. “An introduction to the information age.” City 2 (7): 6–16.
Cohen, J.S., and I. Stewart. 2002. Evolving the Alien: The Science of Extraterrestial Life.
London: Ebury Press.
Chun, D., R. Kern, and B. Smith. 2016. “Technology in language use, language teaching,
and language learning.” The Modern Language Journal 100 (S1): 64–80.
Finn, E. 2017. What Algorithms Want: Imagination in the Age of Computing. MIT Press.
Goodenough, W.H. 1981. Culture, Language, and Society. San Francisco: Benjamin
Cummings.
Hubbard, B.M. 2015. Conscious Evolution: Awakening the Power of Our Social Potential.
Novato, CA: New World Library.
Kauffman, S.A., and S. Johnsen. 1991. “Coevolution to the edge of chaos: Coupled fitness
landscapes, poised states, and coevolutionary avalanches.” Journal of Theoretical Biol-
ogy 149 (4): 467–505.
McCann, P. 2008. “Globalization and economic geography: The world is curved, not flat.”
Cambridge Journal of Regions, Economy and Society 1 (3): 351–370.
McClellan III, J.E., and H. Dorn. 2015. Science and Technology in World History: An Intro-
duction. Baltimore: John Hopkins University Press.
McGregor, P.K. ed. 2005. Animal Communication Networks. Cambridge: Cambridge Uni-
versity Press.
Pinker, S. 2003. “Language as an adaptation to the cognitive niche.” Studies in the Evolution
of Language 3: 16–37.
Raber, D., and J.M. Budd. 2003. “Information as sign: Semiotics and information science.”
Journal of Documentation 59 (5): 507–522.
Searcy, W.A., and S. Nowicki. 2005. The Evolution of Animal Communication: Reliability
and Deception in Signaling Systems. Princeton, NJ: Princeton University Press.
Simon, Herbert A. 1996. The sciences of the artificial, 3rd ed. Cambridge, MA: MIT Press.
Shannon, C.E. 1949. “Communication theory of secrecy systems.” Bell System Technical
Journal 28 (4): 656–715.
Zipf, G.K. 1949. Human Behaviour and the Principle of Least Effort. Boston: Addison-
Wesley Publications.
8
INTERSTELLAR COMPETENCE
Applications of Linguistics and Communicative
and Cultural Competencies to Extraterrestrial
Communication

Sumayya K.R. Granger, Judd Ethan Ruggill, and Ken S.

McAllister

Introduction
Part of what makes xenolinguistics so challenging is that virtually nothing can be
known – or even safely assumed – about the non-human elements of first contact.
Indeed, a good portion of humankind’s eventual and thrilling initial exchanges with
extraterrestrials are likely to be baffling. Foreknowledge of this dilemma is what
has given rise to the planet’s current policies on extraterrestrial contact, most (if
not all) of which are consistently confined to post-detection/disclosure protocols
(e.g., the Rio and San Marino scales); such efforts’ operative premise is “Human
beings cannot know what extraterrestrials are going to say and do; however, humans
can determine in advance their own range of responses.”1 We agree that there will
be many unknowns at first contact, but we contend that the context for that interac-
tion is not necessarily one of them, even considering the non-Homo sapiens side
of the encounter. As specialists in intercultural competence and transdisciplinary
communication and translation, we have often witnessed the process whereby
strangers learn how to communicate with each other: economists and poets, chil-
dren and adults, animals and humans, management and workers, Nigerians and
Germans, progressives and conservatives, academics and office workers, farmers
and urbanites. Without question, the fact that human beings are at the center of all
but one of these discursive conjunctions means that sense-making within them is
a good deal easier than if the interlocutors were, for instance, humans and beings
from exoplanet Kepler-62f. That said, there are a number of useful communication
features that can usually be assumed to be true in any new encounter among intel-
ligent species.2
This chapter begins by identifying a core set of these assumptions, then builds
from them to outline a prolegomenon to alien communication. Along the way, we

DOI: 10.4324/9781003352174-8
 Interstellar Competence 75

attend to several essential components of the xenolinguistic complex: establishing

and maintaining regular contact; the ways in which the constraints of remote con-
tact sketch a roadmap toward (and ethical considerations about) the development
of an adaptive process of learning, understanding, and communication; and design-
ing a working plan for bridging pre-physical and physical contact communication
scenarios. We begin with a discussion of what tool use says about culture and
communication.

Every Tool a Symbol

Regardless of the physical stature or interpretive frame of an alien interlocutor,
we contend that at first contact, all parties will share at least five developmental
features: tool use, symbol use, communication, culture-making, and curiosity.3 The
fact that all parties will be tool users is self-evident; to send and receive interstel-
lar communications strongly suggests a culture that is able to design and build
complex devices. The very existence of such tools also suggests a culture capable
of metonymy – that is, seeing the whole in the part (e.g., the symbol “interstel-
lar transmitter” stands for “planet teeming with intelligent life”). Such complex
symbol production and consumption indicates a communicative species – one that,
given both its inward- and outward-facing communication practices, must also
practice the making of culture. Finally, whatever motives might stand behind any
interstellar messaging (e.g., wonder, expansion, catastrophe), such messages imply
an intelligence that is both curious and expects surprise. Taken together, these fea-
tures clarify the opacity of first contact just enough for us to begin piecing together
a xenolinguistic approach that is both practical and adaptable – core elements in
any successful intercultural exchange.4
For the purpose of context, consider humanity’s brief history of deliberately
attempting to send and receive extraterrestrial messages. The primary goal of send-
ing such signals into space, of course, has been to initiate an intercultural exchange.
Practically, this means: (1) establishing contact, (2) knowing that contact has
occurred, and (3) maintaining contact. It has been more than half a century since
humans sent the “MIR, LENIN, SSSR” messages in Morse code to Venus, and there
has not yet been a recognizable reply.5 It is also unclear if any of the numerous sub-
sequent interstellar messages humans have sent were ever received, or if they were,
what was made of them. Finally, there is the possibility that extraterrestrial beings
have attempted to contact Earth (either in response to the aforementioned messages
or in their own search for distant neighbours), but that their communications have
so far proven undetectable.6 In other words, human attempts at interstellar cultural
exchange do not appear to be advancing quickly, even with innovations such as
SETI’s collaboration with IBM and its Spark application (a project specifically
designed to discover communicative needles in the cosmic haystack).7
Such slow progress, however, is axiomatic for xenolinguistics. The nearest
potentially inhabited planets are many light years away, creating a challenging
76 Sumayya K.R. Granger, Judd Ethan Ruggill, and Ken S. McAllister

hyper-asynchronous communication paradigm. Consider, for instance, the “Cos-

mic Call” message sent to 16 Cyg in 1999.8 Though it has been travelling for almost
two decades, it is still more than half a century from its destination. Even if the
message were received and replied to immediately, Earthlings would not be aware
of achieving Phase 2 of intercultural exchange (“knowing contact has occurred”)
until 2138 at the earliest, more than five generations removed from the original
sender, Charlie Chafer. For this reason (among others), Seth Shostak’s argument
for transmitting “short messages sequentially, but repeatedly” toward “a million
star systems once a day” (Shostak 2011: 367) is compelling. Using 1-micron infra-
red signals set at 1 bit per pulse, such “pings” could daily “convey the equivalent
of the Encyclopedia Britannica to each of these systems.” This would not be a
lot of information relative to the human timeline, but perhaps it would be enough
for an equally advanced species to decode and eventually master the transmis-
sion language. While such a process would still require significant patience and
commitment to extraterrestrial messaging on the part of humanity, this kind of
shotgun approach has the advantage of extending a high number of conversational
invitations, as well as maximizing humanity’s chances to have multiple concurrent
interstellar exchanges.
We emphasize the time scale and the complexity of sending, receiving, and
decoding messages because they speak directly to the five shared developmental
features among all first-contact partners. Specifically, while there are many hypoth-
eses about what makes the best first contact message – math, science, photographs,
or descriptions of humanity’s “exquisite balance of joy and sorrow,” as Douglas A.
Vakoch has proposed (Sutton 2015) – most likely the central transfixing element to
interstellar receivers of initial outbound communications will be the tool of mes-
sage transmission itself. As will be the case when humans receive their own first
interstellar message, an anomaly in the fabric of deep space signals will be detected
first. When a pattern emerges, symbols will be discovered if not decoded. When
the symbols become recursive and systematic, communication may be presumed
among the senders – and perhaps one day, between senders and receivers. Fantasies
of culture will not be far behind, largely devoid of facts and context.
Curiosity will drive the exchange forward, confusion will slow it down, and
understanding will deepen both the exchange and the emerging relationship. Yet in
the beginning, the question that receivers of an alien message will necessarily focus
on is not, “Who are these beings?” or even “What are these symbols?” Rather, it
will be “What is this signal anomaly, and how do we make sense of it?” Marshall
McLuhan’s observation that “the medium is the message” (McLuhan 1964: 7) will
never be truer than upon first contact.

The Importance of Error Correction

The fact that extraterrestrial linguistic interaction will likely entail extended time
frames in the beginning, at least – and possibly multiple, concurrent, and extended
 Interstellar Competence 77

time frames if humans establish regular contact with multiple interstellar intel-
ligences – means that confusion and error will have a spacious petri dish within
which to grow before any clarification can be received. From an Earthly context,
this portends nothing good.9 Given this likelihood, we suggest that xenolinguistics
must not only continue coming to terms with the cunctatious nature of its practice
(as this volume is doing), but also institutionalize a respect for this nature in its
policies and professional practices. To not do so would be to give tacit approval for
rushed xenolinguistic activities, the consequences of which could be devastating
on a planetary scale. In this section, we consider what the institutionalization of a
fastidious xenolinguistics might look like, as well as offer an explanation for why
patience – already the watchword for intercultural exchange on Earth – must be
doubly mandated for interstellar cultural exchange.
To begin, because interstellar messages will not enjoy the benefit of being
exchanged in the same physical space, they will face hyper-asynchronicity (at least
from the human point of view), and will likely be truly attended to only after the
receivers have reverse engineered the communication technology used to encode
and send the message. That is why humans need to think strategically about what
information can be meaningfully conveyed in any potential first-contact message.
As we assess this situation, there are three primary factors to consider: (1) medium,
(2) message robustness, and (3) degree of mutual intelligibility.
The first of these factors, medium, is simply an admonition to recognize that
before the content of a message can be decoded, it must first be discovered, stored,
and studied. Light-based signals may seem fast and efficient to most humans, for
example, but they will not be detectable by beings who lack the ability to appre-
hend and process them.10 Because the medium is the message carrier, interstel-
lar communicators are well-advised to deploy media that offer a wide spectrum
of communicative formats, from visible and invisible (to humans) light, to sound
waves, to physical objects designed to be touched, looked at, and/or listened to.
Second, tempting as it will be (and has been) to tell interstellar beings about the
human animal and its cultures, discerning a recognizable pattern in such foreign
data is likely to be exceptionally difficult. In order to increase the robustness of
first-contact communication, therefore, a programme of discoverable mirroring is
advisable. In such a programme, signals would be designed specifically to stand
out from all other ambient space noise (discoverable), and would mirror informa-
tion that is already familiar and meaningful to the intended recipient (e.g., orbital
patterns or star charts of the recipients’ solar system). By giving first-contact part-
ners something (presumably) easy to recognize, humans will simultaneously create
a differential signal (as compared to the “noise” of deep space), convey Earthly
knowledge of the partners’ general location, and provide a template for an equally
legible reply. The exchange could then proceed along similar lines of comparison
and identification, with each modest change signalling an intelligent and attentive
interlocutor at the other of end of the pathway. Such an intercultural exchange strat-
egy – part baby talk, part “cocktail party effect,” and part interpellation (or “hail,”
78 Sumayya K.R. Granger, Judd Ethan Ruggill, and Ken S. McAllister

as Louis Althusser [Althusser 2001 [1970]: 86] terms it) – could prove a remarkably
efficient way to catalyze a productive (if protracted) conversation. As Daniel Ross
(this volume [Chapter 12]) so elegantly concludes, human beings “should prepare
to teach as much as to learn,” a point that Pepperberg (this volume [Chapter 6])
also makes and expands upon.
Once a minimal form of contact and awareness has been established, including
a working knowledge of what Denise L. Herzing (this volume [Chapter 4]) terms
“metadata,” an important next objective would be to increase contact quality.11 By
working to share – whether by signals or tangible goods – increasingly diverse
and meaningful information, the interstellar relationship would grow, as would the
channel of communication. That is, we predict that, as among humans, the com-
munication channel itself would gradually transform to maximally accommodate
the kinds of messages its users want it to carry (e.g., the telephone). Failure to
establish such robust contact will, again as with humans, likely cause the stagna-
tion of the communication channel, effectively arresting the relationship’s develop-
ment. Detected sequences and patterns from space may well signal intelligence, but
robust communication requires that curiosity never be entirely frustrated or sated.
Such an unfortunate communicative plateau would be akin to castaways waving to
each other from separate islands, aware of each other’s existence but never able to
know more.
Finally, at each one of these communication building stages, there must be a
vigorous effort to discover and correct errors of translation, interpretation, and
cordiality.12 Another common practice for Earth-based intercultural exchange
that would seemingly transfer well to an interstellar variety is the practice of
collaboratively developing a set of ground rules according to which all par-
ties abide. These ground rules might include agreements about which words,
topics, actions, and objects are off-limits, as well as practices that are encour-
aged, such as “When you are offended, make this clear to the offender and per-
mit an opportunity for apology and clarification.” In our experience, it is rarely
too late to apologize and make amends, and doing so can often have the effect
of strengthening a relationship by building trust. Given the fermentation time
between messages in a xenolinguistic context, “rapid” error correction will be
imperative.
As more significant conversation becomes possible, opportunities to connect
and mutually develop will multiply. Concomitantly, establishing a mutually intelli-
gible mode of communication will become increasingly important. In the next sec-
tion, we explore how this process might evolve, particularly in a scenario wherein
extraterrestrial interlocutors are similar enough to humans that consistent inter-
course is possible, but from whom humans differ significantly in terms of physical
form and communication apparatuses. Along with these guidelines for co-creating
a common language given the constraint of divergent (physical) morphology, we
also address the compounding complications that hyper-asynchronicity adds to the
development of a hybrid language corpus.
 Interstellar Competence 79

Mutual Intelligibility: Building a Common Language

In addition to considering interspecies similarity, vigilant mindfulness of – and
respect for – difference is essential to xenolinguistics. Humans, for example, are
routinely confounded by each others’ messages: “Did he mean what he said . . . or
was he kidding?”; “Does she not understand I don’t speak g33k?”; “I can’t even
pronounce that language, let alone understand it.” Interpretive gaps grow more
profound beyond the species – think of the communication challenges between
humans and terrestrial flora, for example.13 Researchers only recently discovered
that beech trees are able to communicate to nearby relatives using chemical sig-
nals, alerting neighbours that they are becoming malnourished, for instance. Using
fungi in their roots, surrounding beeches are then able to alter how soil nutrients
are shared, optimizing the health of the entire stand of trees (Wohlleben 2016).
Given the challenges of interspecies communication in and among Earth’s biomes,
it is reasonable to surmise that interplanetary lifeform communication may well be
even more formidable.14
For one thing, the recipients of humans’ first communiques may exist in an
entirely different corporeal state (e.g., non-solid), may communicate through sen-
sory organs humans do not possess, or have unfathomable body chemistries. How
are humans to guess what kinds of messages will be meaningful? How does one
build a communication bridge with no knowledge of what is on the other side of the
chasm?15 Two fundamental building blocks of intercultural communication theory
provide tentative answers: first, be systematic, creative, and (to the best of one’s
ability) empathic about the kinds of tools that could be deployed to understand
who the other communicant is, how they communicate, and what they communi-
cate about. Second, humans need to make it as easy as possible for interlocutors
to establish the same communicative building blocks. Put simply, humans must
remember that intercultural exchange, on Earth or in the stars, is always about try-
ing to solve two problems simultaneously: (1) What do they mean? and (2) How
can we help them understand what we mean?
We have already discussed, to a limited degree, how the first of these problems
could be addressed; doubtless, other scholars in this volume will have more spe-
cific ideas about such work. Identifying instances of shared knowledge and experi-
ence, using high-performance computing to search for language patterns (first in
ambient space, then in known but untranslated messages), and developing a regime
of patience among all interlocutors are certainly among the most immediate appli-
cable strategies.
The second problem – helping alien contacts understand human messaging –
is arguably what humans have tried to do in the variety of interstellar messages
that have already been sent. By illustrating the human form and its functions,
representing Earth’s galactic location, cataloguing geophysical details ranging
from the subatomic to the planetary, and so on, humans have aimed to say “this
is who we are.” Again, without a functional equivalent for the referents in these
80 Sumayya K.R. Granger, Judd Ethan Ruggill, and Ken S. McAllister

offerings, however, such messages are likely to remain opaque. Given that humans
and extraterrestrials have stellar phenomena in common, at least – nebulae, suns,
black holes, gravity, radiation – these will likely be reliable shared analogues for
early exchanges. Even if terrestrials and extraterrestrials inhabit the universe in
very different ways, mutual experience with the physical world should offer pos-
sibilities for quickly (relatively speaking) establishing rich communication on a
variety of topical entry points. By first identifying as many shared data points as
possible, we believe that, as Slobodchikoff (this volume [Chapter 9]) proposes, a
Rosetta Stone of sorts can be established from which other details can be confi-
dently linked by association.

Top Down or Bottom Up?

While a number of interstellar messages sent to date have emphasized representa-
tions of the atomic and subatomic contexts, and while such contexts are presumably
universal, they violate the second rule of intercultural exchange we proposed in this
chapter; namely, do everything possible to help message receivers understand the
nature of the message itself (i.e., its semantics). Such messages assume that alien
interlocutors will, like humans, be unable to directly experience the activities of
single atoms or even single molecules, yet given universal laws of gravitation and
mass, that seems a safe assumption given that such beings will (like humans) be too
large to be cognizant of such microdynamics. For this reason, we recommend that
initial exchanges emphasize the largest, most unsubtle shared phenomenon, pref-
erably using a combination of nonlinguistic (e.g., diagrams, models, sounds) and
linguistic representations. Once a shared conceptual and terminological vocabu-
lary has begun to emerge, both parties can begin to make associative steps toward
greater levels of abstraction and more subtle sensory experiences.
We also believe that once the paradigm of deductive reasoning has become prac-
ticed, avenues for inductive and abductive reasoning will subsequently emerge.
Written language – including mathematical symbols, formulae, constants, and
curiosities such as Fibonacci sequences (cf. Sutton 2015; Medeiros 2017) – can
be gradually introduced, again, always remaining mindful of how best to make
information (especially new information) comprehensible to others.
Another communicative approach we recommend – one that admittedly requires
the (even slower) exchange of objects, as well as messages – is the creation of
associative information packets (AIPs), physical containers holding a number of
labelled and connected physical specimens. Consider, for example, a carbon AIP
containing: (1) a sample of graphite, (2) a sample of diamond, (3) diagrams of the
molecular and crystalline structures of both materials, (4) a diagram of the atomic
weight of each, (5) the weight of each of these items on Earth, and (6) the mass of
each of these items. Such a packet, designed as it is around the premise of inter-
connectedness, creates a much smaller solution set to the question “what connects
these things” than if they were to appear as discrete items and/or data points. Such
 Interstellar Competence 81

AIPs can be developed for virtually any configuration of information, from genet-
ics to political, social, and family structures to popular culture.
As basic communication between humans and extraterrestrials unfolds through
the identification of shared knowledge and experience, a sensible next step would
be the construction of a Swadesh list or core vocabulary list (CVL), a register
of words with “common meanings, such as ‘mother’, ‘lake’, ‘mountain’, ‘three’,
‘red’, ‘green’, ‘to vomit’, ‘to kill’, ‘dirty’, and ‘dull’” (Calude and Pagel 2011:
1102). Using this common vocabulary – and the shared concepts and phenomena
they represent – humans and aliens would then be able to begin revisiting past con-
versations, creating more intuitive exchanges and assigning key semiotic elements
within them to appropriate words drawn from the CVL. Proceeding in this way, the
initial ad hoc Rosetta Stone would become a consistent and nuanced dictionary.
To imagine what such a list might look like, it is first necessary to consider
what forms of communication would be successful soon after first contact. If, for
instance, the early signals were light-based (e.g., infrared bursts), then humans
would want to consider using (or at least mimicking) these same signals for cre-
ating a preliminary CVL. If early messages were acoustic, the CVL would form
around that mode. The aim of mediumic consistency here is both to expedite com-
munication by keeping all exchanges within a narrowly focused communicative
paradigm, as well as to (hopefully) provide an advantage to interlocutors who –
given their initial ability to receive, identify, and respond to humans’ initial sig-
nals – have tacitly indicated a preference for the exchange’s underlying format.
Moreover, because humans must, as much as possible, remove their linguistic
biases from the exchanges, it would be of paramount importance to create a CVL
that represented the look, feel, and/or sound of the communicants’ native language.
Proceeding slowly and patiently, we believe that in the para-contact moment (e.g.,
the first two millennia), humans should deprioritize attempts to communicate using
full language, opting instead for solidification of an accurate CVL, labelling system
and set of rules by which one idea, object, or concept may be connected to another.
A slow, methodical approach such as this diminishes the danger of what W.V.O.
Quine characterizes as the “gavagai” problem (1969: 32). This problem arises
when a speaker points to an object and utters an unfamiliar word – “rabbit” and
“gavagai” in Quine’s example – and confused listeners must somehow determine
what “gavagai” refers to: the rabbit itself? A running rabbit? A living rabbit? A
brown rabbit? A young rabbit? The rabbit’s legs? The more methodical the process
by which the CVL is assembled, the less likely it will be that subsequent exchanges
will be undermined by communicative misfires. Such patient efficiency is particu-
larly important when messages are being passed over interstellar distances. Despite
what we see in so many science fiction movies and television shows, the experi-
ence of fluent interspecial communication seems highly unlikely until many mil-
lennia after first contact. We suspect, in fact, that communications will plateau for
a long period, perhaps even indefinitely. During this phase, a best-case scenario
might well be that there is a consistent “A1” fluency level, defined by The Common
82 Sumayya K.R. Granger, Judd Ethan Ruggill, and Ken S. McAllister

European Framework of Reference (CEFR) as a level of linguistic competence in

which one:

Can understand and use familiar everyday expressions and very basic phrases
aimed at the satisfaction of needs of a concrete type. Can introduce him/herself
and others and can ask and answer questions about personal details such as
where he/she lives, people he/she knows and things he/she has. Can interact in
a simple way provided the other person talks slowly and clearly and is prepared
to help.
(Council of Europe 2018: n.p.)

Such communications with beings from another world – despite being “very basic” –
promise to be extraordinarily thrilling nonetheless. They have the added advan-
tage of minimizing the chances of misunderstandings (though these are still bound
to occur), while at the same time building confidence in both the mechanisms of
exchange and among interlocutors themselves. Over time, they also promise to lay
a solid foundation for increasingly complex exchanges about ever more sophisti-
cated topics.

What Does It Mean to Be Me?

It is inevitable that as exchanges develop, all parties will begin to discern valua-
tions embedded in topic choices, the CVL, and any other elements of the commu-
nication stream. Fundamentally, such assessments hinge on judgements about what
is good and bad, and consequently entangle not miscommunications but actual
disagreements. This is a highly abstract discursive realm, to put it mildly, and one
that humans themselves often disagree about (cf. Sutton 2015). When combined
with the problematics of nascent linguistic aptitude, xenophobia, and hyper-asyn-
chronicity, a true conundrum is born: how can human beings effectively navigate
contention with a vastly different species that is very far away?
One solution is to use a variation of Optimality Theory, an approach to pho-
nological theory whereby constraints within a language’s sounds compete for
the “best” outcome. In the context of human language, Optimality Theory helps
explain variations among languages (e.g., between Spanish and English) in terms
of what linguistic constraints are in place when spoken language is expressed. For
instance, Alan Prince and Paul Smolensky (1991) and John McCarthy and Alan
Prince (1993) discuss two major types of linguistic constraint – that is, constraints
present within a variety of a language that shape how it is spoken (how the surface
forms produced vary from the underlying forms from which they are derived).
The first of these constraints is faithfulness of the surface form to the underlying
form (i.e., do a word’s underlying constituent parts logically add up to its accreted
appearance?), while the second is markedness (i.e., the extent to which a word is
“well-formed” by syllables of certain shapes and sounds). As languages – human
 Interstellar Competence 83

TABLE 8.1 Illustration of outcomes of competing constraints

/in-/ + /mediate/ Make the result easy to Keep the surface form similar
pronounce to the underlying form

Option 1: [inmediate] No Yes

Option 2: [immediate] (winner) Yes No

ones, at least – evolve, faithfulness and markedness often compete, and the win-
ning option in any given language provides evidence for what qualities of linguistic
expressions are favoured by speakers of the language.
Consider, for instance, the word immediate, formed from the negating prefix /
in-/, plus the root /mediate/; literally, the word means “not mediated,” that is, “right
away.” Yet in English, the word is not “inmediate” but “immediate.” Most simply,
Optimality Theory would represent this phenomenon thusly:
Evident here is that English seems to prioritize word formations that are easier
to pronounce (two adjacent labial sounds, rather than an alveolar immediately fol-
lowed by a labial), over words that clearly reveal their underlying phonemes in
their surface forms. Charting the same word in Spanish, however, reveals the oppo-
site preference: inmediato.
While Optimality Theory is concerned with phonological forms, it is also impor-
tant to consider the possibility of socio-cultural applications of the theory. Geert
Hofstede (2001), for example, proposes that human cultures can be understood as
varying along different continua (e.g., individualism vs. collectivism; low-power
distance vs. high-power distance). At the core, this means focusing on how humans
in groups interact with one another. This kind of Optimality Theoretic approach
might be helpful in conveying to interstellar interlocutors the variance in human
cultures, but variance according to a finite set of parameters. Structuring dialogue
on complex and abstract topics such as culture in this way could be key to avoiding
conflict and extending discussion.

The Danger of Oversharing

We have one last recommendation to offer xenolinguistics from the field of inter-
cultural competence. An underappreciated aspect of negotiating intercultural
exchange is knowing how much to say about what. For reasons that science fic-
tion has elaborated extensively – from weaponized modifications of the human
genome to exploitations of military systems to manipulations of the human psyche
– the indiscriminate sharing of information about life on Earth is highly inadvis-
able.16 This is so, not for fear of planetary conquest but because even small socio-
cultural differences can have dramatic effects on how relationships do or do not
develop. This is another reason to keep initial exchanges relatively neutral, focus-
ing on factual observations about interstellar space until communicative pathways
84 Sumayya K.R. Granger, Judd Ethan Ruggill, and Ken S. McAllister

are sufficiently robust to handle the discovery and processing of more freighted
cultural and personal expressions of meaning. At the same time, the whole proposi-
tion of entering into contact with other lifeforms presupposes a certain openness, at
least on humanity’s part. Despite how vastly different humans may be from beings
beyond the stars, failing to make unflagging efforts to understand and be under-
stood – and doing so thoughtfully, patiently, and generously – will significantly
diminish the probability of developing a persistent and rewarding communication
channel.

Final Thoughts
In this chapter, we have drawn on work from a number of interrelated fields to out-
line an approach to a (highly probable) hyper-asynchronous, physically removed
first-contact scenario. On the chance that first contact is face-to-face, we still recom-
mend the set of practices discussed here, though with appropriate modifications to
accommodate the change from an asynchronous to a synchronous communication
mode. In such a case, a patient, methodical approach to every exchange is essential.
Particularly in the para-contact period – when trust is fragile, anxiety and eagerness
are uneasy emotional partners, and the ground of understanding is, at best, precari-
ous – all parties must remember how volatile new relationships can be. Whether
built near or distantly, communication styles and strategies must be readily adjust-
able based on whatever feedback is discernible. If humans are hindering rather than
helping their interlocutors understand and be understood, then a re-evaluation of
all applicable communication strategies is required. And at every turn, interlocu-
tors must be slow to assume and quick to forgive. A seemingly delightful similarity
might quickly unravel, while an unfathomable difference might snap into clarity
when its context is fully revealed. As a useful reminder, humans will do themselves
a favour by remembering the vast differences even among human cultures, where
miscommunications and dangerous assumptions often yield considerable strife.
Extraterrestrials will be at least as different from humans as human cultures are
from each other. And should interstellar conversationalists make the same mistake
toward humans – unwarranted assumptions; misunderstanding signals; expressing
voluble surprise or dismay at human values, beliefs, and practices; or otherwise
getting the wrong end of the stick – humankind must its redouble efforts to work
patiently to build bridges and cultivate understanding.
We proposed at the start of this chapter that despite the lack of concrete
knowledge of first-contact beings, human beings are not at a complete loss for
sensible approaches to first-contact exchanges. Indeed, we have suggested that
there are at least five important traits humans will share with any alien interlocu-
tor: tool use, symbol use, communication, culture-making, and curiosity. In our
estimation, whether Earthbound or starbound, it is this last feature in particular
that will keep the conversations vital, edifying, and always open to new ideas
and friendships.
 Interstellar Competence 85

Notes
1 Hobaiter et al. (this volume [Chapter 3]) use a Gricean approach to provide a compelling
perspective on assumptions such as these, problematizing what they might mean for the
detection of “intentional communications.”
2 As Daniel Ross (this volume [Chapter 12]) notes, “sufficiently intelligent” is perhaps a
more accurate way to denote the intelligence(s) in question.
3 It is worth noting that communication and culture-making are relevant only for spe-
cies of n > 1. Mono-entities would not necessarily develop communicative or cultural
systems.
4 For precis on the Rio and San Marino scales, see https://iaaseti.org/en/rio-scale/ and
www.setileague.org/iaaseti/smiscale.htm, respectively.
5 For a brief introduction to this project, see https://en.wikipedia.org/wiki/
Morse_Message_(1962).
6 Daniel Ross (this volume [Chapter 12]) astutely points out that human beings may not
have the cognitive ability “for receiving (via language), retaining, and evaluating extra-
terrestrial thoughts, which might be more complex than our own.”
7 See www-01.ibm.com/software/ebusiness/jstart/portfolio/seti.html for more about this
collaboration.
8 For a primer on the Cosmic Call, see www.smithsonianmag.com/science-nature/
annotated-cosmic-call-primer-180960566/.
9 Con Slobodchikoff (this volume [Chapter 9]) makes a similar point, noting that it is not
only spoken language that humans have trouble with – even among native speakers of
the same language – but also written languages and the languages of mathematics and
geometry. The probability of communicative misfires even among our own species thus
makes the possibility of reliable intergalactic communication extremely poor.
10 As Jessie S. Nixon and Fabian Tomaschek (this volume [Chapter 17]) explain, and as
Bridget D. Samuels and Jeffrey Punske (this volume [Chapter 16]) note, even slight
physiological differences in anatomy – differences which may be environmentally
caused – can greatly impact the ability to communicate in comprehensive ways.
11 See Harbour (this volume [Chapter 18]) on the Visible Speech writing system.
12 See Ortner (this volume [Chapter 5]) on interactional calibration.
13 Arik Kershenbaum (this volume [Chapter 2]), offers a survey of the astonishing breadth
of communicative strategies already deployed on Earth, clarifying in the process how
numberless are the possibilities in a universal context.
14 While in this chapter we make a best effort to get “outside of ourselves,” our approach
will necessarily be anthropocentric given the nature of our physical constraints. This is,
indeed, the core challenge of a book about xenolinguistics. Nevertheless, we hope that
the proposals here are abstract or abstractable enough to be of value.
15 See Wells-Jensen (this volume [Chapter 13]) on questioning unconscious assumptions.
16 The precautionary approach recommended here is generally accepted among policy ana-
lysts interested in intelligent interstellar communication. Admittedly driven by a self-
reflexive assessment of how humans treat each other, such an approach considers factors
ranging from local diplomacy to global security. See, for example, Matthews (2019),
Todd and Miller (2018), Wilman and Newman (2018), and Wilson and Cleland (2019).

References
Althusser, Louis. 2001 [1970]. “Ideology and ideological state apparatuses (Notes toward an
investigation).” In Lenin and Philosophy and Other Essays. Translated by Ben Brewster.
New York: Monthly Review Press.
86 Sumayya K.R. Granger, Judd Ethan Ruggill, and Ken S. McAllister

Calude, Andreea S., and Mark Pagel. 2011. “How do we use language? Shared patterns
in the frequency of word use across 17 world languages.” Philosophical Transactions:
Biological Sciences 366 (1567): 1101–1107.
Council of Europe. 2018. “Global scale-table 1 (CEFR 3.3): Common reference levels.”
Common European Framework of Reference for Languages (CEFR) Website, n.d.
www.coe.int/en/web/common-european-framework-reference-languages/table-1-cefr-
3.3-common-reference-levels-global-scale.
Hofstede, Geert. 2001. Culture’s Consequences: Comparing Values, Behaviors, Institutions,
and Organizations across Nations. London: Sage.
Matthews, Luke J. 2019. “A moonshot for extraterrestrial communication.” An-
thropology News, July 12. www.anthropology-news.org/index.php/2019/07/12/
a-moonshot-for-extraterrestrial-communication/.
McCarthy, John, and Alan Prince. 1993. “Prosodic morphology: Constraint interaction and
satisfaction.” Linguistics Department Faculty Publication Series 14. https://scholar-
works.umass.edu/linguist_faculty_pubs/14.
McLuhan, Marshall. 1964. Understanding Media: The Extensions of Man. New York:
McGraw-Hill.
Medeiros, David P. 2017. “Fibonacci and L-grammars.” David P. Medeiros Website, n.d.
http://davidpmedeiros.com/fibonacci-l-grammars.
Prince, Alan, and Paul Smolensky. 1991. “Connectionism and harmony theory in linguis-
tics.” Report no. CU-CS-533–91: Department of Computer Science, University of Colo-
rado, Boulder.
Quine, W.V.O. 1969. Ontological Relativity and Other Essays. New York: Columbia Uni-
versity Press.
Shostak, Seth. 2011. “Limits on interstellar messages.” Acta Astronautica 68: 366–371.
Sutton, Jon. 2015. “Interview [with Douglas Vakoch]: What would you say to an alien?” The
Psychologist: The Website of the British Psychological Association 28: 800–803.
Todd, Peter M., and Geoffrey F. Miller. 2018. “The evolutionary psychology of extrater-
restrial intelligence: Are there universal adaptations in search, aversion, and signaling?”
Biological Theory 13 (2): 131–141.
Wilman, Richard J., and Christopher J. Newman. 2018. Frontiers of Space Risk: Natural
Cosmic Hazards and Societal Challenges. CRC Press (Taylor & Francis Group).
Wilson, Elspeth M., and Carol E. Cleland. 2019. “Should we call E.T.? An ethical-political
analysis of METI.” Theology & Science 17 (3): 382–394.
Wohlleben, Peter. 2016. The Hidden Life of Trees: What They Feel, How They Communi-
cate. Vancouver: Greystone Books.
9
WHY DO WE ASSUME THAT WE CAN
DECODE ALIEN LANGUAGES?
Con Slobodchikoff

Discussion
In our human optimism, we assume that if we received a communication from an
alien species, we would be able to eventually decode that communication, and we
would over time be able to learn the language of the aliens. My contention is that
this is unlikely. We have a number of alien species sharing this planet with us: ani-
mals that communicate with each other, some using sophisticated communication
that is either language or approaches language (Slobodchikoff 2012). So far, we
have not been able to decode most of those languages. Even among our own human
languages, we have not been able to decode and learn a number of languages from
historical times. And yet, we make the assumption that if and when alien contact
occurs, we will be able to understand what the aliens have to say.
First, let me address something about the methodology that we are using to
search for alien signals and, we hope, alien languages. We are currently using mul-
tiple frequencies of radio signals and scans for optical laser pulses, and have come
up empty. So, let’s do a thought exercise. Let’s suppose that we are a culture that
lives in a valley that is surrounded by mountains, and we use smoke signals to
communicate with each other. We are convinced that there are other cultures liv-
ing in other valleys beyond our mountains, and we are looking for signs of their
communication. We know that smoke rises, so the logical assumption is that if we
communicate by smoke signals with each other, and our smoke rises, that other
cultures must be communicating with smoke signals as well, and if we just look for
their smoke we should be able to identify their signals. We search and search, but
it is in vain. We see no smoke signals. Occasionally we see smoke rising above the
mountains that encircle us, but it does not seem to have an organized pattern and
might just come from brush or forest fires in other distant valleys. Eventually, we

DOI: 10.4324/9781003352174-9
88 Con Slobodchikoff

conclude that there must not be other cultures out there, because if there were, we
would surely see their smoke signals. A point that we fail to grasp however, is that
there are lots of cultures in other valleys, but they communicate using cell phones,
a technology that is completely unknown to us. In my opinion, this is a scenario
that we fail to consider in our search for extraterrestrial communications.
But let us say that we do make contact. How are we going to decipher the signals
that we get? What are we going to look for that will tell us that this is a language
that we are encountering and not just some random fluctuations of energy produced
by physical phenomena among the stars?
Here we can turn to how linguists have approached a view of animal language.
In 1960, Charles Hockett, a linguist, proposed a number of design features which
need to be found if we were to recognize that an animal species had language
(Hockett 1960).
Hockett published a list of 13 design features of human language that would be
important to find in animals so that we could say that animals have language. Some
of these design features are found in any system that produces signals. However,
seven of the design features are really key elements that distinguish language from
mere communication. These seven are as follow.

• Semantics. Just as each word in a human language has a distinct meaning, the
signals that animals produce also have to have distinct meanings.
• Arbitrary. An arbitrary symbol has no direct connection to what it represents,
like the word “green” does not tell you anything about what the color green
actually looks like. This is in contrast to an iconic symbol, which represents
some attribute of the thing that it is describing. When you say “bow-wow” to
describe a dog, the “bow-wow” is an iconic symbol for dog, because it repre-
sents an attribute of dogs, namely barking.
• Discrete. Each symbol has to be a discrete unit, just like the words in this sen-
tence are all discrete units.
• Displacement. A language has to provide information about events that occur in
different locations from the speaker or in different time periods, i.e., displace-
ment in either space or time.
• Productivity. A language has to be able to make up new words. For example, the
word(s) “cell phone” did not exist in the English language until recently.
• Duality. Language has to have smaller units that can be combined into bigger
units. Think of how phonemes can be combined into morphemes, or words into
sentences.
• Cultural transmission. There must be a strong component of learning in a lan-
guage. We aren’t born knowing the language that we speak.

These design features are found in human languages. But when we try to apply
them to animals, we are faced with a conundrum. With humans, we can ask people
speaking a language whether something makes sense or not. We can rearrange
 Why Do We Assume That We Can Decode Alien Languages? 89

phonemes or morphemes and ask people whether our rearrangements are meaning-
ful to them. We cannot do that with animals.
With animals, we need a Rosetta Stone to give us the key to unlock their system
of language or communication. The Rosetta Stone is a stone that was found by
the French in 1799 in a village in Egypt that was then called Rosetta (now called
Rashid). It contains three sets of inscriptions of a decree issued in 196 BCE on
behalf of Ptolemy V. The bottom set of inscriptions is in ancient Greek. The mid-
dle set of inscriptions is in Demotic Egyptian, and the top set of inscriptions is in
Egyptian hieroglyphs. Because ancient Greek was a known language, and scholars
were able to recognize that the three sets of inscriptions represented the same text,
Jean-Francois Champollion announced in 1822 that he was able to transliterate
hieroglyphic Egyptian (Ray 2007).
While hieroglyphic Egyptian was transliterated with the Rosetta Stone, other
ancient human languages have not fared as well. Currently, Linear A from the
Minoan culture of around 1500 BCE, Vinca symbols from around 5000 BCE,
Rongorongo from Easter Island, and Olmec Script from around 1200 BCE remain
untranslated (Packard 1974; Fischer 1997; Mora-Marin 2009; Sproat 2010). In
all of these languages, there is no key, like the Rosetta Stone for decoding hiero-
glyphs, that gives linguists and anthropologists any clue about how to approach a
translation.
A Rosetta Stone for decoding animal languages is the context in which different
signals are used. Let me illustrate this with my work in deciphering the language
of prairie dogs (Slobodchikoff et al. 2009b). Prairie dogs are ground squirrels that
live in extensive systems of burrows that are called towns or colonies. Within a
prairie dog town, there might be hundreds of animals, and the ground is partitioned
into territories that are occupied by discrete social groups of prairie dogs. When a
predator such as a coyote or a hawk appears, one or more prairie dogs give an alarm
call which to our ears sound something like a bird chirping. Other prairie dogs
throughout the town run to their burrows to escape the predator.
The entire sequence of predator appearing, prairie dogs alarm calling, and run-
ning to their burrows offered my students and me a Rosetta Stone for decoding
Gunnison’s prairie dog language (Slobodchikoff et al. 2009b). We could videotape
the appearance and behavior of the predator as it approached the colony, we could
record the alarm calls that the prairie dogs produced in response to the predator,
and we could videotape the escape responses of the prairie dogs. Subsequently, on
another day when no predator was present, we could play back the recordings to the
predator that had previously appeared, videotape the escape responses of the prai-
rie dogs, and see if the escape responses of the prairie dogs when no predator was
present were the same as the escape responses when the predator actually appeared.
In a series of experiments, we found that prairie dogs have both different alarm
calls and different escape responses for different predator species. Predators of
prairie dogs include coyotes and domestic dogs, who try to catch the prairie dogs
on the ground, humans who shoot prairie dogs, and red-tailed hawks who swoop
90 Con Slobodchikoff

down on prairie dogs from the air. The escape responses of the prairie dogs differ,
depending on the species of the predator (Kiriazis and Slobodchikoff 2006).
For coyotes, prairie dogs immediately run to their burrows, and stand on their
hind legs at the burrow opening, watching the progress of the coyote through the
colony. Other prairie dogs that were below ground before the coyote appeared
emerge and also join those standing at the lip of the burrow entrances, watching
the coyote. This is related to the hunting strategy of coyotes, who will often find a
concentration of prairie dogs and lie down on the ground near those burrow open-
ings, waiting for unwary prairie dogs to emerge. By watching the progress of the
coyote, the prairie dogs have an excellent idea of where a coyote is likely to lie
down and wait for them.
For domestic dogs, prairie dogs will stand up on their hind legs wherever they
happen to be foraging, and watch the progress of the dog through the colony. Only
if the dog gets close to them do they run for their burrows. However, like with
coyotes, prairie dogs that were below ground emerge from their burrow openings
to stand and watch the progress of the dog through the colony. Dogs are not very
efficient predators of prairie dogs, and unlike coyotes, they do not lie down next to
burrows, so the prairie dogs do not need to find immediate safety in their burrows,
as they do with coyotes, but they do need to see where the dog is going.
For red-tailed hawks, prairie dogs that are in the immediate flightpath of the
stooping hawk will run into their burrows and dive inside, without standing at the
lip of the burrow, as they do with coyotes and with dogs. Other prairie dogs who
are not in the immediate flightpath of the stooping hawk stand on their hind legs
wherever they happen to be foraging and watch the progress of the hawk.
For humans, the entire colony of prairie dogs immediately runs to their burrows
and dives in, without emerging or standing at the lip of their burrows. Prairie dogs
have been hunted since ancestral times by Native Americans, who used bows and
arrows to shoot at them. At the present time, prairie dogs are shot by hunters who
view such killing as sport. Whether they are shot at with bows and arrows or with
rifle bullets, prairie dogs do not have very much time to escape, and become easy
targets if they stand at their burrow lips (Slobodchikoff et al. 2009b).
Just as the escape responses differ, we found that the alarm calls differ in their
acoustic structure to different species of predator. We found that prairie dogs
have acoustically different alarm calls for coyote, domestic dog, human, and red-
tailed hawk.
This gave us the Rosetta Stone to begin to unravel the language system of prairie
dogs. In a series of experiments, we were able to show the following design fea-
tures in the alarm calls of prairie dogs:

• Semantics. Prairie dogs have distinctly different alarm calls for different species
of predators. Furthermore, they can vary the acoustic elements within a call for
a predator species to describe the size, color, and shape of the predator (Slobod-
chikoff et al. 1991; Slobodchikoff et al. 2009b).
 Why Do We Assume That We Can Decode Alien Languages? 91

• Arbitrary. The calls are completely arbitrary, just like human words, with a
series of acoustic frequencies changing as a function of time within an alarm
call (Placer and Slobodchikoff 2000).
• Discrete. Each alarm call is a discrete unit (Placer and Slobodchikoff 2001).
• Displacement. Prairie dogs are able to indicate the presence of a predator that
is far away from them. One experiment showed that prairie dogs were able to
consistently affix an apparent label of a gun to a person who once fired a shotgun
and then subsequently appeared without a gun for the month of the experiment
(Frederiksen and Slobodchikoff 2007).
• Productivity. Prairie dogs are able to coin new “words” for objects that they
have never seen before, such as a black oval, a triangle, and a circle (Ackers and
Slobodchikoff 1999; Slobodchikoff et al. 2012).
• Duality. An analysis of the structure of the alarm calls showed that they were
composed of phonemes, just like human words are composed of phonemes. For
the alarm calls for the different species of predators, most of the same phonemes
were used, but the proportion of phonemes that was used in alarm calls for dif-
ferent species of predators was different (Slobodchikoff and Placer 2006).
• Cultural transmission. An experiment showed that newly-born prairie dogs
emerging from their burrows have a non-specific alarm call to all predators, but
over time, the specificity increases, suggesting that some aspects of the call are
under genetic control and other aspects of the call are determined by cultural
transmission (Slobodchikoff 2010).

Prairie dogs represent one of the best animal language-like systems that has been
decoded, but there are a number of other examples of animals having language-
like properties (Slobodchikoff 2012): sagebrush lizards have been shown to have
a grammatical system in their head bobbing, tail lifting, and arm lifting (Martins
1993); both Japanese tits and American chickadees have been shown to have syn-
tax in their vocalizations (Ficken et al. 1987; Suzuki et al. 2016); and blackbirds
have been shown to have recursion in their calls (Gentner et al. 2006). In all of
these cases, the behavioral context represents a starting point for decoding the lan-
guage of communication system. As with prairie dogs, the behavioral context is
the Rosetta Stone.
However, there are many situations in which a behavioral context is absent.
Imagine that someone who does not know that humans have a language – perhaps
an extraterrestrial alien – is trying to figure out whether humans are able to mean-
ingfully communicate to one another. This alien is able to detect that I am talking
to you on a cell phone. I am holding the phone in my hand and nothing about my
behavior gives a clue as to whether I am speaking a language or I am merely vocal-
izing my emotional state and excitement into a piece of plastic. Similarly, an analy-
sis of your behavior at your end shows that there is no behavioral context other
than the production of noises. This alien could very well conclude that humans do
not have a language but just make a series of vocalizations, perhaps as aggressive
92 Con Slobodchikoff

or mating signals. Unfortunately, this is a position that some biologists and some
linguists take with respect to animal language.
To be fair, it would be possible to analyze these conversations by looking at
the structure of the sounds. Perhaps some sounds occur more frequently than oth-
ers, such as described by Zipf’s Law (Newman 2005). Perhaps some sounds tend
to occur more frequently after a pause, or perhaps some sounds precede a pause,
but nothing about these structural features gives a clue about the meaning of the
conversation. The meaning can be either derived from the behavioral context or
from asking the participants of the conversation, but to ask the participants of the
conversation what they meant, you have to have at least the semblance of either
their language or some other language in common. Applying this logic to decoding
alien languages produces results that are not very hopeful.
But what about communicating with aliens through mathematics? Surely
mathematics are universal, right? Unfortunately, not necessarily. First, there
are the limitations on numbers imposed by different languages. Some languages
count objects as one, two, then many. Some languages go simply from one to
many. The evidence is still ambiguous as to whether languages constrain their
speakers’ perception of external objects or events (Boroditsky 2001). Second,
we tend to think of numbers as discrete integers, such as one, two, three, and
so forth. Our computers treat numbers as zero and one, or “on” or “off.” So
perhaps we can deal with aliens through numbers, or at least deal with them in
sequences of zeros and ones. But what if they did not view numbers as discrete
entities? What if instead of a discrete zero and a discrete one, they viewed
numbers as continuous variables, so that there would be an infinite number of
states between zero and one? We could send out prime numbers, but if aliens
perceived numbers differently from us, our messages of prime numbers would
be meaningless.
Perhaps geometry can help us. After all, a circle is always a circle, a triangle is
always a triangle, and pi is always pi. But we know that our brains are hardwired to
detect certain patterns, which is why our brains can be tricked by illusions (Yarritu
et al. 2015). Our understanding of the abilities of animals has expanded so much
that we now know that different species of animals can perceive the world differ-
ently from us. We cannot see into the ultraviolet range of the spectrum, whereas
bees and some birds can (Lind et al. 2013; Cronin and Bok 2016). We cannot detect
magnetic fields, whereas a number of animals, particularly the those that migrate,
can perceive changes in the magnetic field (Taylor et al. 2017). We have no guaran-
tee that aliens would see a circle as a circle and a triangle as a triangle. Once again,
we might be trapped by the smoke signal analogy.
We might argue that we have things in common with aliens that can be expressed
through mathematics. We know that the world is made up of atoms which can be
counted. We know the time when the Big Bang created our universe. So we have
atoms and time in common, right? However, ever since Einstein formulated his
 Why Do We Assume That We Can Decode Alien Languages? 93

special theory of relativity, we know the time is a somewhat slippery concept and
is not necessarily perceived by everyone in exactly the same way (Austin 2017).
Similarly, the development of quantum theory – with its concepts of superposition,
coherence, and energy states – calls into question whether atoms are discrete objects
or are waves or energy fields (Rosenblum and Kuttner 2011). What we see as dis-
crete objects, aliens might see as waves or continuously fluctuating energy fields.
All of these factors represent possible impediments to us understanding the lan-
guages of aliens, whether we encounter these languages through the detection of
signals or through face-to-face encounters.
However, there is a possible work-around. If aliens are smarter than us, or if
they are more experienced in contact with a variety of extraterrestrial cultures, they
could provide us with a Rosetta Stone that would allow us to communicate with
them at some level. They could teach us the rudiments of their language, or even a
simplified language that would make communication possible. This is something
that we have done with a few chimpanzees, such as Washoe and Kanzi, and Koko
the gorilla, by teaching them American Sign Language or teaching them to respond
to symbols on a keyboard (Gardner et al. 1989; Savage-Rumbaugh et al. 1998;
Hare and Yamamoto 2017). If aliens can teach us something that represents com-
mon ground between them and us, that might provide enough of a Rosetta Stone
for us to begin to communicate with them. Whether we will ever, under those cir-
cumstances, be able to discuss philosophy or higher concepts with them remains an
open question. We have not yet succeeded with our resident aliens, all of the animal
species around us, but we can sweep that lack of success under the rug by assuming
that these animals are too stupid to have language. We can only hope that visiting
aliens will not think the same of us.

References
Ackers, S.H. and C.N. Slobodchikoff. 1999. “Communication of Stimulus Size and Shape in
Alarm Calls of Gunnison’s Prairie Dogs.” Ethology 105: 149–162.
Austin, R.W. 2017. “Gravitational time dilation derived from special relativity and Newto-
nian gravitational potential.” European Scientific Journal 13. http://dx.doi.org/10.19044/
esj.2017.v13.n3p%25p.
Boroditsky, L. 2001. “Does language shape thought?: Mandarin and English speakers’ con-
ception of time.” Cognitive Psychology 43: 1–22.
Cronin, T.W., and M.J. Bok. 2016. “Photoreception and vision in the ultraviolet.” Journal of
Experimental Biology 219: 2790–2801.
Ficken, M.S., C.M. Weise, and J.A. Reinartz. 1987. “A complex vocalization of the black-
capped chickadee. II. Repertoires, dominance, and dialects.” Condor 89: 500–509.
Fischer, S.R. 1997. Rongorongo: The Easter Island Script. Oxford: Clarendon Press.
Frederiksen, J.K., and C.N. Slobodchikoff. 2007. Referential specificity in the alarm calls of
the black-tailed prairie dog.” Ethology, Ecology & Evolution 19: 87–99.
Gardner, R.A., B.T. Gardner, and T.E. Van Cantford. 1989. Teaching Sign Language to
Chimpanzees. Albany: State University of New York Press.
94 Con Slobodchikoff

Gentner, T.Q., K.M. Fenn, D. Margoliash, and H.C. Nusbaum. 2006. “Recursive syntactic
pattern learning by songbirds.” Nature 440 (7088): 1204–1207.
Hare, B., and S. Yamamoto, eds. 2017. Bonobos: Unique in Mind, Brain, and Behavior.
Oxford: Oxford University Press.
Hockett, C.F. 1960. “Logical considerations in the study of animal languages.” In Animal
Sounds and Communication, edited by W.E. Lanyon and W.N. Tavolga, 392–430. Wash-
ington, DC: American Institute of Biological Sciences.
Kiriazis, J., and C.N. Slobodchikoff. 2006. “Perceptual specificity in the alarm calls of Gun-
nison’s prairie dogs.” Behavioural Processes 73: 29–35.
Lind, O., M. Mitkus, P. Olsson, and A. Kelber. 2013. “Ultraviolet vision in birds: The im-
portance of transparent eye media.” Proceedings of the Royal Society B. doi:10.1098/
rspb.2013.2209.
Martins, E.P. 1993. “Contextual use of the push-up display by the sagebrush lizard, Scelopo-
rus graciosus.” Animal Behaviour 45: 25–36.
Mora-Marin, D.F. 2009. “Early olmec writing: Reading format and reading order.” Latin
American Antiquity 20: 395–412.
Newman, M.E.J. 2005. “Power laws, Pareto distributions and Zipf’s law.” Contemporary
Physics 46: 323–451.
Packard, D.W. 1974. Minoan Linear A. Berkeley: University of California Press.
Placer, J., and C.N. Slobodchikoff. 2000. “A fuzzy-neural system for identification of spe-
cies-specific alarm calls of Gunnison’s prairie dogs.” Behavioural Processes 52: 1–9.
Placer, J., and C.N. Slobodchikoff. 2001. “Developing new metrics for the investigation of
animal vocalizations.” Intelligent Automation and Soft Computing 8: 1–11.
Ray, J. 2007. The Rosetta Stone and the Rebirth of Ancient Egypt. Cambridge, MA: Harvard
University Press.
Rosenblum, B., and F. Kuttner. 2011. Quantum Enigma. Second Edition. Oxford: Oxford
University Press.
Savage-Rumbaugh, S., S.G. Shanker, and T.J. Taylor. 1998. Apes, Language, and the Hu-
man Mind. Oxford: Oxford University Press.
Slobodchikoff, C.N. 2010. Talk of the town. BBC Video.
Slobodchikoff, C.N. 2012. Chasing Doctor Dolittle: Learning the Language of Animals.
New York: St. Martin’s Press.
Slobodchikoff, C.N., W.R. Briggs, P.A. Dennis, and A.-M. Hodge. 2012. “Size and shape
information serve as labels in the alarm calls of Gunnison’s prairie dogs Cynomys gun-
nisoni.” Current Zoology 58: 741–748.
Slobodchikoff, C.N., J. Kiriazis, C. Fischer, and E. Creef. 1991. “Semantic information dis-
tinguishing individual predators in the alarm calls of Gunnison’s prairie dogs.” Animal
Behaviour 42: 713–719.
Slobodchikoff, C.N., and J. Placer. 2006. “Acoustic structures in the alarm calls of Gunni-
son’s prairie dogs.” Journal of the Acoustical Society of America 119: 3153–3160.
Slobodchikoff, C.N., A. Paseka, and J. Verdolin. 2009a. “Prairie dog alarm calls encode
labels about predator colors.” Animal Cognition 12: 435–439.
Slobodchikoff, C.N., B. Perla, and J. Verdolin. 2009b. Prairie Dogs: Communication and
Community in an Animal Society. Cambridge, MA: Harvard University Press.
Sproat, R. 2010. “Ancient symbols, computational linguistics, and the reviewing practices
of the general science journals.” Computational Linguistics 36: 585–594.
Suzuki, T.N., D. Wheatcroft, and M. Griesser. 2016. “Experimental evidence for composi-
tional syntax in bird calls.” Nature Communications 7: 10986.
 Why Do We Assume That We Can Decode Alien Languages? 95

Taylor, B.K., S. Johnsen, and K.J. Lohman. 2017. “Detection of magnetic field properties
using distributed sensing: A computational neuroscience approach.” Bioinspiration &
Biomimetics 12. doi.org/10.1088/1748-3190/aa6ccd.
Yarritu, I., H. Matute, and D. Luque. 2015. “The dark side of cognitive illusions: When
an illusory belief interferes with the acquisition of evidence-based knowledge.” British
Journal of Psychology 29 January. doi.org/10.1111/bjop.12119.
10
XENOLINGUISTIC FIELDWORK
Claire Bowern

Preliminaries1
How would we get to work deciphering an alien language when the time comes?
How hard would this be? While our knowledge of human language will only get
us so far, the tools we have developed for linguistic fieldwork and analysis will
be critical.
(Coon 2020: 44)

This chapter is about the tools developed for linguistic fieldwork and analysis and
what is needed to understand extraterrestrial linguistic systems. I discuss what
aspects of human linguistic fieldwork probably transfer to xenolinguistic fieldwork,
what does not, how well prepared we are, and challenges that may arise. I argue
that while fieldwork approaches are, broadly, capable of helping us understand ET
languages, our data storage and archiving systems are woefully inadequate. Such
an upgrade would be beneficial to human linguistic fieldwork, too.
Broadly speaking, “fieldwork” is collecting data for analysis from a natural
context. That is, fieldwork is placed in contrast to laboratory or naturalistic experi-
ments or other types of “intervention,” theoretical models, corpus building from
secondary sources, or meta-studies. It involves the collection of primary data by
researchers directly from the languages users (cf. Thomas 2020 and references
quoted there).
Linguistic fieldwork grows out of an anthropological tradition of ethnography
(Brewer 2000), which combines participant observation (looking at what people
do), interviews (asking people what they do), and immersion and participation
(learning what people do by participating oneself).2 Linguistic fieldwork typically
also includes structured tasks: a range of activities ranging from general questions

DOI: 10.4324/9781003352174-10
 Xenolinguistic Fieldwork 97

to precise translations to open-ended recordings and conversation. This “typical”

fieldwork involves living in the community, though “remote fieldwork” is also
increasingly common (Williams et al. 2021).
Linguistic fieldwork is, these days, usually a collaboration between language
participants and the researcher, with shared negotiated goals and procedures. On
Earth, fieldworkers may travel to the other side of the world or step out their front
door, working on their own language in their own community, or not. Linguists and
anthropologists sometimes use the notion of “alien” in their description of field-
workers (cf. Ly and Spjeldnæs 2021), concentrating on the differences between
fieldworkers who are community outsiders and the language and culture they are
researching. Others, such as Rice (2009), concentrate on the collaborative nature
of fieldwork where linguists and communities come together, with shared goals.
I am assuming that the ET system is a linguistic one. As other chapters in this
book have discussed (e.g., Slobodchikoff, this volume [Chapter 9]; Pepperberg,
this volume [Chapter 6]), it can be no easy task to determine whether a communi-
cation system is “language.” I am also making some assumptions about ET social
structures. For example, I am assuming that linguistic systems arise and are con-
ventionalized through social learning, as well as having properties due to innate
psychology and physiology. While discussions of the language faculty and uni-
versal grammar typically strongly focus on the innate and the properties of the
individual, there are also social properties that shape language. They are not just
a “third factor” (relegated to relative unimportance) but crucially shape the field
experience. Social learning is an inherent part of language, as well as linguistic
innateness.
Some features of language come from human psychology and physiology. The
phonology of signed and spoken languages, for example, is embodied, in that if
our bodies were differently shaped or functioned differently, that aspect of lan-
guage would be different, as well. Other features of language come from human
culture. The lexicon of different languages divide up concepts in limitedly distinct
ways. Some of this comes from human psychology, other aspects from culture.
Languages vary in the way that words for colors are realized, for example (Kay and
Maffi 1999). Some languages make no use of hue for describing items at all. Others
use only a two-term system, whereby a single term covers dark colors and another
covers light. Others have more terms; but while there are thousands of different
languages today, there are not thousands of color term systems or hundreds, or
even tens. The vast majority of the world’s languages can be described with seven
basic categories. A different example comes from words for boats, which reflect not
human psychology but the cultural conventions of craft building.
Other properties of human language serve to distinguish it from non-linguistic
systems (Hockett 1959). It can be difficult to determine whether an example of
“human language” is a linguistic system or not, however. To take one example,
the Voynich Manuscript (cf. Bowern and Lindemann 2021) has been argued about
for more than fifty years: is it an enciphered (already known) human language, a
98 Claire Bowern

constructed language, or a hoax – a non-language? As Slobodchikoff (this volume

[Chapter 9]) and Pepperberg (this volume [Chapter 6]) have argued, it may be quite
difficult to identify ET communication systems as language.
Whether or not the system is “linguistic” – in the sense that it displays prop-
erties of human languages such as recursion, displacement, and duality of pat-
terning, among other aspects – does not actually affect methods for fieldwork and
description. That is, we expect such features to be present in human languages. For
example, when I work on language documentation, I expect there to be ways to
talk about things that are not present in the here and now. I expect that a sentence
will mean the same thing (broadly) from person to person and session to session, if
a word minyaw means “cat” in one session it will not suddenly mean “dog” in the
next one. But if those assumptions do not hold, I will realize very quickly. The lack
of displacement, for example, will emerge from the corpus and from the answers
to questions about things that are not visible. It is usually fairly obvious if there is
no way to directly translate a feature of the linguists’ metalanguage, but it can be
less straightforward to distinguish features that appear in the field language but not
the metalanguage. Just as no human language description is “comprehensive,” ET
linguistic descriptions are likely to miss a lot.

Fieldwork Logistics
As Coon (2020), Macaulay (2004), Bowern (2015), and others have discussed,
language is only one part of a much larger set of fieldwork logistics. There’s travel
to the field site (including any relevant visas and permits), accommodation, com-
pensating consultants for their time, and finding appropriate places to work, among
other things. Fieldwork costs money and there are few funding sources. And then
there is the matter of how to get home again. This section describes some of the
issues around logistics in conducting fieldwork.

Aims and Goals of Fieldwork

The aims of the linguistic fieldwork need to be clearly articulated. What is the
purpose of studying the language? To establish shared communication? For
humans’ cultural and scientific knowledge? As a way of establishing scientific
and cultural reciprocity? To assist ETI communities with language preservation
in the face of human appropriation and assimilation? For easier trade paths? For
the production of language learning or other educational materials? As part of a
program of military interrogation or intergalactic espionage? Evangelism of ET
communities and individuals? As Rice (2022) discusses in detail, “linguistics
has been utilized as a tool for the harmful practice of attempting religious con-
version of Native peoples.” Fieldwork has been conducted for all these reasons
(sometimes simultaneously), and we cannot think about fieldwork in the abstract
 Xenolinguistic Fieldwork 99

without thinking about the goals and consequences of such work (cf. also Charity
Hudley et al. 2018).
Movies like Arrival focus on fieldwork and initial contact. That is, the linguist is
one of a team of scientists and military personnel and the linguist works in part to
further the goals of the military and political operation. They work under immense
time pressure as the fate of the world hangs on their ability to “decode” the lan-
guage. This role combines the military notion of linguist as translator with the
academic idea of linguist who works on language structure.
A more realistic view of linguistic fieldwork is to imagine a larger, long-term
field project that is not tied up with the initial encounter with ETI. Fieldwork takes
time. It involves not just figuring out puzzles, but coming to a shared understanding
with communities. Fieldwork requires a lot of goodwill on both sides to be effec-
tive, particularly on the part of the host community. Such goodwill is unlikely to be
easy to achieve under initial military supervision. If I had to come up with a first
contact “fieldwork” plan, it would not involve the detailed analysis of linguistic
structure. It would be aimed solely at establishing some shared communication and
non-threat.
Bronislaw Malinowski, who is usually credited with developing ethnographic
(particularly participant observation) methods in anthropology, notably begins a
description of how to do fieldwork with what amounts to “find the local white man”
and stay with him.3 While there is often relief in the familiar, a big thing about field-
work is being able to step outside one’s own assumptions about language, culture,
and community (both one’s own and that of others). Linguistic fieldwork is about
language, but it is also more than about language. Language is a means to a shared
understanding that goes well beyond the specifics of particular linguistic systems.
As McKenzie (this volume [Chapter 11]) writes, language is inherently coopera-
tive; so is documenting language. Cooperation is two-way, so linguists should also
be prepared to be representatives for human culture to ET communities.

Shared Language for Fieldwork

Much fieldwork on Earth (though by no means all) is conducted through a lin-
gua franca – either the fieldworkers’ own first language, or a language shared
within at least some of the community. That shared language makes some field-
work tasks easier. Direct translation, for example, is an effective way to increase
vocabulary quickly – the “what is this called?” or “How do you say this?” type
of questioning.
Monolingual and bilingual fieldwork have some of the same tools (see Everett
2001). As Thomas (2020) points out, both “monolingual” and “bilingual” field-
work are multilingual; the main difference is in whose primary language is the
medium for linguistic discussion. Learning the language is a mark of respect and
produces better insights into the language structure (because the fieldworker knows
it better). It allows for better community-oriented work and a deeper knowledge
100 Claire Bowern

of the subject on the part of the linguist. Of course, depending on the nature of the
ETI, such a language may or may not be possible for a human to articulate.
Some fieldwork uses interpreters. That is, the linguist and the research partici-
pants do not share a common language, and an interpreter translates the linguist’s
questions for the language worker (for example, a linguist in Australia might ask a
question in English, which is translated into Kriol, and then answered in an Indig-
enous language). Depending on the circumstances of ETI contact and who learns
each other’s languages, human language interpreters may be involved.

Teams and Individuals

Linguistic fieldwork is often conducted individually. Bowern (2015) discusses the
trope of the lone fieldworker dropped in an “alien” community. The team approach,
assuming there are multiple community members to work with, will likely provide
more progress more quickly, as well as allowing more specialization. However, as
discussed later in this chapter, most methods of field data storage and analysis are
not easily used by teams; they are designed for individuals working alone. This
is a problem for human fieldwork, as well. If the project is conducted as part of a
group, one needs a project manager. This type of fieldwork is particularly common
in English-speaking countries, but other traditions exist. As Thomas (2020) notes,
the canonical Russian field tradition is group-oriented (cf. work in particular led by
Kibrik [cf. Kibrik 2006]).

Remote or In Person?
Fieldwork is typically assumed to be the linguist going elsewhere. Before Malinow-
ski were the ethnographic expeditions (compare Ray and Haddon 1891; Haddon
1935) that involved travel to remote parts of the Empire (British or Russian, for
example) to collect cultural, linguistic, and biological samples from communities.
Of course, long before that were the linguistic and cultural souvenirs from travelers
and “explorers” (compare Busbecq’s reports on Crimean Gothic [cf. Stearns 1978],
for example). Fieldwork is often thought to be about traveling to the other place, to
see how language is used in its natural context. Elwin Ransom in C.S. Lewis’ novel
Out of the Silent Planet does do this. But Louise Banks in Arrival is on Earth, with
all the comforts and resources of Earth. This makes a difference, both in resources
available to the linguist and in the type of data that one is able to collect and the
observations that one is able to make. There is something comforting about being
on one’s home turf. But it really limits what one records (cf. Tsikewa 2021 and oth-
ers for discussions about the artificialness of field methods classes, for example).
“Fieldwork” has been about going to the community, observing and participat-
ing in the community first hand and using those insights and data to understand
language. The COVID-19 pandemic shut down university research and travel,
making in-person fieldwork impossible for months or years (cf. Sanker et al. 2021;
 Xenolinguistic Fieldwork 101

Williams et al. 2021; Sneller 2022). Some linguists used virtual, “at-a-distance”
field methods while unable to travel. These included video linkups with consult-
ants (that is, interaction in real time but through digital connections) and messaging
through communication and social media apps such as WhatsApp. Remote work of
this type presupposes that the means of communication is transferable over video
linkup. Sanker et al. (2021) found many ways in which acoustic data was distorted
through remote recordings, for example.

Other Logistics
Where to stay and nonlinguistic things to take care of while on fieldwork can take a
substantial amount of time, interstellar travel aside. There are the issues of getting
there and getting home again. Where to stay while in the community, how to deal
with other tasks that typically come about. After all, if the linguist is going to the
ET community, they are likely to be as much an object of study and curiosity as ET
visitors are to Earth.

Fieldwork Tools and Methods

As previously discussed, language is embodied: it is produced by particular bod-
ies and minds/brains, so there is no reason to think that a ET linguistic system
will have properties that humans can perceive, learn, or use to communicate with.
Developing solutions to that problem will depend on the specifics of the system
(for example, modulating audio frequencies to make them perceptible by human
ears).

Collecting Data
Fieldwork methods should be broadly applicable to a wide range of ET linguis-
tic systems. However, existing fieldwork tools are strongly constrained by human
physiology and psychology, along with being (unsurprisingly) designed for items
that occur on Earth. Some of these techniques and issues are outlined in what fol-
lows. As Coon (2020) and Wells-Jensen and Spallinger (2020) have previously
described, these methods all require a considerable degree of cooperative attention
and shared understanding in order to succeed, so there are certain cultural pre-
requisites for even doing linguistic fieldwork. However, since social and cultural
learning is part of the definition of language, it is not too farfetched to assume that
if there is enough common understanding for interaction of any type, there will be
enough for language learning and linguistic fieldwork.

Vocabulary collection: A key component of fieldwork is knowing what things

in the world are called in the target language. Such words are typically found
out by either pointing, asking directly, or encountering in the course of other
102 Claire Bowern

work. Such techniques all require that we are able to distinguish “words” in the
language of course. Pointing requires an understanding of what one is eliciting
when pointing to things (Wells-Jensen and Spallinger 2020). The literature is
full of initial errors from this technique (e.g., “nose” recorded as “I” or “finger”
recorded as “you,” for example) but they are usually straightened out fairly
quickly with subsequent work. Monolingual elicitation strategies begin with
vocabulary collection by pointing, and translation-based approaches usually
begin with a wordlist of basic vocabulary (see further Bowern 2015).
Interactions: Through observing interactions it is possible to find out what phrases
are used for greetings and other (culturally dependent) formulaic situations
and interactions, such as describing the weather or wishing someone a happy
birthday.
Translated sentences: If there is a language in common, the linguist can provide
sentences and ask for translations in the target language, or propose sentences
in the target language and ask for corrections and back-translation to a common
language.
Collection of language samples: e.g., recording narrations and conversations and
working through them to identify structures and explore meaning once there is
something of a language in common. There is a large array of semi-structured
and unstructured techniques for obtaining linguistic data that are likely to reveal
variety of morphosyntactic and semantic constructions: causation, location, or
possession, to name a few. One can also use activities to record vocabulary
and grammar, such as narrating what someone is doing. A better technique may
be to be led by the ETI research partners and participant observation. That is,
rather than requiring ETIs to come up with novel descriptions for items that
they may be encountering for the first time on Earth, do field linguistics on the
ETI’s home planet. While this will produce better linguistic documentation, the
logistics may be insurmountable.

Thus contra Slobodchikoff (this volume [Chapter 9]), we have many tools at our
disposal for figuring out the features of human languages; mostly this is because
we are able to communicate reciprocally (even if nonlinguistically) about various
topics, even if we do not know the language. Linguists do not just have to experi-
ment on the research participants, and indeed, the research participant is typically a
partner in the research, not experimented on. They should not be experimented on
without their consent, and discursive linguistic fieldwork works better with shared
common goals and full participation of all. This means that a big part of designing
fieldwork tasks should be consideration for the research participants.

Typological Surveys
Note that many of the lists are shaped by assumptions about the properties that
we expect languages to have. For example, all human languages have ways of
 Xenolinguistic Fieldwork 103

describing events, participants in those events, relationships between those partici-

pants and those events, and attendant circumstances for those events. All languages
have some way of describing events in time and space, though the details vary.
Several chapters in this volume discuss these points as key features of human lan-
guage and as likely to be part of ETI languages, as well.

Fieldwork Equipment
The typical fieldwork setup for recording human language work includes audio
and video recording devices and accompanying equipment (such as microphones),
computer(s), and notebooks. Fieldwork requires some way of storing and retriev-
ing observations. One should not rely on memory, and documentation collections
are made for multiple uses. Language samples are typically recorded in some for-
mat (audio and/or video), transcribed to text, and the text is analyzed. The tran-
scription is accompanied by annotations for features of interest; annotations and
transcriptions are typically time-aligned using software such as ELAN (Wittenburg
et al. 2006).
Recording devices may not work on other worlds, since they are designed for
Earth’s atmosphere and require electricity and batteries that are only obtainable on
parts of Earth. Pencil and paper works in space as long as the human linguists do
not have to wear protective equipment that will impede writing. It is fairly durable
but heavy to transport. Audio devices are also, of course, optimized for recording
particular acoustic frequencies, video for particular light wavelengths. Linguistic
fieldwork is severely underfunded and reliant on the market of such devices for
other purposes (typically amateur music recording and podcasting). ET fieldwork-
ers should be prepared (as fieldworkers usually are) for their recording devices to
fail, and any interstellar fieldwork project will need to be well funded for possibly
custom recording equipment.

Keeping Track of Data

Linguistic field data is stored in a database, so that language samples can be
analyzed, compared, and retrieved. Here we are woefully underprepared for
ETI languages, since current tools, to be blunt, do a poor job of covering even
the diversity shown by human languages. Because linguistic fieldwork data
comes in multiple formats and there are particular requirements for handling
multilingual records, the type of qualitative research organization software is
typically not appropriate for managing linguistic projects,4 but linguistic soft-
ware itself is designed for a narrow set of fieldwork data types and language
structures.
The biggest bottleneck for linguistic work is transcription and parsing. For every
minute of audio or video data recorded, for example, transcription can take 5 times
to more than 100 times as long, depending on the level of detail of the annotation,
104 Claire Bowern

the complexity of the transcription, the quality of the recording, and the experi-
ence of the linguist. Speech-to-text (STT) transcription programs are not usually
usable, because they are custom-designed for languages with more data.5 Thus,
transcription needs to be done manually, by someone familiar with the language. In
team-based ETI fieldwork, expect to need five to ten transcribers for every linguist
directly asking questions.
Linguistic data for language documentation is most useful when it is parsed –
that is, when there is an association between elements in a phrase in the field lan-
guage and a more familiar metalanguage, as in the following example from Bardi
(Northwest Australia, Nyulnyulan family).

Nyirroogoordoo minkal?
ɲɪrukuɖu mi-n-kal?
how 2sg-tr-travel?
“How are you?”
A parsed corpus is essential if anyone other than the linguists who collected the
data in the first place are going to use the materials. Creating these parses can
be done manually or by using a parser that looks up information from a list of
words and morphemes and returns a translation. For some languages, existing soft-
ware works well. But there are plenty of (common!) features of human languages
that are poorly handled by existing parsers. In the preceding example from Bardi,
minkal is not overtly marked for tense. Tense and aspect in Bardi are given by a
combination of tense prefixes and tense/aspect suffixes (Bowern 2012), but in this
verb, neither are overtly marked. Null morphemes can create issues for parsers.
Parsers are also typically unable to elegantly process extensive allomorphy or sup-
pletion, nonconcatenative phenomena, suprasegmental morphology (for example,
morphology marked by tone), or simultaneous exponence as is very frequent in
signed languages. They work on transcribe textual data. Any ETI linguistic analy-
sis is going to have to investigate such phenomena and we have no easy way of
tracking it when we find it.

Archiving and Long-Term Storage

Fieldwork data needs to be stored long-term in formats that will allow retrieval
(Bird and Simons 2021; Henke and Berez-Kroeker 2016). Here, too, we are woe-
fully unprepared. Fieldwork generates a lot of data, so any project working off
Earth will need ways to transmit data back for safekeeping.6 It is also unclear what
archive would host such materials, since Earth-based archives are typically organ-
ized around regions or cultures of humans on Earth. On the principle of LOCKSS
(Lots Of Copies Keeps Stuff Safe), we would want ETI data collections to be housed
in multiple archives and be open access assuming that ETI cultural protocols find
that to be acceptable.
 Xenolinguistic Fieldwork 105

Descriptive Metalanguage
We will need a metalanguage for description. Linguistic fieldwork – talking about
language through language – has developed a complex vocabulary for describing
the languages of Earth. It is not at all clear how much of that metalanguage would
transfer to ETI linguistics.
Throughout the history of fieldwork and description of languages has been
describing phenomena in a model of the familiar and realizing that it does
not fit. For example, the 19th-century grammar writers of Indigenous Austral-
ian languages were unfamiliar with ergativity, and have to figure out ways of
describing the case marking patterns they encounter, either by co-opting famil-
iar terms from traditional grammar (that do not fit) or by inventing new ways
of describing the phenomenon. This is likely to be a problem for describing ET
languages.

Other Considerations

Linguistic Training
What training do fieldworkers currently get? Tsikewa (2021) is a recent review. She
finds, just as Macaulay (2004: 194) did fifteen years earlier:

While we generally do a very thorough job of teaching [our linguistics students]

how to elicit and analyze data, we often forget to tell them that there is a per-
sonal and practical side to fieldwork that can very well derail their research if
they are not prepared for it.

More recent books, including Bowern (2015) and Meakins et al. (2018), do devote
space on the parts of fieldwork that do not involve data elicitation directly. But
fieldworkers wear many hats and are not equipped for extraterrestrial fieldwork
(considering the amount of training that astronauts get). Perhaps fieldwork from
Earth (either remote or with ET community members here) is more practical and
realistic, but loses a lot of the nuance of in situ work. It might also lead to inaccura-
cies in the language documentation. It makes it hard to know what aspects of the
language are the properties of the individual and what are the properties of a larger
social group or ET linguistic community.
Most fieldwork these days is not ex nihilo – there is at least something for
most languages, or related languages. We know something about the families
that languages belong to.7 Fieldworkers very seldom start entirely from scratch
these days. There are probably not many people, even those with extensive field-
work training, who have started to work on a language where there is no infor-
mation whatsoever about the language, its close relatives, or its geographical
neighbors.
106 Claire Bowern

Ethics
Fieldwork research has numerous ethical considerations.8 We cannot assume that
it will just be fine to rock up and learn ET languages. Human cultures and indi-
vidual humans have a wide range of views when it comes to the appropriateness of
outsiders learning their languages and using them elsewhere. Fieldwork requires
consent: active, continually negotiated, and informed consent (Rice 2006; D’Arcy
and Bender 2023). Informed consent is difficult (if not impossible) to achieve with-
out some understanding of shared goals and a metalanguage for communication.
There are metalinguistic beliefs – folk linguistic theories about language – that
may assist with linguistic fieldwork, or may prevent it. The dominant Western view
of language is utilitarian. People do not own languages, for example. Languages
are equivalent, they are part of culture (and signal features of culture). But there are
other views of language, where it comes from, who can speak it, and so on, that are
often difficult to access directly but are important for acting ethically.
Fieldwork brings other aspects of ethics and interaction, too. These apply to any
human–non-human contact (that is, anyone involved in work with ETIs). Field-
workers could bring diseases, for example. They are likely to be just as much a
subject of curiosity and interest as ETIs on Earth.

Language Variation
Thus far, I have been mostly describing fieldwork as it is “traditionally” done, with
relatively more focus on features that apply across all (or most) users of a language,
rather than the parts of a language where individuals tend to differ. I have not dis-
cussed other aspects of language description, such as language surveys of contexts
of use, language censuses, or other types of community survey that investigate
what languages are used and where. I have not touched on the investigation of lin-
guistic variation. While there are many points in common between sociolinguistic/
variationist methods and “fieldwork” methods for language documentation, they
tend to target different aspects of language, produce different data, and – crucially –
involve work with quite different numbers of community members. All of this will
be needed, too.
Second, while I have here been talking about fieldwork with an ET language,
there is no reason to expect a single homogenous linguistic system. Fieldwork in
multilingual situations is the norm, and we have no reason to assume that any ET
encounter will involve a single system.

Fieldwork on METI (Messaging ETI)

So far, most of this chapter has been about fieldwork as interaction, but there are
techniques that fieldworkers can be helpful for when considering METI and decod-
ing extraterrestrial messages.
 Xenolinguistic Fieldwork 107

Such work is considerably harder than fieldwork. Human language records are
incredibly difficult to decipher when divorced from their context of use, or when
immediate and somewhat guided interaction is not possible (or where scaffolded
questioning is not possible). It is really only possible when we either know what
language is likely to underlie it (e.g., in the case of a cipher), where we have a
“Rosetta stone” (that is, enough parallel translation that we have anchoring points
and can extrapolate from what we already know to what we do not know), where
there is enough inferred context that we can make a guess as the types of structures
that are present, or where we have records from a closely related language. In addi-
tion, we also need enough data to be able to test theories.

Conclusions
We are poorly prepared for extraterrestrial fieldwork. Quite apart from the logis-
tics of a field trip beyond Earth, linguists lack both training and – especially – the
tools to do this well. This is a highly skilled part of linguistics, requiring extensive
domain knowledge, but linguistics training tends to compartmentalize early: to
particular subsystems of language (the sound system or the syntax, for example).
Crosslinguistic and cross-field expertise is not as common as it used to be.
Where we really need more infrastructure, however, is in the fieldwork tools.
Fieldwork recording devices and data organization software is very heavily skewed
towards a small part of the human language space. We cannot even use these tools
adequately on the array of human language data that linguists routinely collect. The
major tool for linguistic organization is designed for Bible translation. It cannot
store audio, let alone be adapted for other modalities that might be encountered.
Thus, to summarize, language data collection is not straightforward, and it is
just scratching the surface of what good linguistic fieldwork actually is. Field-
work relies on a bunch of shared assumptions about how the world works, about
human interaction, and about language – none of which can be taken for granted
in a ETI encounter. It requires sustained interactions and the building of shared
understanding, and our tools are woefully inadequate to store the data we are
likely to obtain.

Notes
1 I write this as a researcher with twenty years’ experience in doing linguistic fieldwork and
teaching about field linguistics and language documentation, and conducting research on
archival collections. My work involves spoken languages, mostly from Australia and
belonging to Indigenous Australian families (cf. Bowern 2023 for more details). I work
in a paradigm of fieldwork that centers communities.
2 Note that for ethnographic anthropology, the mean method is participant observation,
but linguistic fieldwork usually has a more structured element, as well.
3 Malinowski (2007: 46) writes: “Since you take up your abode in the compound of some
neighbouring white man, trader or missionary, you have nothing to do, but to start at
108 Claire Bowern

once on your ethnographic work.” For discussion of precursors to “linguistic fieldwork”

traditions outside of ethnography, see Thomas (2020).
4 Annotation software like NVIVO allows annotation of documents and collections of
audio and video snippets, but has no way of handling parsed text or collating dictionary
materials.
5 This is slowly changing, but the incorporation of usable STT into a field workflow is still
a long way off.
6 Presumably, such issues have been at least somewhat addressed by transmission from
space telescopes, but these procedures are not currently part of linguistic workflows.
7 It is worth noting that the vast majority of the languages spoken today go back to around
150 languages spoken 10,000 years ago, most which were spoken in one of three areas
where agriculture developed. For all the linguistic diversity still around today, it is a
small slice of the total linguistic diversity in human history, and there is no reason to
believe that it represents the total area of the design space of human language.
8 Some of these ethical considerations are primarily “regulatory” – that is, they stem
from regulations that govern research through federally funded institutions like most
universities.

References
Bird, Steven, and Gary Simons. 2021. Towards an Agenda for Open Language Archiving |
IDEALS. Accessed September 9, 2022. www.ideals.illinois.edu/items/119580.
Bowern, Claire. 2012. A Grammar of Bardi. Berlin and Boston: Mouton de Gruyter.
Bowern, Claire. 2015. Linguistic Fieldwork: A Practical Guide. Dordrecht: Springer.
Bowern, Claire. 2023. The Oxford Guide to Australian Languages. Oxford: Oxford Univer-
sity Press.
Bowern, Claire, and Luke Lindemann. 2021. “The linguistics of the Voynich manuscript.”
Annual Review of Linguistics 7: 285–308.
Brewer, John. 2000. Ethnography. London: McGraw-Hill Education (UK).
Charity Hudley, Anne, Christine Mallinson, Mary Bucholtz, Nelson Flores, Nicole Hol-
liday, Elaine Chun, and Arthur Spears. 2018. “Linguistics and race: An interdisciplinary
approach toward an LSA statement on race.” Proceedings of the Linguistics Society of
America 3: 1–14.
Coon, Jessica. 2020. “The linguistics of Arrival: Heptapods, field linguistics, and Universal
Grammar.” In Language Invention in Linguistics Pedagogy, 32–48. Oxford: Oxford Uni-
versity Press. https://doi.org/10.1093/oso/9780198829874.003.0004.
D’Arcy, Alexandra, and Emily M. Bender. 2023. “Ethics in linguistics.” Annual Review of
Linguistics 9 (1): null. https://doi.org/10.1146/annurev-linguistics-031120-015324.
Everett, Daniel. 2001. “Monolingual fieldwork.” In Linguistic Fieldwork, edited by Paul
Newman and Martha Ratliff. Cambridge: Cambridge University Press.
Haddon, Alfred Cort. 1935. Reports of the Cambridge Anthropological Expedition to Torres
Straits. CUP Archive.
Henke, Ryan E., and Andrea L. Berez-Kroeker. 2016. “A brief history of archiving in lan-
guage documentation, with an annotated bibliography.” Language Documentation 10: 47.
Hockett, Charles F. 1959. “Animal ‘languages’ and human language.” Human Biology. JS-
TOR 31 (1): 32–39.
Kay, Paul, and Luisa Maffi. 1999. “Color appearance and the emergence and evolution of
basic color terms.” American Anthropologist 101: 743–760.
Kibrik, Aleksandr. 2006. “Collective field work: Advantages or disadvantages?” Studies in
Language 30: 259–279.
 Xenolinguistic Fieldwork 109

Ly, Annelise, and Ingrid Onarheim Spjeldnæs. 2021. “Strategies to survive on foreign turf:
Experience sharing and reflections from two apparent aliens in the field.” In Field Guide
to Intercultural Research, 205–216. Edited by David S. A. Guttormsen, Jakob Lauring,
and Malcolm Chapman. Cheltenham: Edward Elgar Publishing.
Macaulay, Monica. 2004. “Training linguistics students for the realities of fieldwork.” An-
thropological Linguistics 46 (2): 194–209.
Malinowski, Bronislaw. 2007. “Method and scope of anthroplogical fieldwork.” In Eth-
nographic Fieldwork: An Anthropological Reader, edited by G.M. Robben and Jeffrey
Sluka. Oxford: Blackwell.
Meakins, Felicity, Jennifer Green, and Myfany Turpin. 2018. Understanding Linguistic
Fieldwork. Abingdon, UK: Routledge.
Ray, Sidney H., and Alfred C. Haddon. 1891. “A study of the languages of Torres Straits,
with vocabularies and grammatical notes (part I).” Proceedings of the Royal Irish Acad-
emy (1889–1901) 2: 463–616.
Rice, K. 2006. “Ethical issues in linguistic fieldwork: An overview.” Journal of Academic
Ethics 4 (1): 123–155.
Rice, K. 2009. “Must there be two solitudes? Language activists and linguists working to-
gether.” In Indigenous Language Revitalization: Encouragement, Guidance, and Les-
sons Learned, edited by J. Reyhner and L. Lockard, 37–59. Flagstaff: Northern Arizona
University.
Rice, Mskwaankwad. 2022. “Power and positionality: A case study of linguistics’ relation-
ship to Indigenous peoples.” Proceedings of the Linguistic Society of America 7 (1):
5295. https://doi.org/10.3765/plsa.v7i1.5295.
Sanker, Chelsea, Sarah Babinski, Roslyn Burns, Marisha Evans, Juhyae Kim, Slater Smith,
Natalie Weber, and Claire Bowern. 2021. “(Don’t) try this at home! The effects of record-
ing devices and software on phonetic analysis.” Language 97: e360–e382.
Sneller, Betsy. 2022. “COVID-era sociolinguistics: introduction to the special issue.” Lin-
guistics Vanguard 8: 303–306.
Stearns, MacDonald. 1978. Crimean Gothic. Analysis and Etymology of the Corpus. Sara-
toga, CA: Amma Libri.
Thomas, Margaret. 2020. “The monolingual approach in American linguistic fieldwork.”
Historiographia Linguistica 47 (2–3): 266–302. https://doi.org/10.1075/hl.00078.tho.
Tsikewa, Adrienne. 2021. “Reimagining the current praxis of field linguistics training: De-
colonial considerations. Language. Linguistic Society of America 97 (4): e293–e319.
https://doi.org/10.1353/lan.2021.0072.
Wells-Jensen, Sheri, and Kimberly Spallinger. 2020. “Extraterrestrial message construc-
tion: Guidelines for the use of xenolinguistics in the classroom.” In Language Inven-
tion in Linguistics Pedagogy, 239–250. Oxford: Oxford University Press. https://doi.
org/10.1093/oso/9780198829874.003.0014.
Williams, Nicholas, Wilson D.L. Silva, Laura McPherson, and Jeff Good. 2021. “Covid-19
and documentary linguistics: Some ways forward.” Language Documentation and De-
scription 20: 359–377.
Wittenburg, P., H. Brugman, A. Russel, A. Klassman, and H. Sloetjes. 2006. “ELAN: A
professional framework for multimodality research.” Proceedings of LREC 2006, Fifth
International Conference on Language Resources and Evaluation, European Language
Resources Association (ELRA), Luxembourg, 1556–1559.
11
INVESTIGATING THE FOUNDATIONS
OF MEANING IN A XENOLANGUAGE
Andrew McKenzie

Introduction
This chapter explores how we might discover the nature of meaning in extrater-
restrial languages, and what that may or may not tell us about how extraterrestrials
conceive their world. Any discovery will build off of three distinct threads of analy-
sis that have only intertwined in the last generation of research. First, we have the
linguistic task of inducing generalizations about observed morphemes; next, the
philosophical goal of understanding how knowledge is encoded and transmitted;
finally, we have the modern linguistic theory of discovering the nature of human
linguistic knowledge as a cognitive object. Each of these strands of modern seman-
tics plays a key role in understanding what we might find out.
Given an alien species, how do we engage in understanding not only the forms
of their languages, but the meanings?1 Clearly, we will need a solid footing in truth-
conditions. We can reasonably assume they have objects like utterances, indexi-
cal forms that relate to them, and also referential forms that point out the objects
around them and elsewhere. Yet we must also understand if and how they quan-
tify over things, express modal claims, or combine modifiers. We must observe
how they organize their discourses, and seek whether their languages reflect that
organization. We must also explore the nature of their semantic ontologies, down
to the most basic levels. Do they distinguish entities that ‘are’ from events that
‘happen’? Do they conceive of modal claims in terms of possible worlds? How do
their languages reflect their cognition of mereology? Do their languages show them
to treat causation differently from us? Do they share similar spatial and temporal
awareness?
These questions and others are not merely curious inquiries of linguistic struc-
ture. In many ways, a speaker’s semantic (and pragmatic) behavior can reveal their

DOI: 10.4324/9781003352174-11
 Investigating the Foundations of Meaning in a Xenolanguage 111

underlying patterns of cognition, as well as their cultural habits and mindsets. Yet
we must take care not to overstate the ways that cognition, language, and culture
are intertwined. So, as we speculate about what meaning xenosemantics might
involve, we must consider a method to find it.

The Truth About Truth

At its heart, semantics is the bridge between the word and the world. Under a
modern generative framework, it links the core linguistic component to the concep-
tual-intentional module of the mind. Under more cognitively oriented approaches,
semantics is simply a part of how we organize thought. In either case, once we
separate the pieces of language that have no near relation to meaning, and we filter
out non-linguistic thought processes and cultural habits, we end up with our field
of study: some body of knowledge held by linguistic beings.
Ascertaining this knowledge is difficult. Any generalization requires us to iron
out variations, but we have rarely – if ever – focused on discerning which varia-
tions even matter, the way we have for sound. When a person emits a speech sound,
a listener’s ears pick up a signal, and the listener’s mind catches upon a variety of
phonetic cues, adjusting for variations within and across speakers. In doing so, that
mind situates that sound within a class (a phoneme) that thus relates it by opposi-
tion to other classes. Linguists who easily worked out how we do this with sounds
also argued that doing the same for the semantics was a fool’s errand. Leonard
Bloomfield pointed out nearly 100 years ago that linguists have to assume that the
concept linked to a word in one person’s mind is the same as that in another per-
son’s mind. How could one possibly know whether my lexical item ‘cat’ actually
means the same as yours? If our personal lexemes do not mean the same thing, how
can we be certain that they are even the same lexical item? He called this assump-
tion the fundamental assumption of linguistics, suggesting a problem that could
not be overcome (Bloomfield 1933). Observation makes us wonder, though, if this
assumption is really problematic. Speakers rarely have difficulties using words like
‘cat’ with each other, so we know either that speakers make this assumption all the
time or that these lexical items are unproblematically non-identical. Perhaps they
vary in ways similar to the ways that non-meaningful linguistic categories vary,
and minds adjust accordingly.
Adjustment does not entail that speakers always agree, but that is also the case
with sounds. U.S. Supreme Court Justice Potter Stewart famously wrote in a judi-
cial opinion for an obscenity case that defining ‘pornography’ was difficult, but
suggests that what we might call semantic intuition guides us toward understand-
ing what the term might comprise (Jacobellis v Ohio, United States 1964: my
emphasis).

I shall not today attempt further to define the kinds of material I understand to be
embraced within that shorthand description, and perhaps I could never succeed
112 Andrew McKenzie

in intelligibly doing so. But I know it when I see it, and the motion picture
involved in this case is not that.

Participants in legal proceedings might obviously want something more pre-

dictable to guide their actions, but outside the narrow proscriptions of the law, “I
know it when I see it” is a very ordinary method that humans use to categorize
their world semantically. The fact that each person sees ‘it’ differently, and thus
knows ‘it’ differently, is not controversial; it even underpins de Saussure (1916
[1995])’s notion of langue, Chomsky (1986)’s notion of e- vs. i-language, and so
on.2 Variations like this do not bother speakers generally, and should not bother
any xenolinguists.
Being a crucial component of linguistic behavior, “I know it when I see it”
is thus a fundamental component of understanding what we can find out about
semantic meaning in humans, and undoubtedly other species as well. We cannot
directly observe what two people’s meaning of ‘cat’ is, but we can empirically test
for sameness by observing whether or not they always ‘see it’ in the same contexts.
If two speakers agreed 100% of the time on whether a number of given objects
were cats or not, we can reasonably conclude that their known meaning of ‘cat’ was
identical to one another’s.
We can draw an analogy to an optometrist, who cannot observe what their
patients are perceiving. Instead, they can only draw conclusions based on the
patients’ verbal behavior: answering “yes” or “no,” reading figures on a Snellen
test, saying when the ball comes into view, etc. Even advanced scanning devices
cannot answer these questions. The doctor can induce what a patient can and can-
not see clearly, and conclude that their visual acuity is the same as anyone else who
gave the same answers at the same stages of the eye test. This in turn leads to an
accurate prescription, even if the people with the same prescription still have slight
variations in vision.
With semantic knowledge, the path to understanding what an alien friend knows
involves finding out when it ‘sees it,’ and that leads us to truth-conditions. As Lewis
(1970) pointed out succinctly, to know the meaning of an expression is to know the
conditions that make it true. Linguistic knowledge of an expression’s meaning is
the knowledge of those conditions, even if a linguist still cannot quite tell exactly
what that knowledge contains.
One might wonder how well truth-conditions can be applied to anything besides
assertions. What makes a question or a command true? Clearly Lewis overgeneral-
ized. That said, the meanings of these other speech acts still rely on truth-condi-
tions. If I ask you “Does Maren drink coffee?” I am essentially laying out for you
a proposition (Maren drinks coffee), but instead of asserting that it has a value of
‘true’; I am instructing you to assert its truth-value to me. The answer still depends
on the same world conditions that would make “Maren drinks coffee” true. If I tell
you “Get Maren some coffee,” my command will only be satisfied if the proposi-
tion that you get Maren some coffee becomes true.
 Investigating the Foundations of Meaning in a Xenolanguage 113

Alternately, some semanticists have instead formulated meaning in terms of its

effect on the speech context: what the speech participants know and have in mind
(Kamp 1981; Heim 1982). Even here, though, the particular effects on the context
is largely rooted in truth-conditions or easily linked back to them.
Besides truth-conditionality, another key aspect of linguistic meaning is com-
positionality: the meaning of a complex expression is built solely out of the
meanings of its parts. Just as speakers do not generally memorize the forms of
entire sentences, they do not generally memorize the meanings of entire proposi-
tions. Instead, we build them each time, starting with the structure built by some
level of the syntax. The mechanisms for composing meaning depend tightly on
the denotations of lexical items, so most of semantic research eventually boils
down to that.

Semantics and Perception

Much of the semantic research into the lexicon over the years has also delved into
some kind of psychology, because it is fairly obvious that lexical meanings directly
relate to how our minds have ‘carved out’ the spaces of reality around us, the way
our vision delineates the borders of objects, and so on. What is less obvious is what
those relations are, how deep they go, and what ‘carving’ even means.
Whitney (1867) figured that “separate articulated signs of thought” help humans
make sense and organize the world around them. More importantly, humans could
share these realities far and wide.

not only were we thus assisted to an intelligent recognition of ourselves and the
world immediately about us, but knowledge began at once to be communicated
to us respecting things beyond our reach.
(Whitney 1867: 13)

This organization is quite evident in terms denoting human-assigned categories,

which are relations that only exist in our knowledge. These include family rela-
tions, which mix biological and cultural notions. We see it in the names of places,
like membership in a mountain range or a continent. Certain objects only exist in
our knowledge, like the constellations in the night sky, or the red dot from a laser
pointer (a collocation of instants of light), and their names reflect a recognition of
our world.
This conception of language taming our thoughts reappears over the years.
Bréal (1897: 271 et seq.), who coined the term sémantique, argues that language
is a translation of reality whose real value comes from how it objectifies our pre-
existing thought by making vague ideas solid enough to transmit.

No doubt it must be the case that the idea came first: but this idea is vacillating,
fleeting, difficult to transmit; once it’s incorporated into a sign, we are more sure
114 Andrew McKenzie

of possessing it, of handling it as we wish, and of communicating it to others.

Such is the service rendered by language: It objectifies thought.3

Saussure’s seminal work went a clear step further with his concept of signe,
which includes a linguistic representation and a mental concept. This is not a ref-
erential link like we would imagine; the one does not exist without the other, so
whatever the actual nature of things, linguistic signs help us carve out the mental
spaces for them in our minds. For de Saussure ([1916] 1995: 99) this leads to con-
clusions we might today consider ‘Whorfian.’

it is clear that only the combinations consecrated by the language system appear
to us to properly fit reality.4

This sort of intricate link between the meaning of linguistic items and the mental
conceptions of humans did not catch on very well among structuralists who soon
dispensed with Saussure’s notion of signe for empirical reasons. Bréal, for his part,
had already emphasized how semantic changes in natural language demonstrated the
flexibility of human thought processes, and their independence from language. Lan-
guage helps us pinpoint and share clear ideas, but it is not a requirement to have them.
Americanists avoided these questions altogether in their main line of research,
eschewing semantic inquiry on positivist grounds. We cannot observe mental
states, so leave that for later research. And in any case, tying one’s linguistics to
particular theories of psychology runs a risk as those come and go. Instead, they
focused on inducing generalizations about the observable usage conditions of the
welter of unheralded morphemes they discovered as they trekked around the globe
documenting languages. This approach is very much like the modern reliance on
observing truth-conditional behavior discussed previously, and their findings with
this method have proven to provide their most lasting results.
Nonetheless, they often came to assume a tight link between meaning and
worldview. Famously, Edward Sapir would claim such a tight link that the seman-
tics directly shaped cognition, as much as vice versa.

Human beings do not live in the objective world alone . . . but are very much at
the mercy of the particular language which has become the medium of expres-
sion for their society . . . the ‘real world’ is to a large extent unconsciously built
up on the language habits of the group . . . . We see and hear and otherwise
experience very largely as we do because the language habits of our community
predispose certain choices of interpretation.
(Sapir 1929: 209–210)

Sapir’s student, Whorf, argued (1956) that a Hopi speaker’s sense of time was dis-
tinct from an English speaker’s because the Hopi language’s temporal semantics
was not built on the same notion of how time progresses.
 Investigating the Foundations of Meaning in a Xenolanguage 115

Such speculative claims, now called linguistic determinism, were not empirically
backed, either by linguistics, psychology, or a cursory examination of multilingual
speech communities. For one thing, Whorf does not seem to have actually asked any
Hopis about their time concepts, or even conducted basic linguistic documentation.
We know now that Hopi temporal semantics is built on the same ordering of times
that every other language is (Malotki 1983). More recently, experimental ‘neo-Whor-
fian’ research finds extremely minor differences in aspects like color discernment or
directional orientation, yet even those results turn out to involve lexical selection
rather than actual differences in perception (Li and Gleitman 2002; Li et al. 2011).

Looking Into the Lexicon

We can be confident that we will discover linguistic facts that reflect differences
in our friends’ cognition, rather than shape it.5 We can begin to explore those dif-
ferences just by asking about lexical items. However, we must be careful with
our methods, because they were designed with human experience in mind. For
instance, one classic idea is to begin fieldwork with a lexical list, such as a Swadesh
list: more than 100 words rooted in universal human experience, including local
person pronouns like ‘you,’ basic elements of matter like ‘water,’ and body parts
like ‘arm.’

Pronouns and numerals are occasionally replaced . . . but such replacement is

rare. The same is more or less true of other everyday expressions connected with
concepts and experiences common to all human groups, or to the groups living
in a given part of the world during a given epoch.
(Swadesh 1950: 157)

Swadesh designed this list for research in historical linguistics, by eliciting words
unlikely to be borrowed. Yet it proves useful for the first moments of fieldwork
because the linguist is more likely to elicit something than by choosing words at ran-
dom. After all, we are asking about experiences common to any human. A ‘hit’ is not
completely likely, though, because a surprising number of languages lack a ‘word’
for ostensibly simple universal concepts – or they have several words where English
has one, making the translation inexact. More frequently, there are mismatches in
lexical spaces from differences in encoding – or, what is an ordinary noun in one lan-
guage is only used as a finite verb with agreement in another, or what is a free word in
English turns out to be a bound stem in the other language, and cannot be expressed
without some other stem to carry it. Still, out of hundreds of words on these lists,
there are bound to be a healthy number of hits that can kickstart a fieldwork process,
simply due to the relative universality of the human experiences the list evokes.6
One immediately sees how a list rooted in human experience can be a prob-
lem with non-Earthlings. Even with a small alphabetical selection of 25 words
(Table 11.1), which might get 20 hits with a human language, I can only see a
116 Andrew McKenzie

TABLE 11.1 Selection of words to elicit from Swadesh (1950)

egg flower green here lake

eye fog hair hit laugh
Far foot hand hunt leaf
father good head husband left hand
fire grass heart ice leg

handful that we could expect an alien would be able to translate at first (in bold-
face). Several others we could expect them to at least understand, but some, like
‘fog,’ ‘flower,’ ‘grass,’ ‘laugh,’ ‘hand’ . . . perhaps those are foreign to them.
Granted, we would not have to use linguistic evidence alone to see what they
perceive, nor should we. Our friends would hopefully consent to a wide variety of
psychological and medical tests to help get a sense of their perception. For linguis-
tics, though, the first days will be tricky. Trickier still, in fact, because the list also
assumes the speaker and linguist already share one common language to use for
inquiry.

Referring to Reference
Although we cannot assume that xenolanguages would refer to or describe things
the same way we do, we can healthily assume they contain methods of reference to
the objects of the world. In this way, we can rely at first on pointing at things in the
room, hoping for a match. Unfortunately, we then reach Quine (1969)’s gavagai
problem: We cannot know, at least at first, that our hope is true. Imagine that I point
at a sitting dog wagging its tail and say “dog.” Do they know I am speaking about
the entity and not some part of it, or some group it is a part of (mammal, animal)?
Or do I mean its color, its furriness, its being alive, the act of sitting, its happiness,
cuteness, puppy-dog eyes, odor, food, or what? Do they even understand that I am
describing a particular object, instead of a generic concept? Or do they think I am
asserting possession (“mine!”), or even offering the dog, saying “you can have it”?
Consequently . . . what would their translation mean?
Experience shows that speakers of distinct languages eventually surmount
mutual unintelligibility, as the existence of pidgins demonstrates. Even the paltry
vocabulary lists of colonial merchant-explorers stand as a testament to overcoming
this hurdle. Jacques Cartier’s 1545 expedition records enough words from the vil-
lages along the St. Lawrence River (including the name of a village, Canada), that
linguists today can tell that the inhabitants spoke an Iroquoian language distinct
from any still known. This list consisted of body parts, person types (man, woman,
child, etc.), flora and fauna, tools and implements, and so on (Cartier 1863). After a
while, with basic vocabulary and a lot of help from gestures, the expedition and the
locals were able to communicate about certain kinds of information (Huchon 2006).
 Investigating the Foundations of Meaning in a Xenolanguage 117

This history suggests to us that groping around the lexicon can get us started,
barring massive differences in ontological perception and awareness. In doing so,
we can start to gain some linguistic evidence about how our friends mentally com-
partmentalize the world around them.

Onto the Ontology

We would need that evidence, because the way our semantics organizes objects
might differ from how our friends’ semantics does. While few semantic univer-
sals have been explored in depth, the nature of semantic investigation has led to a
number of practical assumptions about fundamental universals of human language.
Many of these revolve around the ontology of basic elements (Rett 2022). Building
off Montague, Davidson, and so on, semanticists generally agree at least on the dis-
tinction between entities and events.7 We generally include truth-values, time inter-
vals, and possible worlds, as well. Some semanticists also include situations, kinds,
degrees, and locations, depending on their approach and the theoretical question at
hand. As far as I know, no one has made a full ‘semantics’ of a single language with
all the fully accepted or well-supported components, so it does remain to be seen if
these all truly fit together.
In a type-theoretic semantics, simple atomic objects are members of one of these
ontological sets, and complex expressions denote functions mapping from one set
of objects to another (including itself). We do not question that semantic meaning
deals with truth-values, entities, or events, because those are clearly distinct kinds
of objects. At least, they seem clearly distinct. In all these years, we have not devel-
oped any real criterion for distinguishing them beyond their use in language. Link
(1983: 303f.) suggests that “our guide in ontological matters has to be language
itself.” Entities are described by nouns and adjectives, while events are described
by adverbs and adverbials. Verbs and thematic relations relate entities to events. In
the absence of criteria, we think of entities being objects that ‘are’, while events
are objects that ‘happen’, but it is hard to actually define that difference. So we rely
on entailments and morphological distinctions involving proforms, quantifiers, and
modifiers. For instance, in English, it can be used to refer to antecedent events.

1 “The mayor was caught red-handed and it sank his re-election bid.”
it = the event of the mayor being caught (red-handed)
2 “The silo exploded and I saw it happen.”

it = the event of the silo exploding

In many ways, events and entities can be treated in similar ways as far as plurality
and mereology are concerned (Link 1983; Bach 1986; Krifka 1990). They even
overlap in famous instances, like (3) following, which could no doubt be elicited
118 Andrew McKenzie

from our friends with minor tweaks to the vocabulary. We may find a sentence like
this is not ambiguous for them.

3 “Four thousand ships passed through the lock.”

= Not necessarily 4,000 distinct boats, but events of ships passing
One point of distinction is that relations between events and entities are asym-
metrical. A verb relates entities to events. The entity’s role in the event is deter-
minative― switch it out and the event is gone. In contrast, the event only plays a
small role in determining the entity. If I see you, you contribute far more to this
seeing event being the way it is than it contributes to you being the way you are.
This asymmetry holds even if the event significantly affects the theme – say, I dis-
integrate you,8 your contribution to this event’s being how it is from start to finish
dwarfs the event’s contribution to how you were from start to finish. It also holds if
the event is nominalized: with the creation of the sculpture, the sculpture is a key
component of this event, while the event is only one small part of the sculpture’s
history, as many other events may happen to it afterwards.
Also, the parts of entities that are themselves entities can have parts that
have properties the whole cannot have. Bach (1986: 13) offers this case as an
example:

4 “The gold making up Terry’s ring is old but the ring itself is new.”

Another key difference is causation. The parts of events cause the whole event in
ways that parts of entities do not cause entities. If Jenna climbed Mount Everest,
that event is composed of a large number of sub-events, each of which contributes
causally to the whole. Indeed, we tend to ignore potential sub-events that are not
causal contributors, like stopping to scratch an itch or chatting idly with someone
along the way. On the other hand, none of Jenna’s parts cause her – not her arm
nor her heart. None of those parts’ parts cause them, either, not even down to the
cellular level.
Any language whose ontology we have explored behaves similarly, although
that is not a large set. We might find that alien languages do not work this way at
all; e.g., that the equivalent of ‘old’ cannot apply to the components of something
that is the equivalent of ‘new.’ The preceding differences seem to suggest that peo-
ple distinguish some semantic objects by concepts like causation – but we must not
assume that all species would do the same.

Speaking of Speech Acts

Setting aside the mode of communication our friends employ, and the things they
might talk about, we can safely presume that they will deliver it via speech acts,
or an analog thereof; we can still call them speech acts (Austin 1962). Natural
 Investigating the Foundations of Meaning in a Xenolanguage 119

languages are also observed to vary little in the sorts of speech acts available,
at least in a broad sense: Assertions, questions, apologies, promises, and so on.
We might wonder if our friends would have speech acts we have not observed,
especially performatives. Searle and Vanderveken (2005) point out the lack of
a possible speech act in human languages like “I hereby persuade you,” with a
performative effect, because performatives are rooted in social acceptance. Other
creatures might have that power, though, and if they could use it on us, I suppose
I would agree with them.
We can also assume as a hypothesis that their languages will have indexical
items that express relations toward these speech acts, based on the universality in
human languages. A language without indexicals could logically exist, but indexi-
cality not only makes things far more efficient; it also ropes in the self-aware com-
ponents of our psychology.
Relations between entities and speech acts known as person are features of all
human languages, and a large amount of research finds that cross-linguistic person
marking boils down to the same small set (see Cysouw 2011 for a summary). The
speaker is distinguished from the addressee, though the two can be lumped together
in inclusive plurals. Everything else is being talked about. The small cross-lin-
guistic variation allows for derivation via a powerful feature geometry (Harley
and Ritter 2002), and the tracking of changes over time. We might imagine other
relations, though. We do not, for instance, observe languages that distinguish direct
addressees from other (potential) listeners, themselves distinct from things being
talked about. We also see that the persons are tied to each particular speech act or
utterance, not to entire conversations (first to speak, first to reply, most important
person, etc.). In alien languages, we may well find some other setup, and if it is as
universal for them as our person setup is for us, it may well reflect some aspect of
their cognition.

Composed With Compositionality

The meaning of entire clauses is built from the meanings of its parts, and a large
body of research has sought to see how that works, building off Frege’s idea of
using functional application. A predicate is a function that takes a simple object
as its argument. However, that is not sufficient. Heim and Kratzer (1998)’s well-
accepted formulations of various compositional rules only number five, and
only a handful of narrow types of conjunction have been added to them (Kratzer
1996; Chung and Ladusaw 2004). With this limitation on composition modes,
von Fintel and Matthewson (2008) ask if compositionality can be shown to be
universal for humans. If xenolanguages are not so compositional, it may throw a
serious wrench in our attempts to learn about them. Compositionality fundamen-
tally affects fieldwork, because it allows us to bootstrap upon previous findings
by substituting out single words or morphemes and comparing the meanings of
sentences.
120 Andrew McKenzie

Being Pragmatic
This chapter has focused on semantics, but we must not forget the pragmatics for
understanding usage in a xenolanguage. A significant amount of communication is
indirect. Speech act participants work tirelessly to create and fill deliberate gaps,
but in doing so, they rely on several types of tacit knowledge: cultural background,
personal experience, observations of other conversations, and so on.
We should certainly expect surface pragmatic principles to vary, as we do among
linguistic cultures and subcultures. We should not expect a pragmatics as extreme
as that of the fictional Tamarians who appeared in an episode of Star Trek: The Next
Generation. These humanoids spoke only in cultural references to heroic figures
(e.g., “Shaka, when the walls fell” describes events of failure). Whatever we get,
we must take care to distinguish pragmatic from semantic elements of meaning.
When we scrape away cultural layers, we find basic principles of pragmatics
that we usually call Gricean maxims (Grice 1975). These maxims are rooted in the
assumption that speech act participants are cooperating, even when we are flouting
the maxims. Put another way, the mere act of language is inherently cooperative.
Should we expect the same from our friends, or might the act of a xenolanguage be
inherently something else?

Notes
1 We would also do well to ask: How can we help them discover how our languages work?
2 Never mind what it might mean to know anything in the first place.
3 Translated by the author from French: “Sans doute il a fallu que l’idée précédât: mais
cette idée est vacillante, fugitive, difficile à transmettre; une fois incorporée dans un
signe, nous sommes plus sûrs de la posséder, de la manier à volonté et de la commu-
niquer à d’autres. Tel est le service rendu par le langage: il objective la pensée.”
4 Translated by the author from French: “il est clair que seuls les rapprochements con-
sacrés par la langue nous apparaissent conformes à la réalité.”
5 If the ‘language of thought’ hypothesis holds – of us and of extraterrestrials – whereby
thoughts are built in a separate mental language distinct from the spoken one, then these
questions can all be applied to that.
6 A similar approach involves semantic primes, argued to be the fundamental building
blocks of linguistic meaning and language-built thought (Wierzbicka 1972; Goddard
1999). The same issues come up, but worse because many of those primes are somewhat
abstract.
7 Entities are also called individuals, while events are sometimes called eventualities
which are then divided into events and states. Here we will use the broad event to cover
eventualities including states, even though there is mounting evidence that states are
distinct (see Maienborn 2011 for a discussion).
8 Sorry about that.

References
Austin, J.L. 1962. How to Do Things with Words. Edited by J.O. Urmson and Marina Sbisà.
Oxford: Clarendon Press.
Bach, Emmon. 1986. “The algebra of events.” Linguistics and Philosophy 9: 5–16.
 Investigating the Foundations of Meaning in a Xenolanguage 121

Bloomfield, Leonard. 1933. Language. Chicago, IL: University of Chicago Press.

Bréal, Michel. 1897. Essai de Sémantique. Paris: Hachette.
Cartier, Jacques. 1863. Bref récit et succincte narration de la navigation faite en MDXXXV
et MDXXXVI par le capitaine Jacques Cartier aux îles de Canada. Reprint of original
from 1545. Paris: Librairie Tross. www.gutenberg.org/cache/epub/12356/pg12356-im-
ages.html.
Chomsky, Noam. 1986. Knowledge of Language: Its Nature, Origin, and Use. New York:
Praeger.
Chung, Sandra, and William Ladusaw. 2004. Restriction and Saturation. Cambridge, MA:
MIT Press.
Cysouw, Michael. 2011. “The expression of person and number: A typologist’s perspective.”
Morphology 21: 419–443. doi:10.1007/s11525-010-9170-5
de Saussure, Ferdinand. 1916 [1995]. Cours de linguistique générale. Paris: Payot.
Goddard, Cliff. 1999. “Building a universal semantic metalanguage: The semantic theory
of Anna Wierzbicka.” RASK – International Journal of Language and Communication
9–10: 3–35.
Grice, Paul. 1975. “Logic and conversation.” In Syntax and Semantics 3: Speech Acts. Ed-
ited by Peter Cole and Jerry Morgan, 41–58. New York, NY: Academic Press.
Harley, Heidi, and Elizabeth Ritter. 2002. “Person and number in pronouns: A feature-geo-
metric analysis.” Language 78 (3): 482–526.
Heim, Irene. 1982. The Semantics of Definite and Indefinite Noun Phrases. PhD thesis. Uni-
versity of Massachusetts Amherst, Amherst, MA.
Heim, Irene, and Angelika Kratzer. 1998. Semantics in Generative Grammar. Amsterdam,
the Netherlands: Kluwer.
Huchon, Mireille. 2006. Le français au temps de Jacques Cartier. Rimouski, QC, Canada:
Tangence.
Kamp, Hans. 1981. “A theory of truth and semantic representation.” In Formal Methods in
the Study of Language, edited by Jeroen Groenendijk, Theo Janssen, and Martin Stokhof,
277–322. Mathematical Centre Tracts 135. Amsterdam: Mathematisch Centrum.
Kratzer, Angelika. 1996. “Severing the external argument from its verb.” In Phrase Struc-
ture and the Lexicon, edited by Johan Rooryck and Laurie Zaring, 109–137. Studies in
Natural Language and Linguistic Theory, vol. 33. Dordrecht, the Netherlands: Springer.
doi:10.1007/978-94-015-8617-7_5.
Krifka, Manfred. 1990. “Four thousand ships passed through the lock.” Linguistics and Phi-
losophy 13: 487–520.
Lewis, David. 1970. “General semantics.” Synthese 22 (1/2): 18–67.
Li, Peggy, Linda Abarbanell, Lila Gleitman, and Anna Papafragou. 2011. “Spatial reasoning
in Tenejapan Mayans.” Cognition 120: 33–53.
Li, Peggy, and Lila Gleitman. 2002. “Turning the tables: Language and spatial cognition.”
Cognition 83: 265–294.
Link, Godehard. 1983. “The logical analysis of plurals and mass terms: A lattice-theroetical
approach.” In Meaning, Use and the Interpretation of Language, edited by Rainer Bäu-
erle, Christoph Schwarze, and Arnim von Stechow, 303–323. Berlin: de Gruyter.
Maienborn, Claudia. 2011. “Event semantics.” In Semantics: An International Handbook
of Linguistic Meaning, edited by Claudia Maienborn, Klaus von Heusinger, and Paul
Portner. Berlin and Boston: de Gruyter Mouton.
Malotki, Ekkehart. 1983. “Hopi time: A linguistic analysis of the temporal concepts in the
hopi language.” In Trends in Linguistics: Studies and Monographs, edited by Werner
Winter. Vol. 20. New York: Mouton.
122 Andrew McKenzie

Quine, W.V. 1969. Ontological Relativity and Other Essays. New York: Columbia Univer-
sity Press.
Rett, Jessica. 2022. “A typology of semantic entities.” In Linguistics Meets Philosophy,
edited by Daniel Altshuler. Oxford: Oxford University Press.
Sapir, Edward. 1929. “The status of lingusitics as a science.” Language 5 (4): 207–214.
de Saussure, Ferdinand. 1916 [1995]. Cours de linguistique générale. Paris: Payot.
Searle, John, and Daniel Vanderveken. 2005. Speech acts and illocutionary logic.” In in
book: Logic, Thought and Action, edited by Daniel Vanderveken, 109–132. Dordrecht,
The Netherlands: Springer.
Swadesh, Morris. 1950. “Salish internal relationships.” International Journal of American
Linguistics 16 (4): 157–167.
Quine, W.V. 1969. Ontological Relativity and Other Essays. New York: Columbia Univer-
sity Press.
United States, Supreme Court. 1964. Jacobellis v Ohio, 378 U.S. 185. June 22. Accessed
November 30, 2022. https://supreme.justia.com/cases/federal/us/378/184/.
von Fintel, Kai, and Lisa Matthewson. 2008. “Universals in semantics.” The Linguistic Re-
view 25: 139–201.
Whitney, William Dwight. 1867. Language and the Study of Language. New York: Scribner.
Whorf, Benjamin Lee. 1956. Language, Thought, and Reality. Cambridge, MA: MIT Press.
Wierzbicka, Anna. 1972. Semantic Primitives. Translated by Anna Weirzbicka and John
Besemeres. Frankfurt, Germany: Athenäum Verlag.
12
A LINGUISTIC PERSPECTIVE ON THE
DRAKE EQUATION
Knowns and Unknowns for Human Languages
and Extraterrestrial Communication

Daniel Ross

Introduction
Given the extensive range of potential communication systems that may exist in
the universe, how likely is it that humans would be able to communicate with
aliens? By breaking this question down into its component parts, we can attempt to
generalize from our knowledge of human languages to predict what extraterrestrial
languages might look like. Do extraterrestrials exist? Are they intelligent, and do
they have languages? Can we detect and perceive their signals? Can we understand
their languages?
In 1961, Frank Drake set the stage for the search for extraterrestrial intelligence
(SETI), and today’s messaging extraterrestrial intelligence (METI), by proposing
an equation as a thought experiment about the existence of extraterrestrial intel-
ligence. In the same way, we can attempt to formalize questions about xenolin-
guistics with an equation including the unknown values of the sub-questions of
extraterrestrial communication. After establishing an equation as a starting point,
this chapter surveys current knowledge of human languages, drawing from sources
such as Hockett’s design features, to begin the discussion on possible values for the
terms. Despite the many unknowns of xenolinguistics and possible variation, the
resulting analysis provides an optimistic outlook: It may be possible to communi-
cate with extraterrestrials, at least in principle, if we can overcome certain practical
difficulties such as via technology and assuming mutual understanding of a shared
goal of communication (see also Minsky 1985).

Questions of Extraterrestrial Communication

Without data, the study of xenolinguistics is a challenge. Where do we begin? There
are several different ways that we can approach this question. We can address it

DOI: 10.4324/9781003352174-12
124 Daniel Ross

conservatively (e.g., Ross 2016): say little, with confidence. We can address it spec-
ulatively (as in science fiction, cf. Meyers 1980; and some research, e.g., Chomsky
1983, 2000): say more, but unreliably. Or we can study the question itself, which
is the focus of this chapter. We begin by asking what we know, and then consider
what else we want to know and how to proceed. There are few – if any – certain-
ties but many possibilities. The conservative approach does not get us far enough,
but a speculative approach is too open. Instead, we could consider probabilities, to
evaluate what may be likely properties of extraterrestrial communication.
We must begin, of course, with what knowledge we do have, based on human
language and animal communication here on Earth, attempting to generalize expec-
tations for extraterrestrial communication and identify points of potential variation.
Charles Hockett proposed 16 design features used to contrast human languages
and animal communication (Hockett 1960; Hockett and Altmann 1968; inter alia),1
and these will be one starting point for the current research, referenced throughout
this chapter, where HDF refers to “Hockett Design Feature” (numbered following
Hockett and Altmann 1968). Although these properties in some cases also provide
insight into possible extraterrestrial communication systems we will see that some
are more relevant and others less certain.
One obvious question is what a particular signal or language will be like, but for
practical reasons, this is likely not the place to start. It is too difficult to predict spe-
cific properties because there are too many variables and too many unknowns. More
importantly, it is unlikely that we would encounter any particular, expected com-
munication system, as we will only (at least first) encounter one of many possible
signals. Strategically, then, it is best to be prepared for a variety of communication
types. Otherwise, we should wait and react when a specific signal is detected, espe-
cially because our questions will be more immediately answered then and we will
likely have plenty of time to analyze the signal, given the already great time delays
at interstellar distances even with speed-of-light communication. In the less likely
scenario of face-to-face contact, we would still need time to learn the arbitrary
details (e.g., vocabulary) of the language of any visitors anyway. What we can –
and should – do now is research how to identify signals for effective communica-
tion (or prepare our own). The more general question that research on xenolinguis-
tics must address, and the focus of this chapter, is whether (and how) we would be
able to understand them.

The Drake Equation

In November 1961, Frank Drake organized a discussion of the Space Science Board
of the National Academy of Sciences on Extraterrestrial Intelligent Life and Inter-
stellar Communication in Green Bank, Virginia. He set the stage for research that
still continues today based on an equation composed of seven terms with unknown
values (Table 12.1). This thought experiment identified factors relevant to the
search for intelligent life in our galaxy (Drake 1961, 1962, 2013; Shklovskii and
 A Linguistic Perspective on the Drake Equation 125

TABLE 12.1 Drake’s equation

N = R* × fp × ne × fl × fi × fc × L

N Number of detectable civilizations in the Milky Way

R* Rate of star formation
fp Fraction of stars with planets
ne Number of planets per star suitable for life
fl Fraction of such planets where life emerges
fi Fraction of life that evolves to become intelligent
fc Fraction of civilizations that are detectable (radio broadcast)
L Length such civilizations survive, remaining detectable

Sagan 1966; Rospars et al. 2013; Vakoch and Dowd 2015; inter alia). Importantly,
the goal was not to actually calculate an answer, but to emphasize the importance
of the unknowns and encourage research on these topics.
When multiplied together, these unknown values estimate the number of planets
in our galaxy hosting intelligent life that we might contact or be contacted by. In the
years following the proposal, many variants have been suggested, but research on
the original parameters has continued and advanced substantially. We have a much
better understanding of some terms such as the rate of star formation or the number
of exoplanets in the galaxy, while others such as the frequency of life becoming
intelligent remain mysterious.

A Drake Equation for Linguistics?

Can we do the same for xenolinguistics? Questions about the topic are as old as
the Drake Equation itself (e.g., Freudenthal 1960; Wooster et al. 1966), but much
of the previous work has not been connected or followed a specific direction and
has instead generally repeated the same ideas, not substantively progressing overall
yet mentioning some important topics (e.g., Hockett 1955; Warner 1984; Minsky
1985; Holmer 2013). There are several questions we might ask: Are there xenolan-
guages? How many? This is unknown but already embedded – at least by impli-
cation – in Drake’s original equation. What are these xenolanguages like? This
question is too specific and multifaceted to be used to formulate a general equation.
Instead, let us ask a more fundamental question, introduced earlier: Can we under-
stand them? This is an important question, and actually encompasses those more
specific questions and others. As with Drake’s equation, the goal is not to calculate
an answer but to guide ongoing research. Some general questions to consider: (1)
Do intelligent extraterrestrials exist?; (2) Will we interact with them?; (3) Can we
detect and perceive their signals?; and (4) Can we understand their languages?
With these sub-questions, we can try to determine our likelihood of communicating
with extraterrestrials, as shown in Table 12.2.
126 Daniel Ross

TABLE 12.2 Equation for xenolinguistics

C=N×E×P×U

C Estimates the number of extraterrestrial civilizations with which we may be able

to communicate effectively
N Number of intelligent extraterrestrial civilizations; “answer” to Drake Equation
E Probability of encountering or observing them
P Probability of detecting and perceiving their signals
U Probability of understanding their communication

The first two terms are directly related to the original Drake Equation and SETI
goals; these are not narrowly linguistic questions. The first term, N, is the “answer”
to the original Drake Equation, and therefore links these research questions to
the ongoing work of SETI in general. Drake’s equation may need to be adjusted
slightly to estimate this term relevantly (for example, including contact other than
radio signals, whether face-to-face or other interstellar signals such as optical
SETI/METI), and also by explicitly accounting for the likelihood (or assumption)
that sufficiently intelligent extraterrestrials will have a relevant language for con-
sideration. Several of Hockett’s design features are also applicable at this point
– HDF6: Specialization (communication for communication), HDF7: Semanticity
(reference, meaning), HDF8: Arbitrariness (conventional form-meaning pairings),2
and HDF15: Reflexiveness (language can discuss language). There may be species
in the universe with more basic communication systems, like animals on Earth, but
as argued in Ross (2016), it is those with advanced communication systems that
we are likely to encounter and that are therefore most relevant for discussions of
xenolinguistics.3 We might refer to this as contact-facilitating intelligence. Such a
species would likely be as intelligent as we are, or more so. Indeed, with HDF15 as
a likely feature, we might also assume there are extraterrestrial linguists out there
wondering about how species on such exotic planets such as Earth communicate,
and it would be reasonable to prepare not only for the possibility of encountering
xenolanguages, but also extraterrestrial linguists who share with us the goal of
mutual communication. We should prepare to teach as much as to learn.
The second term, E, is the probability of encountering them, or the probability
of contact, whether via interstellar signals or face-to-face interaction. This is not
a strictly linguistic question either. There are social factors, and many unknowns
in its calculation. However, we can again begin with the original Drake Equation
and make an assumption that if there are extraterrestrial civilizations broadcast-
ing signals (or exploring space, eventually including our solar system), contact is
likely for at least some fraction of our nearest neighbors. This question will not
be addressed in detail in this chapter, but some linguistic factors require a brief
discussion here. Paul Grice’s Cooperative Principle (1967, 1975) states that speak-
ers intend to communicate, and also recognize this in their interlocutors (see also
 A Linguistic Perspective on the Drake Equation 127

Hobaiter et al., this volume [Chapter 3]). Note that this does not assume behavioral
or social cooperation (e.g., arguing or threatening) nor truthfulness (e.g., lying),
merely that the communication is purposeful and comprehension is expected:
speakers work together to create shared meaning. Similarly, in most cases of extra-
terrestrial contact, this should apply although we cannot rely on it in the case of
invasion, or if we are merely objects of study. HDF10: Displacement (discussion
beyond the here and now) would also apply. Any broadcasting (or space-faring)
species must have an interest in what is “out there” and have ways of referring to
it in speech. This also provides a shared context for humans and any species we
might encounter. Beyond this, the remaining questions about this term are more
for SETI/METI in general, not xenolinguistics narrowly, so for now let us assume
there will eventually be contact and focus on whether that encounter would lead to
successful communication.
We therefore set aside questions of the existence of intelligent extraterrestri-
als with languages and the likelihood of contact. The remaining two terms are
the focus of this chapter: Perception and Understanding. Will we perceive their
signals? Will we understand their language? Assuming a context of appropriate,
relevant interaction with two cultures seeking communication, we will return to
these questions after first turning to an overview of the grammatical properties of
human language.

Generalizations From Human Languages

The only experience from which we as linguists can generalize is research on
human languages, as well as comparisons with animal communication systems.
Yet even for human language, linguists agree about almost nothing at an abstract,
theoretical level. There are many open and often controversial questions for the
field: What, if anything, is innate? Do languages share the same structure? How
do we account for cross-linguistic variation? Confusingly, differing opinions (and
alleged conclusions) are voiced equally strongly by objectively qualified research-
ers, and we have not reached a consensus about most topics. These problems are
most apparent when comparing different theories, but the underlying problem is
more profound, down to the basic empirical level. For example, various levels of
structure that are fundamental to even defining different subfields (e.g., “X is the
study of Y”) are controversial. “Words” cannot be defined or identified in a con-
sistent way cross-linguistically (Haspelmath 2011). Syntax is the study of sentence
structure, but “sentence” is rarely defined explicitly and there are good arguments,
for example, that there are grammatical principles operating at something like the
paragraph level (Mithun 2008; Longacre 1979). And are “phonemes” (the indi-
vidual units of sound in phonology) theoretically or cognitively real? For example,
archiphonemes have been introduced to explain the unidentifiable units in merger
contexts (when two or more candidates for a phoneme would not be contrastive
in a given environment, such as /sPɪn/‘spin’ where //P/ might stand for either /p/
128 Daniel Ross

or /b/), and it is unclear how cross-linguistic comparisons of phonology can be

made, given that phonemes are language-specific categories (Maddieson 2018).
Similarly, even morphologists disagree about whether morphology is a distinct tier
of grammatical structure or simply where phonology and syntax meet.
As for making generalizations, almost nothing is shared by all languages, or at
least there are no easily observed – especially surface-level – “universals” (Evans
and Levinson 2009; inter alia). Haspelmath (2007) has gone as far as claiming
that there are no cross-linguistic categories at all. Interestingly, and from a very
different perspective, Chomsky’s Minimalist Program also drastically reduces the
components assumed to be shared by all languages (cf. Ross 2021). Regardless, we
should not expect alien languages to all be the same as human languages, nor each
other. The best we can do is keep an open mind about possibilities for extraterres-
trial communication and look at what features from human language are likely to
generalize and why, such as Hockett did with his design features.

Likelihood of Successful Communication

Returning to the equation proposed in Table 12.2, we must now consider the final
two terms: Perception and Understanding. In fact, these terms should each be
expanded into several components for the following discussion.
(PN + PT) × UG × UM × UC

Perceiving the Signal

We have two distinct opportunities to perceive extraterrestrial communication. PN
represents the probability of our perception of their natural modality of communi-
cation (whether auditory, visual, chemical, or otherwise).4 PT represents the prob-
ability of using technology to encode the language, from writing5 to radio signals.
Either approach would work, so there is no need for both, and this becomes a sum
of probabilities.6 Note that this means that the equation estimates our overall ability
to achieve successful communication of some sort, not whether we will be able to
communicate easily or directly and in the same variety of ways that they do. This is
analogous to being able to either speak or write Chinese, but not both.
Natural Perception (PN) is the first scenario most people would consider for
talking with aliens, but it is also least likely. Even assuming they speak orally,
could we hear the range of frequencies they produce? And could we respond?
More generally, there are many more possibilities for medium of communica-
tion: electromagnetic radiation (including but not limited to visual light), pres-
sure waves (including sound), chemical, tactile, and more. Our speech production
abilities and physical anatomy have co-evolved and are specialized. Even if we
could perceive it, it is unlikely we would be able to produce a natural response.
Consider whale songs, which are similar to spoken language but outside the range
of frequencies we produce (and perceive). There are cases of imitation, such as
 A Linguistic Perspective on the Drake Equation 129

parrots “speaking,” but at the very least these introduce a strong “accent” and also
rely on similar environmental factors in evolution (e.g., atmospheric composition)
and even distantly related genetics (having a tongue, mouth, head, etc.). In fact,
perception might be so difficult that it could affect term 2 (E: probability of initial
contact), because we might not recognize that there is a signal at all. Given the
possible variation in signals, there are three increasingly unlikely requirements:
(1) recognizing the signal (hearing, seeing, etc.); (2) being able to perceive its
components (contrasts); and (3) being able to reproduce the signal. The outlook
for PN is not optimistic, and it is unlikely that we could “speak” their languages
in the typical sense. Given the additional unlikelihood of experiencing natural
communication despite interstellar distances between civilizations, we may wish
to concentrate on PT, except to the extent that a decoded radio transmission might
itself correspond to properties of PN.
Perception via Technological Means (PT) is a viable alternative, either tech-
nology-aided reception/interpretation of natural signals or technological encoding.
For example, any wavelength of electromagnetic radiation, or frequency range of
pressure waves, etc., can be recorded and manipulated to make it perceivable to
us, similar to recording and replaying whale songs at an adjusted frequency so we
can hear them well. The remaining difficulty would be perceiving the necessary
contrasts (e.g., phoneme-equivalents) in the signal. Alternatively, we could entirely
bypass the modality by using written, digital, or other encoding to transmit the
message. Furthermore, assuming the most likely scenario of interstellar communi-
cation at a distance of many light-years, the most likely signal is one that travels at
the speed of light, resulting in a bottleneck restricting the channel to radio waves
or a limited number of similar technological means, such that we may even be able
to predict what sort of signal would be sent. The outlook for PT is much better than
PN, and even likely to succeed.
Regarding likely modalities, Hockett’s design features are a starting point.
HDF3 (Rapid Fading) emphasizes the importance of signals being momentaneous
so that communication can proceed quickly, with one symbolic unit rapidly follow-
ing another, permitting complex signals (Galantucci et al. 2010). More generally,
linearization occurs in a relevant dimension (Ross 2016, 2021). In the case of spo-
ken language, that dimension is time, but spatial dimensions can function similarly,
such as the way symbols are distributed across a page in writing. Chemical signals
do not easily linearize and therefore are less likely to encode complex messages
(see also Kershenbaum, this volume [Chapter 2]). Another relevant principle is
HDF4 (Interchangeability), such that speakers can also hear, although again we
may not be able to do both as the extraterrestrials do. Some features may not apply,
such as HDF2 (Broadcast Transmission and Directional Reception: the signal can
be perceived widely but perceivers can identify its origin) or HDF5 (Complete
Feedback: speakers perceive their own speech); these are likely, but do not apply
to all possible systems, such as chemical injection from speaker to hearer,7 or touch
which would be felt only by the interlocutors. These scenarios are less relevant to
130 Daniel Ross

communication at interstellar distances, which would likely be encoded by electro-

magnetic radiation or similar means, although in interpreting the signal, we should
be aware of the possible effects of different natural modalities.
Potential sources of fundamental variation are introduced by other design fea-
tures of human language. HDF9 (Discreteness) describes distinct units that make
up the speech signal, more like the 1s and 0s of digitally encoded audio than the
analog signal of radio waves. HDF13 (Duality of Patterning) describes the two-
tiered structure of human language: sound patterns and grammatical patterns, inter-
acting at the level of word (or morpheme). It is difficult to even imagine a language
that violated or altered these principles, much less what it would be like to speak
and understand such a language. Note, however, that human sign languages pre-
sent a potential example of violating HDF9 and a challenge for research:8 these
languages are generally unwritten because it is difficult to design a writing system
that captures subtle uses of the physical signing space not strictly constrained to
discrete contrasts comparable to spoken languages.

Understanding Extraterrestrial Communication

Understanding is divided into three parts: understanding the language grammati-
cally (UG), understanding the message cognitively (UM), and understanding in a cul-
tural context (UC). Regarding the grammatical component (UG), we must determine
whether the grammatical structure of extraterrestrial communication is parsable,
and relevantly similar to our own languages so as to be comprehensible. Human
languages are already highly variable, and even allowing for differences, could we
expect that at least some “simple” expressions might overlap in structure? HDF11
(Openness) refers to the ability for language to create new expressions through
modification (e.g., compositionality). As argued in Ross (2016), and now followed
by several others (cf. Roberts et al., this volume [Chapter 15], contra Chomsky
1983, 2000; Samuels and Punske, this volume [Chapter 16]),9 we can add this to
a short, provisional list of truly “universal” features of intelligent communication.
Any sufficiently intelligent communication system would include structures for the
combination of concepts, which is a fundamental part of the mental manipulation
of the world through language, and at the same time, given the otherwise extreme
degree of diversity already represented here on Earth, it is possible that human lan-
guages may not be so unusual in the universe, either. This would also suggest that
at least limited translation (including via paraphrase) between human and alien
languages should be possible.
Some major components of human languages are likely (McKenzie [Chapter
11, this volume]), including: the ability to describe actions/processes (verbs) vs.
entities (nouns); the identification of who did what to whom (argument structure:
cf. Langacker 1986); description and modifiers (adjectives, etc.); and complex
expressions (coordination, embedding).10 There are also some likely challenges,
especially if they are more intelligent than we are:11 consider again extraterrestrial
 A Linguistic Perspective on the Drake Equation 131

communication systems with non-discrete contrasts (violating HDF9) or more than

two tiers of grammatical structure (violating HDF13), as mentioned previously.
Another area for complex grammar comes from metalinguistic structures; that is,
grammatical devices are used to overtly modify other grammatical constructions.
We often use language to discuss language itself (cf. HDF15), but through this pro-
cess, we can also develop systematic ways of extending the grammar. For exam-
ple, even in early, pre-language human communication, restricted to single-word
utterances without grammar, the juxtaposition of two utterances would have had
contextual significance (e.g., “Danger! Predator!” or “Food! Eat!”), and eventu-
ally the linguistic code itself was extended to allow for multi-word utterances of
increasing complexity via the combinatorial structure of compositionality (Ross
2016, 2021; Rizzi 2016). Other types of structure-building allow us to expand the
grammar even further with more complex constructions. Circumlocution is a type
of metalinguistic usage, when we do not quite identify the right words or expres-
sion for our communicative purposes, and sometimes this usage spreads; further,
in usage, metalinguistic expressions are often inserted as parentheticals, like “. . . I
think . . .” within a sentence but not yet directly part of its structure. Metalinguistic
commentary, circumlocutions, and reanalysis of existing forms can lead to lan-
guage change when they conventionalize (Ross 2021).12

1 bahay na maganda (Tagalog: Philippines)

house lnk beautiful
‘beautiful house’ (Scontras and Nicolae 2014: 21)
2 Mom and dad visited for the holidays.
3 John and Mary ate pizza and pasta, respectively.

As shown in Table 12.3, these metalinguistic devices can be described in a hier-

archy of several levels for human languages. Basic modification is obligatorily
encoded with a linking element in Tagalog, as in ex. 1, while logical operators may
add layers to a simple sentence, as in ex. 2, and additional grammatical devices may
act as parsing instructions for the listener to understand the sentence in a particular

TABLE 12.3 Metalinguistic structures in human languages

Level Example constructions

1 Basic compositionality (Ross 2016; Hauser et al. 2002; inter alia), as well as overt
“linkers” marking modification: ex. 1.
2 Logical operations, such as negation (“X is not the case”), coordination,
embedding, etc., expanding and combining phrases or sentences: ex. 2.
3 Complex, layered logical expressions modifying existing expressions and giving
explicit information about interpretation to the listener: ex. 3.
132 Daniel Ross

way, as in ex. 3 where specifically (and only) John ate pizza, while Mary ate pasta.
Notice that each consecutive level builds on the previous one. We must wonder
whether there are intelligent extraterrestrials out there with other features beyond
our imagining and current comprehension. Indeed, this is one possible dimension
of variation, along with violations of HDF9 and HDF13, that we may not yet be
ready to understand, requiring further research.
Understanding the Message (UM) would require cognitive abilities sufficient for
receiving (via language), retaining, and evaluating extraterrestrial thoughts, which
might be more complex than our own. However, we should share some common
ground, including basic expressions that may translate universally. We also almost
certainly share some common interests such as interstellar signals, or space travel,
which may be enough overlap to facilitate some first-contact communication.
Understanding from a cultural context (UC) relies on having some shared
perspectives and motivations. While their culture may substantially differ from
our own, our shared interests will provide a bridge for some topics of discussion
(again, consider HDF10: Displacement, discussion beyond here and now). There
are, however, many less certain aspects of linguistic culture that do not necessarily
interfere with our ability to effectively communicate with them. HDF12 (Tradi-
tion) establishes cultural, rather than genetic, transmission as the way a specific
language is passed from one generation to the next. Cultural transmission is likely
to be encountered in extraterrestrial communication systems, as the product of the
natural development of communication through usage. Alternatives are not incon-
ceivable, such as an intelligently designed, perfected communication system for
an intelligent species that has moved beyond traditional communication, or per-
haps one that developed entirely genetically through evolution including all details
down to vocabulary and pronunciation, which might be an indication of a differ-
ent type of intelligence, less creative and more instinctive.13 HDF16 (Learnabil-
ity) states that speakers of one language can learn another, which seems relatively
likely within species, but less likely across species, such that at best what we would
hope for with inter-species communication would be something like a more exag-
gerated effect of second language learning.14 HDF14 (Prevarication) describes the
ability to manipulate the truth, e.g., to lie. It is unclear whether and to what extent
these features would be found in extraterrestrial languages. Regarding pragmatics
in general, we might expect it to apply similarly, given the Cooperative Principle,
from which many other components of pragmatics can be derived.15
In summary, for understanding grammar, there would likely be enough overlap
for at least basic messages; for understanding the message cognitively, our shared
interests will lead to some similar thoughts; and understanding in a cultural context
will be challenging, but shared interests will also help.

Outlook
Successful communication with extraterrestrials depends on many factors,
but the possibility of cognitive capability for communication is supported,
 A Linguistic Perspective on the Drake Equation 133

TABLE 12.4 Preliminary values for xenolinguistics equation

C = N × E × (PN + PT) × UG × UM × UC

Underlined (C, N, E) Unknown

Italic (PN) Unlikely
Bold (PT, UG, UM, UC) Likely (at least partially)

highlighting the need for further research. Assuming there are intelligent extra-
terrestrials (N) and we will eventually encounter them (E), perception through
technological means is likely possible (PT), and understanding is likely to be
at least partially effective in the relevant domains (UG, UM, UC). This means
that overall the probability for successful communication using extraterrestrial
languages – or for them to communicate using ours – is a relevant possibility,
as shown in Table 12.4.
The equation proposed in this chapter describes only the possibility for
communication, not the difficulty or means of learning to communicate. I
have not discussed deciphering a signal we receive (or learning a language
via face-to-face communication). The two key steps for this will be in iden-
tifying the contrastive units within the signal, as already briefly discussed, as
well as understanding the intention and meaning behind the message (on the
difficulty of this task, see Slobodchikoff, this volume [Chapter 9], as well as
Pepperberg, this volume [Chapter 6]). This is of course facilitated in face-to-
face communication because trial and error in interaction can lead to additional
information, but will be hindered by the essentially one-sided discourses pos-
sible at interstellar distances; on the other hand, as mentioned previously, this
will provide us with relatively more time to analyze and decode such a trans-
mission, given that even a generation of research by humans would be only a
minor delay for communication at a distance of 100 light-years, for example.
We also must hope that whoever created the transmission designed it to be
intelligible, beyond the mere announcement that there is someone out there
transmitting.16 Regardless, in ideal circumstances, with an optimally designed
signal carrying enough information to allow analysis and decoding, it seems
likely that we could, in principle, learn to interpret an extraterrestrial language,
although actually achieving this could be difficult for practical reasons (see
Slobodchikoff, this volume [Chapter 9]).
Despite the extremes in possible variation and the many unknowns, the outlook
is optimistic. If we do encounter intelligent extraterrestrials, there may be no inher-
ent cognitive barriers to achieving at least basic communication. Therefore, human
languages may not be so unusual for the universe, either. It is my hope that the dis-
cussion presented here is a starting point for continued research on xenolinguistics
so that some of the unknowns may eventually become knowns. In turn, this will
also lead to a better understanding of our own languages here on Earth.
134 Daniel Ross

This volume provides a foundation for continued research on the language-

specific terms of the equation proposed here, as follows.

E: Pepperberg (Chapter 6) and Slobodchikoff (Chapter 9) discuss strategies and

challenges for first contact. Hobaiter et al. (Chapter 3) consider the effects
of intentionality in communication on interaction. Bowern (Chapter 10) dis-
cusses the applicability of fieldwork on human languages to extraterrestrial
contexts.
PN: the question of natural modality of communication is surveyed by Kershen-
baum (Chapter 2) and Nixon and Tomaschek (Chapter 17), including effects
of physical environment. Wells-Jensen (Chapter 13) specifically considers the
role of blindness in the development of possible extraterrestrial communication
systems.
PT: Herzing (Chapter 4) also discusses natural modalities as well as technological
means to interpret them, and Harbour (Chapter 18) considers the utility of writ-
ing as a tool in METI.
UG: Roberts et al. (Chapter 5) and Samuels and Punske (Chapter 16) explore
implications and applicability of theories of human language on extraterrestrial
grammar.
UM: McKenzie (Chapter 11) considers potential overlap between conceptual mean-
ing in human and extraterrestrial languages. Sperlich (Chapter 14) examines the
possible effects of artificial intelligence in extraterrestrial communication.
UC: Granger et al. (Chapter 8), Ortner (Chapter 5), Berea (Chapter 7), and Nixon
and Tomaschek (Chapter 7) discuss diversity, expectations, and interaction from
a cultural perspective, as well as resulting differences in grammar and cognition.

Overall, these contributions tell us to expect diversity yet predict some over-
lap with certain familiar properties of human languages. They also demonstrate
the interrelatedness of the different factors, in turn emphasizing the importance of
interdisciplinary research.

Notes
1 For Hockett’s own perspective on these questions, see also Hockett (1955), at least for
some speculative ideas.
2 Beyond those ideas that can easily be expressed with imitative forms, symbolic com-
munication is required, and this in turn can lead to and support more complex structure
(cf. Roberts et al. 2015; Little et al. 2017).
3 An exception would be the scenario of humans as traveling to exoplanets and encounter-
ing life face to face, where a wider range of less sophisticated communications systems
would be expected.
4 HDF1 (Vocal-Auditory Channel) must be expanded, as well as for manual-visual signed
languages on Earth.
 A Linguistic Perspective on the Drake Equation 135

5 See also Harbour (this volume [Chapter 18]) for discussion of writing systems applied
to extraterrestrial communication.
6 Technically, this part of the equation should be specified as PN + T or PN + PT − PN & T to
avoid inflating the probability if both possibilities have a high likelihood. In practice,
especially in scenarios such as communication at interstellar distances that restrict our
options, we might focus on only one possibility.
7 Compare for example lateral gene transfer (e.g., spreading antibiotic resistance in bacte-
ria: Gyles and Boerlin 2014).
8 See also Kershenbaum (this volume [Chapter 2]) for discussion of non-discreteness in
animal communication.
9 My analysis was originally proposed at a conference workshop in 2016, followed by the
2018 meeting which resulted in the current volume and where others presented similar
ideas; Ross (2016) has also been submitted for the proceedings of the first workshop, but
publication of that volume has been delayed beyond that of this volume.
10 Human languages differ in how they render these functions, and an important question to
consider is to what extent human languages already span the possibility space of gram-
matical strategies that might be found in the universe.
11 Conversely, would it be easier for them to learn our languages? An important question is
whether variation in grammatical structure is primarily hierarchical, with more complex
languages for more intelligent species, or multi-dimensional, such that the only overlap
might be in the more basic elements. The answers to these questions are unknown, but
important for continued research on the topic.
12 Related to this point, Peterson (2018) argued that analogy may be a more important
feature of human language than even compositionality, because it is what allows us to
take existing structures and adapt them to new usage, leading to the grammaticalization
of new grammatical constructions to fill communicative needs, and also along the way
adding intricacy and even irregularity to the grammatical system, which are hallmarks
of human language. These features are also likely in extraterrestrial communication (see
also later in this chapter on HDF12).
13 See also Sperlich (this volume [Chapter 14]) on possible relationships between artificial
intelligence and extraterrestrial communication.
14 Incidentally, the premise of the movie Arrival (2016) should thus be questioned, which
is to say – without spoiling the ending of an interesting movie – that humans are unlikely
to become native-like speakers of exotic alien languages.
15 Another topic not addressed here is the diversity of languages within a species or planet,
a question for which Hook (1999) relevantly suggests “a Drake Equation for linguistic
diversity” based on sociolinguistic parameters.
16 For design considerations see Harbour (this volume [Chapter 18]).

References
Chomsky, Noam. 1983. “Interview: Noam Chomsky: The revolutionary linguist who be-
lieves we have a ‘language organ’ in our brain also thinks there are some things no
amount of learning can teach.” Interviewed by John Gliedman. Omni 6 (2), November:
113–118, 171–174.
Chomsky, Noam. 2000. New Horizons in the Study of Language and Mind. Cambridge:
Cambridge University Press.
Drake, Frank. 1961. “Extraterrestrial intelligent life and interstellar communication: An in-
formal discussion.” Space Science Board, National Academy of Sciences, Green Bank,
Virginia, November.
Drake, Frank. 1962. Intelligent Life in Space. New York: Macmillan.
136 Daniel Ross

Drake, Frank. 2013. “Reflections on the equation.” International Journal of Astrobiology 12

(3): 173–176. https://doi.org/10.1017/S1473550413000207.
Evans, Nicholas, and Stephen C. Levinson. 2009. “The myth of language universals: Lan-
guage diversity and its importance for cognitive science.” The Behavioral and Brain
Sciences 32 (5): 429–448. https://doi.org/10.1017/S0140525X0999094X.
Freudenthal, Hans. 1960. LINCOS: Design of a Language for Cosmic Intercourse, Part I.
Amsterdam: North-Holland.
Galantucci, Bruno, Christian Kroos, and Theo Rhodes. 2010. “The effects of rapidity of
fading on communication systems.” Interaction Studies 11 (1): 100–111. https://doi.
org/10.1075/is.11.1.03gal.
Grice, H. Paul. 1967. “Logic and conversation.” William James Lectures, Harvard University.
Grice, H. Paul. 1975. “Logic and conversation.” In Syntax and Semantics 3: Speech Acts,
edited by Peter Cole and Jerry L. Morgan, 41–58. New York: Academic Press.
Gyles, C., and P. Boerlin. 2014. “Horizontally transferred genetic elements and their role
in pathogenesis of bacterial disease.” Veterinary Pathology 51 (2): 328–340. https://doi.
org/10.1177/0300985813511131.
Haspelmath, Martin. 2007. “Pre-established categories don’t exist: Consequences for
language description and typology.” Linguistic Typology 11 (1): 119–132. https://doi.
org/10.1515/LINGTY.2007.011.
Haspelmath, Martin. 2011. “The indeterminacy of word segmentation and the nature of
morphology and syntax.” Folia Linguistica 45 (1): 31–80. https://doi.org/10.1515/
flin.2011.002.
Hauser, Marc D., Noam Chomsky, and W. Tecumseh Fitch. 2002. “The faculty of language:
What is it, who has it, and how did it evolve?” Science 298 (5598): 1569–1579. https://
doi.org/10.1126/science.298.5598.1569.
Hockett, Charles F. 1955, May. “How to learn Martian.” Astounding Science Fiction 55 (3):
91–106.
Hockett, Charles F. 1960. “The origin of speech.” Scientific American 203: 89–96. https://
doi.org/10.1038/scientificamerican0960-88.
Hockett, Charles F., and Stuart A. Altmann. 1968. “A note on design features.” In Animal
Communication: Techniques of Study and Results of Research, edited by Thomas A.
Sebeok, 61–72. Bloomington: Indiana University Press.
Holmer, Arthur. 2013. “Greetings earthlings! On possible features of exolanguage.” In
The History and Philosophy of Astrobiology: Perspectives on Extraterrestrial Life and
the Human Mind, edited by David Dunér, 157–183. Newcastle upon Tyne: Cambridge
Scholars.
Hook, Peter Edwin. 1999. “A Drake equation for linguistic diversity.” October. www-per-
sonal.umich.edu/~pehook/drake.html.
Langacker, Ronald W. 1986. “Settings, participants, and grammatical relations.” In Pro-
ceedings of the Second Annual Meeting of the Pacific Linguistics Conference, November
1986, University of Oregon, Eugene, Oregon, edited by Scott DeLancey and Russell S.
Tomlin, 1–31. Eugene: Department of Linguistics, University of Oregon.
Little, Hannah, Kerem Eryılmaz, and Bart de Boer. 2017. “Signal dimensionality and the
emergence of combinatorial structure.” Cognition 168: 1–15. https://doi.org/10.1016/j.
cognition.2017.06.011.
Longacre, Robert E. 1979. “The paragraph as a grammatical unit.” In Syntax and Semantics
12: Discourse and Syntax, edited by Talmy Givón, 115–134. New York: Academic Press.
 A Linguistic Perspective on the Drake Equation 137

Maddieson, Ian. 2018. “Is phonological typology possible without (universal) categories?”
In Phonological Typology, edited by Larry M. Hyman and Frans Plank, 107–125. Berlin:
De Gruyter. https://doi.org/10.1515/9783110451931.
Meyers, Walter Earl. 1980. Aliens and Linguists: Language Study and Science Fiction. Ath-
ens: University of Georgia Press.
Minsky, Marvin. 1985. “Why intelligent aliens will be intelligible.” In Extraterrestrials:
Science and Alien Intelligence, edited by Edward Regis, 117–128. Cambridge: Cam-
bridge University Press.
Mithun, Marianne. 2008. “The extension of dependency beyond the sentence.” Language 84
(1): 69–119. https://doi.org/10.1353/lan.2008.0054.
Peterson, David J. 2018. “Dothraki, language creation and other fascinations.” Paper pre-
sented at the METI International Symposium on Xenolinguistics at the National Space
Society’s 37th Annual International Space Development Conference (ISDC) 2018, Los
Angeles, May 26.
Rizzi, Luigi. 2016. “Monkey morpho-syntax and merge-based systems.” Theoretical Lin-
guistics 42 (1–2): 139–145. https://doi.org/10.1515/tl-2016-0006.
Roberts, Gareth, Jirka Lewandowski, and Bruno Galantucci. 2015. “How communica-
tion changes when we cannot mime the world: Experimental evidence for the effect
of iconicity on combinatoriality.” Cognition 141: 52–66. https://doi.org/10.1016/j.
cognition.2015.04.001.
Rospars, Jean-Pierre, Florence Raulin-Cerceau, and Rocco Mancinelli, eds. 2013. “The
Drake equation.” International Journal of Astrobiology 12 (3, Special issue).
Ross, Daniel. 2016. “Modification as a universal property of intelligent communication.”
Presented at the the SETI Plenary Session at the National Space Society’s 35th Annual
International Space Development Conference (ISDC), San Juan, May 18.
Ross, Daniel. 2021. Pseudocoordination, Serial Verb Constructions and Multi-Verb Predi-
cates: The Relationship between Form and Structure. Ph.D. dissertation, University of
Illinois Urbana-Champaign, Urbana. https://doi.org/10.5281/zenodo.5546425.
Scontras, Gregory, and Andreea C. Nicolae. 2014. “Saturating syntax: Linkers and modifica-
tion in tagalog.” Lingua 149: 17–33. https://doi.org/10.1016/j.lingua.2014.05.005.
Shklovskii, I.S., and Carl Sagan. 1966. Intelligent Life in the Universe. San Francisco:
Holden-Day, Inc.
Vakoch, Douglas A., and Matthew F. Dowd, eds. 2015. The Drake Equation: Estimating the
Prevalence of Extraterrestrial Life through the Ages. Cambridge: Cambridge University
Press. https://doi.org/10.1017/CBO9781139683593.
Warner, Richard. 1984. “Exolinguistics: State of the art.” Omni 6 (11), August: 91.
Wooster, Harold, Paul L. Garvin, Lambros D. Callimahos, John C. Lilly, William O. Davis,
and Francis J. Heyden. 1966. “Communication with extraterrestrial intelligence.” IEEE
Spectrum 3 (3): 153–163. https://doi.org/10.1109/MSPEC.1966.5216745.
13
COGNITION, SENSORY INPUT, AND
LINGUISTICS
A Possible Language for Blind Aliens

Sheri Wells-Jensen

Introduction
Reading and writing about extraterrestrial languages, before we have discovered
even a single extraterrestrial microbe, is a radical act of hope.
Every chapter here quickens with the belief that we could become conversa-
tional partners – perhaps even friends – with intelligent beings born on planets
circling other stars. Our task here is to facilitate that eventual conversation, to lay
the groundwork for its success and anticipate the problems we might face.
For many of us, the first step in this process is to examine human languages and
push back the barriers between ourselves and what we know about how humans
began to speak. Linguists have struggled with this question for decades, sometimes
with collegial gentleness and sometimes without – but always with vigor. We have
wanted to determine specifically what part (if any) of the human ability to use lan-
guage is innate and separate from cognition. Answers to this question have ranged
from “quite a lot” (Chomsky 1986) to “maybe nothing” (Everett 2016, 2017; Evans
2014) to “we just can’t really know” (Hauser et al. 2014), and we are not much
closer to determining the truth than was the Pharaoh Psamtik in 700 BCE, who is
said to have isolated two children with the purpose of discovering the first, most
fundamental language.
The innate component, popularly known as the “language instinct” (Pinker
1994), is the first of three factors pertinent to language acquisition and develop-
ment (O’Grady 2012; Chomsky 2005). This language-specific mutation, if it exists,
would have arisen in the human genome 50,000–150,000 years ago, coincident
with the evolution of Homo sapiens.
“Second-factor variables” are the experiences babies have with language. It
amounts to their exposure to fluent speakers and their experiences as active learners.

DOI: 10.4324/9781003352174-13
 Cognition, Sensory Input, and Linguistics 139

Everyone agrees that this is necessary, but how exactly children utilize this input
to learn languages, and what kind of assistive bootstrapping and strategizing they
need, remains a topic of ongoing research (Fletcher and MacWhinney 2017)
“Third-factor variables” include characteristics of the environment such as
surrounding noise level, air or water pressure, amount of gravity, and chemical
makeup of the atmosphere.
Also included here are traits within the language learners such as overall body
shape, short term memory ability, computational power, amount and kind of mem-
ory storage, and kinds of sensory inputs.
The question of the relative influences of these three factors is crucial to every
chapter in this volume. If human language is deeply influenced by body shape and
the environment in which we find ourselves, then we might expect roughly bipedal
beings from rocky worlds similar to our own to share some of our cognitive (and
thus linguistic) characteristics (Minsky 1985). Maybe we could learn to speak to
them. On the other hand, beings who do not share the humanoid body plan or
who come from very different physical environments may “speak” languages that
would be forever incomprehensible to us (Haden-Elgin 1984).
If language is not closely woven into the physical, biological, and cognitive envi-
ronment – that is, if first-factor variables are significant – we would arguably have an
equally wide set of possibilities but for a different reason. If the genetic mutation that
brought language to humans is unique, then it is possible that humans would never
be able to learn any language that did not evolve on Earth. However, if our language
instinct is standard for intelligent beings, we might be able to speak to anyone at
all, regardless of the characteristics of their home world, their kind of cognition, or
their body plan. (It’s worth mentioning here that, of course, that there is no reason
to assume that there is only one “right” answer to this question; it is easy to imagine
a universe in which some species rely heavily on a “language instinct” while others
develop language as a natural extension of their nonlinguistic cognitive processes.)
Given that we do not know what comprises Factor 1, and we agree that Factor
2 is important (if not completely understood), one of the things we can do while
we wait for an exemplar of an extraterrestrial language is choose a third-factor
variable, carefully isolate it, and work through how it might influence language
structure and use.
Some of this third-factor work has been begun by our best science fiction writ-
ers, accustomed as they are to world building that is both creative and logically
consistent. We might, as Mary Doria Russell (1996) has done in The Sparrow,
explore a cultural scenario on a planet where a predator and prey species both
evolve to sentience. Or, it might be useful, as Becky Chambers (2014) has done
in the Wayfarer series, to explore how standards of reference to gender and ability
evolve when several species live and work together. Or, as James Cambias (2014)
has done in his novel The Darkling Sea, imagine the cultural and linguistic effects
of being a large, water-dwelling squid-like creature so far beneath the surface of an
ocean that there is no usable light.
140 Sheri Wells-Jensen

As I personally have no experience (as a human) being squid-shaped, but sub-

stantial experience (as a blind human) living in the world without chronic reliance
on visible light, I will pursue this third-factor question: What specific effect, if any,
would species-wide blindness have on the structure of a language?

Defning and Isolating a Single Third-Factor Variable

The most important thing to establish before starting on this process is that this is
not a list of adaptations that blind humans need in order to live and work success-
fully in human society. Blind humans have held nearly every conceivable job on
Earth, from teacher to athlete, general, carpenter, parent, judge, and college profes-
sor. What we are investigating here is how beings who had no exposure to sight,
left on their own without sighted influence, might use language to maximize the
amount and kind of information they receive from one another.
To isolate a single variable, we start by holding everything else constant on our
hypothetical planet.
That is, Earth is the starting point.
In our case, we assume that our blind aliens evolved on the surface of a rocky
world much like ours, with a large moon, oxygen atmosphere, and plentiful food
and water, and that they are humanoid with roughly human physical abilities. (For
more on what might be constant across cultures, consult Granger et al., this volume
[Chapter 8] and Herzing, this volume [Chapter 4].)
Although I am confident that one could build a convincing case for the survival
and advancement of a totally blind species (Wells-Jensen 2016), whether or not
this is a likely scenario is not the focus here; our purpose is to isolate and explore
one third-factor variable that could conceivably make a significant difference in
language evolution and follow that logical trail. (Consult both Kershenbaum, this
volume [Chapter 2], and Pepperberg, this volume [Chapter 6], for ways in which
sensory modality and embodiment might influence cognition.)
Thus, we take blindness as a developmental influence seriously, and we allow the
blind aliens to develop reasonable (even ingenious) non-visual skills in response to
their environment, but do not grant them superpowers. This difference between skills
and powers is essential. Our blind aliens will be able to hear, smell, and feel at more
or less the level of humans. They may train their existing abilities more cleverly and
deliberately and may pay more attention to auditory or tactile information, but the
level of sensory input itself will not differ from that which is typical for humans.
Most blind people use some degree of passive sonar to detect openings in a cor-
ridor as they walk along (Kish 1982), but unlike Daredevil, they regrettably cannot
“hear” the heartbeats of their conversational partners or listen in on the blood roar-
ing through their veins. Similarly, a blind human, with training, can learn to read
braille rapidly and quickly, using a skill set that an untrained sighted person cannot
readily duplicate (Miller 1997), and this early exposure to increased tactile stimula-
tion can affect the allocation of resources in the human brain. That is, while blind
 Cognition, Sensory Input, and Linguistics 141

and humans and sighted humans have the same brains, congenitally blind people
allocate brain resources differently than sighted peers (Bedny 2017; Burton et al.
2002). This does not convey superpowers, but it does speak to the strong influence
of natural circumstances over brain allocation.
In addition, when necessary (or simply desirable), humans can train themselves
to accomplish impressive feats of memory (Chalmers and Humphreys 2017); some
chess masters, for example, can retain the positions of all the pieces on a chess
board and can, in fact, play several games of “blindfold” chess at once (Saariluoma
1991; Saariluoma and Kalakoski 1998).
The difference between spatial ability (shared by all humans) and visual ability
(something sighted people have, and blind people presumably do not) is important
here. Much of what we generally identify as “visual” ability is, in fact a complex
mix of mental abstractions (found in the parietal lobe) and stored knowledge of
the properties of objects (found in the temporal lobe) (Kosslyn et al. 2006). Visual
knowledge is a subset of spatial knowledge, and we know that spatial awareness in
blind people positively correlates with amount of exploration and physical activity
(Schmidt et al. 2012), meaning that the ability to hold spatial information in the
mind improves with practice.
For example, I know, sitting in my front yard (although I have never seen these
things), that I am roughly thirty feet from the street that runs north and south in
front of my house. I know that my front door is behind me, off to my left, and
that the apple tree is forward and to my right. I can track the passage of a loudly
complaining chickadee as she passes more or less directly over my head, flying
parallel to the street. Similarly, I can hold in my mind the form of the chickadee,
remembering her two clawed feet, her legs, and the body above them, and think
accurately about how her wide-open mouth is on the opposite end of the body from
her tapered tail.
I can compare the size of the chickadee to the very interested cat sitting beside
me, and I can tell by resting my hand on the cat’s head that he, too, is moving his
attention across the sky, following the bird’s path. Sima et al. (2013: 1) have con-
firmed experimentally what blind people know instinctively: “The visual mental
image is a specification of a part of a spatial mental image” (consult also Kosslyn
et al. 2006).
All this notwithstanding, much of daily life can be effectively navigated with lit-
tle to no spatial information (Hull 1990) – and this, as it turns out, is generally not
a problem, at least in the life of a relatively sedentary 21st-century person. Spatial
representations might only be called upon when needed. For example, despite my
spatial description of my front yard, I do not create or maintain representations of
physical data I do not need.
If I am playing chess, I hold the grid of the board in my mind and understand
the relationships of the pieces to each other. If I am faced with an opponent in a
snowball fight, I use sound cues to track their movements with reference to where
I am waiting.
142 Sheri Wells-Jensen

But if a colleague knocks on my office door and says hello, I hear and interpret
the knock, identify her voice, and know her to be standing in the hallway just
beyond my door without imaging her knuckles, the door frame, or any part of her
body. As we are chatting, I make no effort to picture her body size, face, dress, or the
position of her arms and legs. However, when she stands up and offers me a sheaf
of papers (signaled by our conversational context, by sounds of clothing against
skin and of a chair moving or creaking as weight shifts off of it), I deploy spatial
knowledge and can judge accurately where to reach to receive the proffered pages
from her hand, but I do not bother to construct a representation of her extended
arm, or the rest of her body, or her abandoned chair, unless these become important.
Thus, while spatial information is generally retrievable from the environment, it is
not as essential in most situations as one might imagine. We also want to tease apart
the inherent usefulness of vision from the ways in which humans have constructed
our shared environment to privilege vision over other sensory inputs.
For example, if we want to communicate what is behind a series of closed doors
in an office building, we generally place visual representations on the doors in the
form of printed words and pictures, and potential clients are expected to find their
way to their goal by searching through these possibilities. We do this because it
is assumed that the vast majority of the people entering the office building would
be sighted. However, this is only one possible way of communicating the relevant
information. Signs on the doors could be tactile instead, or the carpet or floor cover-
ing in the hallway could contain tactile cues, detectable by the feet. Each door could
carry a small audio beacon explaining its contents. It could be standard to provide
a succinct directory to anyone entering the building to avoid the problem of search-
ing for the right door. Each business, when it advertises itself to the public, could
include detailed instructions on how to locate its entrance. Building contents could
be organized such that similar businesses are placed adjacent to one another, or
there could be standard places within each building for specific kinds of services –
the restaurants always on the highest floor and the shoe stores at the bottom.
In fact, if one or more of these factors were employed in addition to the visual
cues, the office building would be easier for everyone to use. (For more discussion,
consult Davis 2017.)
Finally, although we are collectively relying to some degree on imagination as
we progress through this scenario, research has shown that a sighted person cannot
duplicate the sensory, emotional, or practical world of a blind person by closing
his eyes or donning a blindfold. The perceptions, abilities, and reactions of a blind
person are not the same as those of a sighted person who is not looking (Burgs-
tahler and Doe 2004; French 1992; Silverman et al. 2014; Silverman 2015). This
is primarily due to the sighted person’s lack of adaptive skills; a blind person is
accustomed to working around barriers such as the office building’s doors and has
innumerable strategies and hacks to find her way.
The troubles the blindfolded sighted person has are also partly due to the anxi-
ety felt by most sighted people under a blindfold, and partly because of cultural
 Cognition, Sensory Input, and Linguistics 143

misperceptions about blindness. Perhaps because sight is so useful, these misper-

ceptions about blindness are very deeply woven into all human cultures.
Although blind humans competently and safely cross busy streets, care for chil-
dren, hold jobs at universities, and are in charge of homes and offices, the pre-
dominant cultural stereotype is that blind people are slow, passive, and incapable
of grace. Rather than an athlete, parent, or even college professor, most sighted
people imagine blind people as elderly relatives, beggars, or, on a good day, blues
musicians (Wells-Jensen et al. 2021).
To rationally consider how blindness would shape language and culture, we also
have to untangle blindness from its cultural baggage.
What remains, then, as a useful definition of blindness is a particular set of gaps
in useful information.

• Location of objects at a distance when those objects are not emitting noises and
cannot be easily detected through passive sonar (especially important if these
are dangerous, such as a hole in the ground or overhanging branches that could
strike a passerby on the head).
• Identification of small objects at a distance.
• Location of two or more objects with respect to one another at a distance.
• Identification of individual people at a distance.
• Detection of what a person is doing if that action is not audible (this might be
especially important for any situation where passive observation leads to signifi-
cant cultural cohesion, safety, or advancement of scientific knowledge).

A Blind Alien Language

Having established what blindness is, and the gaps it might cause in information
flow for the blind aliens, we turn to a discussion of how language might fill those
gaps.
I do not in any way mean to imply that blind people on Earth are linguisti-
cally impoverished, or that they need these enhancements to manage their affairs.
Rather, this is an exploration of what blind people could develop without sighted
influences.

Content Words: Nouns, Verbs, and Adjectives

Minsky (1985) explains that any language will have some way of referring to
objects (“nouns”) and some way of referring to events (“verbs”). As discussed by
Roberts et al. (this volume [Chapter 15]), in order to be usable, the language would
also need a method of “predication,” a means to express the relationships between
these nouns and verbs.
Following is a possible list of items that appear in all human languages which
might be different extraterrestrially:
144 Sheri Wells-Jensen

Verbs for Seeing and Not Seeing

A trivial but relevant point is that the language would lack the word “blind”; as
there is not seeing, there can be no blindness. And, of course, such polysemous
extensions as “see” meaning “know” and “blind to” meaning “ignorant of” would
be absent. This would be of little consequence; the reader may not have noticed up
until now this chapter’s use of “consult” in place of the usual academic usage of
“see” in parenthetical citations.

Color Terms
These would obviously be absent (at least until the advent of technology that
reveals them). Similarly, words for amounts of light and darkness would be
absent.
This does not, however, preclude words in the semantic domain of “shadow,”
meaning the situation whereby a body interposes itself between a radiation source
and a receiving surface, as infrared radiation and sound waves can be blocked and
redirected in ways similar to visible light. Similarly, this obviously does not pre-
clude words for “day” and “night,” as the absence of the sun is detectible without
vision. Daytime would be warmer, but not brighter.
There are many places where the blind aliens would need words not present in
Earth languages.

Properties of Sound
Because human perception skews toward vision, there are detectible qualities of
sound which lack unique lexical expression in Earth’s languages. For example,
although the difference in timbre between a trumpet and a flute is easily identified,
there is no single word describing this.

Words for Kinds and Qualities of Echoes and Acoustics of Ambient Spaces
If you step into a cathedral, an open space in a library, a living room, or a closet,
little effort is needed to identify that the quality of sound is different, yet there are
no commonly used words in human languages to describe these differences.

Function Words and Grammar

This is where things become interesting. We might imagine a deeper, more com-
plex system of function vocabulary (words like prepositions and pronouns) and an
equally expanded and more complex set of grammar rules to go with them.

Directional Systems
Perhaps the most common kind of system of directions in human languages is rela-
tive: left, right, forward, backward, up, and down with respect to the speaker. This
 Cognition, Sensory Input, and Linguistics 145

is often problematic and usually frustrating, as these are easily confused when try-
ing to describe locations or events from the hearer’s perspective, frequently result-
ing in cries of “No, your other left!”
One alternative in some human languages is to use a common external refer-
ence point, as is done aboard ships, where “bow,” “stern,” “port,” and “starboard”
always mean the front, back, left, and right with respect to the ship itself, rather
than to speakers or hearers aboard the ship.
In some human languages, speakers regularly keep track of, and use, the cardi-
nal directions (north, south, east, and west) to locate objects indoors as well as in
more natural settings (Harrison 2007). Speakers might say, “The cup is on the east
edge of the table” or “Go through the door on the north wall.”
Other human languages use a stable environmental feature as the anchor point.
The Oroha language, spoken in the Solomon Islands, uses “up the beach” and
“down the beach” as other languages use “upstream” and “downstream.”
A blind species might use a cardinal direction, a salient noise source, the slope
of the terrain, or some other set of factors.

Deictic Expressions
Earth languages use “deictic” forms to indicate relative position in space, usually in
relation to the speaker and listener. In languages like English, word pairs such as “this/
that,” “these/those,” and “here/there” lexicalize a two-way distinction, whereas other
languages like Japanese have three relative positions: “kono” (this one near me), “sono”
(that one near you) and “ano” (that one far from both of us). Slightly more complex sys-
tems of deixis, some of which involve visibility and invisibility, also exist (Dixon 1972).
Certainly, we could imagine a more granular system including single words or
affixes indicating “at a position roughly equidistant between us” or “at a position
toward a third conversational participant.”
In combination with direction words, this could allow one speaker to efficiently
tell another everything she knows about where the object under discussion is.
One might even imagine the construction of a rough XYZ grid, with the origin
of the axis located at a fixed point, allowing speaker and listener to locate them-
selves and their interlocutors, including the extent to which an object or person of
interest is above or below a fixed point.
This would employ concentration and memorization skills humans do not regu-
larly tap, but a single deictic word (or even a suffix) could allow a speaker to describe
an object’s location efficiently and unambiguously. Subsequently that point in space
(or an object occupying that point in space) could be referred to using a pronoun.
Some of this information would obviously be optional but having the potential
to communicate extra locations would be extremely useful and probably represent
a survival advantage. Here is an example.

1 We were harvesting apples and we discovered a chickadee nest with a bird in it,
and while we were there, the cat ran past.
146 Sheri Wells-Jensen

If we define the following set of location words . . .

XYZ-1 = the position of the speaker in the yard, relative either to the addressee
or to some other fixed point.
XYZ-2 = the location of the tree (which locates the person with respect to the
tree, closer or further from the addressee, or indicates on what side of the tree
the person is standing).
XYZ-3 = the location of the nest in the tree (how high or low in the branches,
and how close to the trunk).
XYZ-4 = an idea of the vector (speed and direction) of the cat.

. . . then our initial example sentence could be elaborated as follows.

2 We (XYZ-1) were harvesting apples (XYZ-2) and we discovered a chickadee

nest (XYZ-3) with a bird in it, and while we were there, the cat ran past (XYZ-4).

This would also allow the location of the nest (XYZ-3) to be used later as a
deictic pronoun indicating the nest in a subsequent narrative and allow XYZ-1 to
refer to the group of people who found the nest.

Evidentiality
Some human languages provide a set of affixes indicating the reliability of an utter-
ance. These range from direct experience (“I-know-because-I-myself-was-there”)
to doubt (“I-think-that-I-once-heard” or “some-people-say”). This would be valu-
able as a means of establishing trust between speakers or for conveying the speak-
er’s confidence that the information is correct.
We might then have something like the following.

3 We (XYZ-1) were harvesting apples (XYZ-2, EV-1) and we discovered a chick-

adee nest (XYZ-3, EV-2) with a bird (EV-1) in it, and while we were there, the
cat (EV-2) ran past (XYZ-4, EV-3).

Where:
EV-1 = Definite: This is my tree. I know where I was and can say for sure. I also
know it was a bird.
EV-2 = Reasonably sure: I am no expert, but I do know the noise chickadees make
and I know there are some around here, so it is a good bet that it is a chicka-
dee nest. Also, I know my cat was around at the time, so it is a good bet that
it was the cat I heard.
EV-3 = Uncertain: I was a little distracted, so I think the cat went that way at that
speed, but I did not touch him and there was some other noise around at the
time so I cannot be sure.
 Cognition, Sensory Input, and Linguistics 147

Personal Pronouns
Social status, animacy, and gender are often marked on human-language pronouns.
There are a variety of ways of usefully expanding the information content of such
a pronoun system.
A system of exclusive and inclusive marking (“we-including-the-listener” and
“we-excluding-the-listener”), as employed in some human languages, would take
the place of eye contact and hand gestures. Together, these two factors could sim-
plify information exchange when two groups meet, and could facilitate communi-
cation within a single group.
Furthermore, the number marking on pronouns could be more precise than
in Earth languages, referring (when possible) to the actual count of individuals
present.
For example,

4 We (XYZ-1, PRO-4-INCL) were harvesting apples (XYZ-2, EV-1) and we

(PRO-2-EXCL) discovered a chickadee nest (XYZ-3, EV-2) with a bird (EV-1)
in it, and while we (PRO-3-EXCL) were there, the cat (EV-2) ran past (XYZ-4,
EV-3).

Where:
PRO-4-INCL = All four of us, including the addressee.
PRO-2-EXCL = Two of us, not including the addressee.
PRO-3-EXCL = Three of us, not including the addressee.

This would mean that four people (including the addressee) were harvesting
apples, when two of them (not including the addressee) found a nest and three of
them (again not including the addressee) were around to hear the cat.
Pronouns might also include directional or deictic information. Thus, we might
find distinct pronouns meaning “you-north-of-me,” “the-three-of-you-ahead-of-
me,” or “you-of-mixed-gender-and-lower-status-currently-uphill-from-me.”
Gender and social status might be included on animate pronouns and other
qualities marked on inanimate ones. These could take the form of affixes denot-
ing material, size, shape, texture or whether the speaker is currently holding the
object.
Thus, we might have a single pronoun meaning “The-large-wooden-object-
near-you-in-the-downhill-direction.”

Adverbials
First among the set of available adverbials might be a set of words for body posi-
tions and common movements. Human dancers, martial artists, and yogis employ
specialized terminology precisely describing poses or movements of the arms, legs,
148 Sheri Wells-Jensen

hands, feet, and head. Without the remote access to other people’s bodies through
vision, it would be useful for such terms to exist in the everyday lexicon of the
blind aliens. Think of the number of times an instructional video says, “Hold the
needle like this, and wrap the yarn around it this way.”
One might also imagine a more nuanced set of what we might think of as tra-
ditional Earth adverbials, perhaps with finer gradation as shown in what follows.

5 We (XYZ-1, PRO-4-INCL) were harvesting BODY-SHAPE-ARMS-UP

EFFORT-3 apples (XYZ-2, EV-1) and we (PRO-2-EXCL) discovered a chicka-
dee nest (XYZ-3, EV-2) with a bird (EV-1) in it, and while we (PRO-3-EXCL)
were there EFFORT-2, the cat (EV-2) ran past (XYZ-4, EV-3) EFFORT-1.

Where
BODY-SHAPE-ARMS-UP = One of a set of words denoting specific body positions.
EFFORT-3 = With some degree of serious effort.
EFFORT-2 = Less effort but still working.
EFFORT-1 = Minimal effort: the cat is perceived to be running easily. One could
imagine a body shape for the cat as he runs, but if the speaker has no direct
knowledge of the cat’s movement, this could be eliminated or added along
with one of the evidentials indicating doubt.

Prepositions
One function of prepositions is to locate one object with respect to another. Thus,
one object might be in, on, above, under, or beside another. Any of the following
could be expanded to give additional useful information about one object’s rela-
tionship to another.

“in” 1: Nondescript – equivalent to English “in” when the speaker does not know
or does not wish to say something more specific.
“in” 2: Partly located within another body but with access to the outside, e.g., a bird
in an open nest with her head sticking out; one of the default prepositions for
describing something resting in your palm.
“in” 3: Fully located within another body but with access to the outside, e.g., a bird
in an open nest with her head down or an object resting in cupped hands.
“in” 4: Wholly within another object, e.g., a bird buried in a pile of branches or
inside a nest with a woven top, or something held in a closed fist.
“in” 5: Equivalent to “in” 4, except that the nest, hand, or other container is imper-
meable to sound. One could not hear the bird even if it chirped.
“in” 6: Equivalent to “in” 4, except the container is open in some places, like a bird
in a cage.
“in” 7: Equivalent to “in” 4, except that the container has openings wide enough for
fingers to penetrate to touch whatever is inside.
 Cognition, Sensory Input, and Linguistics 149

6 We (XYZ-1, PRO-4-INCL) were harvesting BODY-SHAPE-ARMS-UP

EFFORT-3 apples (XYZ-2, EV-1) and we (PRO-2-EXCL) discovered a chicka-
dee nest (XYZ-3, EV-2) with a bird (EV-1) in-3 it, and while we (PRO-3-EXCL)
were there EFFORT-2, the cat (EV-2) ran past (XYZ-4, EV-3) EFFORT-1.

Multi-Channel Speech
In cases when all this (and presumably more) additional information were used, the
length of utterances could become unworkable.
It might make sense, then, for this species to adopt a parallel system of com-
munication to carry some of this extra information. Noises made with the hands
or feet, voice quality, auxiliary noises made with the voice, or even an artificially
constructed paralinguistic object like a clicker might be essential. (Consult Ortner,
this volume [Chapter 5], for human analogs.)

Choosing What to Say

How language is used in context is as important as what is said. In this case, the
amount and kind of information added to the basic sentence would depend on a
variety of factors including formality, how familiar the addressee is with the envi-
ronment of the story, and the particular point the speaker wants to make.
In addition, the amount and kind of supportive “back-channeling” might be
quite different than in human languages. Whereas it is common for some human
speakers to offer occasional verbal affirmation (such as “mmhm” or “yeah”) while
another is speaking, this has limited utility beyond establishing a social bond and
encouraging the speaker to continue.
However, this back channel could also be used by the blind aliens to keep the
group updated on what is happening. For example, since body language is not
directly accessible, a listener might want to express honest indications of their
level of attention or to offer brief narrations of the things they are doing while
listening.
What is included in these narrations would comprise an intricate network of
culturally governed responses. It might be very difficult for outsiders to follow
these parallel tracks of information and even harder to calculate what they should
say when they are listening.
Furthermore, a crucial cultural norm among the blind aliens might be that accu-
rate information is more important that saving face or pleasing the addressee. This
calculus of courtesy, involving as it does an unfamiliar set of cultural imperatives,
might be very difficult for humans to learn.

Conclusion
Everything in this chapter could be wrong.
150 Sheri Wells-Jensen

Blind scholar Georgina Kleege puts it this way:

I know what it means to be sighted because I live in a sighted world. The lan-
guage I speak, the literature I read, the art I value, the history I learned in school,
the architecture I inhabit, the appliances and conveyances I employ, were all
created by and for sighted people.
(Kleege 1999)

She does not know, and I do not know, as blind individuals in a sighted world,
what the world of a completely blind species would be like – none of us can know
what any alien civilization would be like, even if – as has been attempted here – we
try to proceed in a logical way to keep as many variables constant as possible.
But the point is not to “guess right.” The point is to make the guesses. Our job
before contact is to refine the art of asking and answering all manner of questions.
Replace the word “blind” with anything you like: water-dwelling, nine-legged,
short-lived, long-lived, agoraphobic, claustrophobic, insectoid, ovoid, three-
tongued, highly combustible, etc, and play out the scenario, trying to work out
what impact this would have on language.
That is how we will uncover some of our own unconscious assumptions about
life, intelligence, and language – and that is what will move us further along the
path of preparing for contact with intelligent beings that are truly different from us.

References
Bedny, Marina. 2017. “Evidence from blindness for a cognitively pluripotent cortex.”
Trends in Cognitive Sciences 21.
Burgstahler, Sheryl, and Tanis Doe. 2004. “Disability-related simulations: If, when and how
to use them in professional development.” Review of Disability Studies: An International
Journal 1: Online.
Burton, H., et al. 2002. “Adaptive changes in early and late blind: A fMRI study of Braille
reading.” Journal of Neurophysiology 87: 589–607.
Cambias, James. 2014. The Darkling Sea. NewYork: Tor Books.
Chalmers, Kerry, and B. Humphreys. 2017. Thinking about Human Memory. Cambridge:
Cambridge University Press.
Chambers, Becky. 2014. A Long Way to a Small and Angry Planet. Great Britain: Harper
Voyager.
Chomsky, Noam. 1986. Language: Its Nature, Origin and Use. New York: Praeger.
Chomsky, Noam. 2005. “Three factors in language design.” Linguistic Inquiry 36 (1): 1–22.
Davis, Leonnard. 2017. Beginning with Disability: A Primer. Abingdon, UK: Routledge.
Dixon, R.M.W. 1972. The Dyirbal Language of North Queensland. Cambridge: Cambridge
University Press.
Evans, Vyv. 2014. The Language Myth: Why Language Is Not an Instinct. Cambridge: Cam-
bridge University Press.
Everett, Daniel. 2016. Dark Matter of the Mind: The Culturally Articulated Unconscious.
Chicago, IL: University of Chicago Press.
 Cognition, Sensory Input, and Linguistics 151

Everett, Daniel. 2017. How Language Began: The Story of Humanity’s Greatest Invention.
New York: Liveright.
Fletcher and MacWhinney. 2017. The Handbook of Child Language. Blackwell.
French, Sally. 1992. “Simulation exercises in disability awareness Training: A critique.”
Disability, Handicap & Society 7 (3): 257–266.
Haden-Elgin, Suzette. 1984. Native Tongue. New York: DAW Books Inc.
Harrison, K. David. 2007. When Languages Die: The Extinction of the World’s Languages
and the Erosion of Human Knowledge. Oxford: Oxford University Press.
Hauser, Marc, Noam Chomsky, and W. Tecumseh Fitch. 2014. “The mystery of language
evolution.” Frontiers in Psychology 5: 1–12.
Hull, John. 1990. Touching the Rock: An Experience of Blindness. London: SPCK Publishing.
Kish, Daniel. 1982. Evaluation of an Echo Mobility Training Program for Young Blind Peo-
ple. MA Thesis. University of Southern California, Los Angeles.
Kleege, Georgina. 1999. Sight Unseen. New Haven, CT: Yale University Press.
Kosslyn, S.M., W.L. Thompson, and G. Ganis. 2006. The Case for Mental Imagery. Oxford:
Oxford University Press.
Miller, Susanna. 1997. Reading by Touch. Abingdon, UK: Routledge.
Minsky, Marvin. 1985. “Why intelligent aliens will be intelligible.” In Extraterrestrials, sci-
ence, and alien intelligence, 117–128. Cambridge: Cambridge University Press.
O’Grady, W. 2012. “Three factors in the design and acquisition of language.” Wiley Interdis-
ciplinary Reviews: Cognitive science 3: 493–499.
Pinker, Steven. 1994. The Language Instinct: How the Mind Creates Language. New York:
HarperCollins.
Russell, Mary Doria. 1996. The Sparrow. Ballantine Books.
Saariluoma, Pertti. 1991. “Aspects of skilled imagery in blindfold chess.” Acta Psycho-
logica 77 (1): 65–89.
Saariluoma, Pertti, and V. Kalakoski. 1998. “Apperception and imagery in blindfold chess.”
Memory 61: 67–90.
Schmidt, Susanna, Carla Tinti, Micaela Fantino, Irene Mammarella, and Cesare Cornoldi.
2012. “Spatial representations in blind people: The role of strategies and mobility skills.”
Acta Psychologica 142: 43–50.
Silverman, Arielle. 2015. “The perils of playing blind: Problems with blindness simula-
tion and a better way to teach about blindness.” Journal of Blindness Innovation and
Research 5.
Silverman, Arielle, Jason Gwinn, and Leaf Van Bowen. 2014. “Stumbling in their shoes:
Disability simulations reduce judged capabilities of disabled people.” Social, Psycho-
logical, and Personality Science 6 (4): 464–471.
Sima, J.F., H. Schultheis, and T. Barkowsky. 2013. “Differences between spatial and visual
mental representations.” Frontiers in Psychology 4: (240).
Wells-Jensen, Sheri. 2016. “Alternative perceptual systems and the discovery of basic astro-
nomical phenomena.” International Space Development Conference, San Juan, Puerto
Rico, National Space Society.
Wells-Jensen, Sheri, Emily Michaelson, and Mona Mincara. 2021. “How blind professors
win the first day.” In Picture a Professor: Intersectional Teaching Strategies for Inter-
rupting Bias about Faculty and Increasing Student Learning, edited by Jessamyn Nau-
haus. Morgantown, WV: West Virgina University Press.
14
THE DESIGN FEATURES OF
EXTRATERRESTRIAL LANGUAGE
A Domain-General Approach

Darcy Sperlich

Introduction
Delimiting what the language of an extraterrestrial intelligence (ETI) may consist of
seems to be an impossible task – we have no empirical evidence for ETI even existing.
Consequently, we must engage in blue-sky thinking about their biology, cognition,
and culture. There are numerous factors to consider for a potential ETI language – for
example, does the ETI have a speech organ? How does the ETI’s memory system
work? Do ETIs live in a society? In order to focus our thinking, we need to demys-
tify what a non-human language could be like by making some sound assumptions
based on the best example we have – human language. One way to delimit the current
topic is to claim that language itself is a result of domain-general brain phenomena,
as advocated by the Emergentist approach (e.g., Christiansen and Chater 2016a), and
not a result of an innate narrow language faculty consisting of Merge (e.g., Berwick
and Chomsky 2016, see also Roberts et al., this volume [Chapter 15], and Samuels
and Punske, this volume [Chapter 16], for proponents of this viewpoint).1 The Emer-
gentist approach provides us a context in which to understand ETI’s language, as the
assumption that language arose from a language-ready brain helps us consider what
a probable versus improbable ETI language may look like. Here, our discussion is
mainly focused on natural ETI language, in the sense that it is a result of evolution, but
we cannot discount a biological ETI language that has been engineered. The unknown
here is that we have little idea of what direction a non-natural transformation will take.
The next section reviews what the factors are behind language and its features,
followed by sections which respectively: sketches human language to prepare for
our look at ETI language; focuses on the two areas of ETI language, natural lan-
guage versus engineered; reviews what actual ETI language data we might acquire
in our lifetime; and concludes the chapter.

DOI: 10.4324/9781003352174-14
 The Design Features of Extraterrestrial Language 153

What Is Language?
Gallons of ink have been spilled on this topic, and a single section cannot hope
to do it justice. Nevertheless, we know that language is uniquely human, so any
description of language must account for what we find in humans. We also know
that non-human terrestrial animals lack language, although they do have commu-
nication systems (e.g., see Slobodchikoff on prairie dogs, this volume [Chapter 9]).
One language identification method is Hockett’s design features (Hockett 1960;
Hockett and Altmann 1968), which proposes 16 features that define human lan-
guage.2 Hockett asserted that languages must have all of these features in order for
language to differ from animal communication systems, which are missing some of
these features. Animal communication has certain clusters of features, but, as dis-
cussed by Reboul (2017), the “core” fundamental features of language seem to be
Hockett’s semanticity (sentences have a variety of meanings), discreteness (units
are made up of smaller units), and decoupling (referring to situations not in the here
and now).3 Moreover, hardly any of the other features seem unique to human lan-
guage (Coleman 2006). Thus, we might conclude that any language must express
these core factors, including that of an ETI. Consequently, at a minimum, ETI will
share these core features of human language, and probably a lot more, as we find
with animal communication systems.4
What is language for? While it is used for both communication and internal
thought (cf. Goddard and Wierzbicka 2014), it is uncertain whether language for
communication developed from language of thought or vice versa (cf. Reboul
2017). Another important question is whether language has its origins in Merge
(suggesting a language of thought came first), or developed as a result of repurpos-
ing domain-general processes? The approach adopted here is that language devel-
oped for communication, as the derivatives of the “language is innate” hypothesis
is improbable. Reading Hauser et al. (2002) and the ensuing debates (cf. Christian-
sen and Chater 2015 for recent debate) over the years about the narrow language
faculty involving recursion via a genetic mutation (the key being the feature of
Merge), presents a species-specific scenario that cannot seriously be extended to
ETI, who, we can say with a high degree of confidence, would not come from an
Earth-like environment or evolutionary context.5 At a minimum, ETI would need
to have the same/similar mutation occur (in a brain that is ready for language) to
allow for language to emerge. Following Hornstein and Boeckx’s (2009) discus-
sion on how the faculty of language came about via a “couple of adjustments” to
our cognition and computation, ascribing the same idea to ETI’s emergence of
language (as contrasted against natural selection), one did not leave Earth.6 Unless
one posits a human-like scenario for the entire universe, such a perspective over-
simplifies the problem of language evolution by providing a general “cosmos the-
ory of language.”7 It could be possible to try and obtain Merge in a different way;
however, the theorized extent to which is it responsible for language remains a
moot point.
154 Darcy Sperlich

The approach here is in line with Evans and Levinson (2009) and Christiansen
and Chater (2016b), who maintain that language is a product of culture and cog-
nition, which give rise to the diversity of human languages.8 Humans are highly
adaptive, and language reflects this (see Bybee 2009 discussing the various differ-
ent factors that shape language). It is simpler to postulate that language is a result
of a domain-general mechanism that is adapted for language, and this helps us to
focus our inquiry on what language is. For instance, language universals in ETI
would arise from form and function correspondences and cognitive solutions, as
we see in human languages (Cristofaro 2010; Givón 2018). A mediating view is
that humans try to deliver their message while not overloading their communica-
tion channel vis-à-vis Information Theory (e.g., see Pellegrino et al. 2011 compar-
ing information rates between languages). These are some of the restrictions that
have guided our language development.
This brief sketch shows that any language will have to meet Hockett’s core fea-
ture criteria, and is a result of domain-general processes. I now discuss cognition
and culture as related to the human language system.

Human Language Development

The exact nature of human language is what linguists study, and there are many dif-
ferent theories that try to account for it. The domain-general approach to language
accepts that language has followed a bio-evolutionary approach, in that systems
like vision, audio, attention, and memory have been extended to language devel-
opment. The language system itself is covered by any introductory linguistic text
that provides an overview of the major areas of phonetics, phonology, morphol-
ogy, semantics, pragmatics, and syntax. As importance is placed upon the cognitive
and cultural perspective, a processing perspective is discussed, along with cultural
influences on language development.9
A recent theory constraining language evolution due to processing is Hawkins
(2004, 2014), which details how the processor works as efficiently as possible for
maximum economy, with cross-linguistic patterns as the result. If processing does
affect the shape of human grammar, as Hawkins argues, then we expect the same
for ETI.10 Another thesis by Christiansen and Chater (2016a) posited that there is
a limited window for the uptake of information before it is lost, which is why the
system must work around it.11 Moravcsik (2010: 73) has observed that there is
phonological and semantic reduction in frequently occurring phrases, which eases
production. An important pragmatic aspect is that our language system seems to be
adapted to overcome communicative bottlenecks as discussed by Levinson (2000),
who notes that when speaking, our information-exchange rate is painfully slow
(i.e., restricted by our rate of speech, see also Nixon and Tomaschek, this volume
[Chapter 17]). In order to help overcome this problem, pragmatic inferences come
cheaply as the hearer will be able to infer meaning (utterance meaning), as com-
pared to the literal meaning of the words that were said (sentence meaning). Thus,
 The Design Features of Extraterrestrial Language 155

along the lines of Hawkins (2004, 2014), our grammars are influenced by the fac-
tors that influence the design of our processing system.12
Culture can also have an impact (Chafe 2018). As Moravcsik (2010: 88) pointed
out, language change can be influenced by social factors, and not just for reasons of
general economy. For instance, Lupyan and Dale (2010) discussed how the com-
plexity of a language’s morphosyntax is affected by group size, geographic spread
of the language, and how much language contact has occurred. Furthermore, socio-
linguistic research has helped immensely in our understanding of how language is
molded by its users in society (politeness, for example; see Holmes and Wilson
2017), but we are not able to investigate these aspects for ETI unless we first gain
an insight into ETI’s culture and society.
In sum, human language is understood here to be a result of domain-general
processes, guided by evolutionary pressures. Culture is an important aspect for
human language development, but we will not cover this aspect concerning ETI.

ETI Language

Natural ETI Language

If Darwin’s evolutionary theory works in the right direction, then, as Ulmschneider
(2006) suggested, these natural laws of selection would be present on other Earth-
like planets. Thus, as languages have evolved due to such processes here, there is
hope language would have had similar starting points in extraterrestrial spaces.13
Combining what we know about language theories with language instantiation in
an intelligent species (Homo sapiens), we can now place rough limits on an ETI’s
language, based upon their cognitive apparatus. Again, the caveat has to be made
that any cognitive system must be anchored in culture and society. However, since
there are too many possibilities about what an ETI’s culture and society would look
like, we thus are missing a piece of the puzzle.14 What can be said is that if ETI can
send messages to us (or receive them), this is an indication of a societal product
involving cooperation, and therefore, language must be complex enough to help
organize such an endeavor.
Considering cognition in and by itself, we are also limited in our understand-
ing, in the sense of what embedded language systems look like and how they are
processed. Questions such as “Do ETI have emotions?”; “Are they empathic?”;
and “Do they have humor?” do not seriously enter our discussion. All are
answered “yes” for humans, as there is a close link between language and emo-
tions (Lindquist et al. 2015); language is used to express other’s viewpoints (e.g.,
via logophors, cf. Huang 2000), and jokes are told which are not assessed for truth
values (Barbe 1995). Disregarding these aspects for ETI, what would the linguis-
tic system look like?
Assuming ETIs have concepts about the universe they live in, they need to relate
these concepts to some kind of symbols, which language provides. Such symbolic
156 Darcy Sperlich

tokens combine syntactically, accordingly generating phrases and sentences to

express propositions. We do not know how ETIs communicate physically. For
example, it could be through sound, and thus there would be a phonology; it could
be through smell, whereby the linguistics of olfaction would come into play (cf.
Kershenbaum, this volume [Chapter 2], for discussion on different modes). Would
they have a pragmatic inferencing system? That is debatable, because it depends on
the processing power of ETI, which we shall come to later in the chapter. First, let
us look at some proposals for the lexicon, syntax, and semantics of an ETI.
Regarding the lexicon, would they have a word for every conceivable thing and
situation? Possibly not, as one thinks about/discovers many new things in a day –
but they might have a superior vocabulary (depending on memory constraints),15
suggesting that they would not rely on pragmatic enrichment as much as humans
do (e.g., understanding “black” in different contexts; stereotypically “black” in
“black pen” indicates that the ink is black, while “black” in “black cow” refers to
its hair). Assuming that our ETI are logical creatures, one might also assume that
they have boundaries between words well defined, and not have fuzzy semantic
fields where humans are concerned.
Before we discuss the possible syntax, as with humans, ETI could also have
many different languages. Or, if they are further along the evolutionary line, they
might have chosen to speak one language – which presents us with the theoreti-
cal problem that our conclusions may not be ETI-typologically valid. In any case,
there is not much to comment on ETI’s syntax – it depends on what theory of
syntax one prefers,16 but there will be a system of combining words into phrases,
and phrases into sentences. Would there be the canonical subject, verb, object? If
ETI views the world as a cause–effect one, then a probable result is that ETI will
semantically conceptualize actors and undergoer for example, with a consequence
that relations might hold like subject = actor/agent, object = patient/undergoer and
verb = transitive/intransitive.17 Therefore, we might expect these types of relational
categories to exist.18 Linked to the yet-undiscussed processing powers of ETI, it
is likely that the word order of ETI language would be fixed rather than mixed,
allowing for more straightforward logical/computational processing (see Cristo-
faro 2010 for discussion) – as is the trend among human languages.19
We assume that the semantic system, which involves the meaning of words and
how they form complex meanings in phrases and sentences along with proposi-
tions, exists in their linguistic architecture. That is, they must have a system of
meaning in order for the message to be logically construed.20 It might be the case
that the ETI allows only a completely logical interpretation of what they say, in
that the words and phrases mean what they mean at the semantic level and are not
affected by context at all. Wittgenstein (1922) proposed such a language, so while
this might be considered an advantage in terms of processing, it would be inflexible
and complex, and as such does not reflect the fuzzy logic inherent in human lan-
guages (perhaps Wittgenstein would not have renounced his previous work when
considering ETI). Thus, how important is pragmatics for a logical ETI?
 The Design Features of Extraterrestrial Language 157

A system of inferencing is part of human cognition and language. Humans have

inferencing systems to speed up communication (Levinson 2000) – we do not have
the luxury of time to spell out exactly what we mean in a purely logical way. In
the case of ETI however, if the communication corridor allows for relatively quick
exchange of messages (in a “logical” format), then the need for a pragmatic infer-
encing system is reduced. This is not to say that they would not be able to under-
stand an inference, but they would not make much use of it in their language. Thus,
we need to say a few words about their cognitive system.
Given the domain-general approach adopted for language, it is clear that pro-
cessing limitations in humans have formed language in a way that it is processed
efficiently with the resources at hand.21 Think of an analogy between a car (lan-
guage) and a road (the processing system); if the road is a muddy one, then my
1-liter engine front-wheel drive car will have many problems as it was not designed
for such conditions. However, if I drive my off-road V8 3-liter 4×4, then there will
be no problems as the conditions affected the design of the car. The point here is
that language is affected by the environment it develops in. With ETI, we assume
that their processing speed and capacity is at minimum similar to ours – which in
turn will produce a language system not dissimilar to our own. But if we find that
their processing ability is akin to a hyperloop, their language will reflect that design
parameter. If they have much processing power, it allows for a very powerful and
complex system – one that does not need to be economical because of the vast
resources at hand (think of the historical accident of English having both regular
and irregular verbs, and how the system must deal with this uneconomical fact).
In summary, just as we have discussed the building blocks of human language,
we find similar blueprints fashioning an ETI language.

Engineered ETI Language

What does it mean to have an engineered language? We are referring not to lan-
guages such as Esperanto, which is created, but one wherein the system is inherently
subconscious, but in which ETI (or perhaps humans later on) have: a) discovered
exactly how the system functions and b) how it is represented in the brain, and c)
has modified its neural architecture, which has affected the language system in some
way. With our language as a domain-general phenomenon, increasing the cognitive
capacity of the ETI (for example, improving their memory systems) would allow
their processor to take onboard far more information to digest. A second possibility
is that they are able to directly modify language itself – humans can do this, too, but
it takes a long time and mostly occurs at the subconscious level (e.g., grammati-
calization). For instance, ETI could decide to insert numbers after every pronoun
in order to track reference, or they could selectively link words to particular senses
immediately, without using their general cognitive mechanisms – the possibilities
are endless. Another possibility is that ETIs edit their genes (if they have them) so
that language becomes well and truly innate – at least in an engineered sense.22
158 Darcy Sperlich

Another perspective on engineered language relates to that of artificial intel-

ligence (AI), which is powered by the ubiquitous computer languages. Of course,
however, they are programming languages and do not fit in with language as dis-
cussed in this chapter. For instance, an AI machine can be fed (incomplete) rules
of English as we find with applications that communicate with us to find a cheap
flight, or a chatbot on the internet; Searle’s (1980) Chinese room thought experi-
ment comes to mind, and by following a rule book external to us is not considered
language. In essence, AI language is static (human language is ever changing), and
thus shares the features of a dead language. In this sense, any “language” of an AI
as we now know it is not language as discussed here. Thus, can AI actually have
language? If we consider that the functions of language are to communicate and
organize thought, it will fail on both accounts. AI certainly does not have thought,
and therefore does not have anything to express to a communicator about any pos-
sible thoughts.23 Moreover, Searle (2014) argued strongly against computers hav-
ing beliefs and intentions; consequently, any possibility of a computer language
becoming a natural language does not follow. Following this argumentation, only
when AI is conscious could we consider it as having language, as it will now have
an intrinsic motivation to communicate its thoughts. What would the ramifications
be if we were to conclude that the ETI was in fact an AI robot that gained sentience
(with thought and motivation)? In terms of language, not much, because this brings
us in a full circle back to an engineered ETI language – we cannot predict what
the result would be. We cannot even be sure whether the AI robot cum ETI will be
conscious of its programming because consciousness generally does not entail that
one understands the architecture of one’s own system.
In summary, having an engineered language opens up Pandora’s box in that
we are uncertain of the direction it will take, but we can be sure that processing
capacity (as found affecting human language) is a key factor in consideration of
the design.

Language in ETI Signals

On a final empirical note, what is the most probable situation whereby we can assess
first-hand an alleged ETI language? Certainly, with the SETI and METI programs,
first contact will be an ETI message (which implies intentionality; cf. Hobaiter et
al., this volume [Chapter 3]). Will we be able to decode it? With great difficulty
(as Slobodchikoff, this volume [Chapter 9], thinks) – there are dead languages on
Earth (e.g., Linear A), which still defy analysis. Furthermore, dead languages are a
product of human culture, whereas ETI will be something else entirely. To be fair
however, the writers of these languages probably did not consider their language
would become extinct, and did not think to include a dictionary. A glimmer of
hope is that ETI, on the other hand, may have thought about this problem when
sending messages, and constructed the message as a fully self-contained and self-
explanatory whole – teaching us their language within their message (cf. Sperlich
 The Design Features of Extraterrestrial Language 159

forthcoming).24 As pointed out by Granger et al. (this volume [Chapter 8]) how-
ever, one problem of communication is the inferencing aspect: a message is sent
with a specific intent, but it is interpreted in another way. If the alien species does
make use of inferences as humans do in their language, then it is apparent that they
have a pragmatic system in place which amplifies meaning beyond the semantics
expressed by the message. Hence, if the ETI indicates a misunderstanding which is
beyond the semantics of the message, then pragmatics is clearly evident – we can
then use this to deduce the state of their linguistic, cognitive, and physical system,
to a certain extent (cf. Sperlich forthcoming, for discussion on how processing and
physical restrictions have given rise to pragmatic systems).25
In terms of the possible conversations that might ensue, we can speculate that
the speed of the conversation might be considerably enhanced beyond what is
now thought, given advances in quantum entanglement technology (in the sense
of extending the distance over which two photons can be entangled). It might be
possible that when we receive the message, an entangled photon will be attached
and alert the ETI to its receipt, which in turn would generate more messages in our
direction. On another note, Pepperburg (this volume [Chapter 6]) suggests that we
should be open to ETI using repetition of signal to encode meaning, rather than just
an individual unit. In human language at least, repetition is found, for example, to
express marked messages such as “Is he a friend, friend?” Unless the signal is using
repetition to express linguistic morphological reduplication, for example (which
still would be an individual unit), this strategy is both uneconomical linguistically
(a single unit is preferred over several repetitions to represent a unit of meaning)
and technically (as there will certainly be signal loss which negatively affects the
chances of receiving the exact number of repetitions required).

Conclusion
While there are many possibilities, it makes sense to focus on the probable. Here,
sketching likely ETI language(s) by providing limitations on the ETI’s natural lan-
guage, specifically assuming a domain-general Emergentist approach, allows us
to scale the qualities of the language system and its processing. Assuming that
language is constrained by processing, there is a chance that their language has
developed in a similar manner and would be comprehensible to us. However, rais-
ing the processing bar of the ETI only increases our chances of not understanding
anything due to the information surge – it might still be analyzable by humans
but require enormous computational power and time. After all, language is still
a system. This is further complicated by the scenario that ETI has engineered its
language itself, which would have an inconceivable impact. The question of AI
was raised, discussing the fact that AI does not have language, and if AI were able
to gain consciousness, we would be put in a similar situation of understanding an
engineered language. In any case, we can only make an educated guess as to what
an ETI language would look like from a natural evolutionary perspective, because
160 Darcy Sperlich

once engineering is applied, anything is possible. Finally, like any ETI receiving
a message from Earth (and vice versa), they will have great difficulty in decipher-
ing the language of our message, if it does not at the same time teach the linguistic
system it is based upon.26 However, we can be certain that we will be dealing with
the first non-human language Homo sapiens have ever encountered. Thus, it is
imperative that we continue the search for ETI signals whatever the consequences
may be, rather than being isolationists with a Commodore Perry–like Black Fleet
on the horizon.

Notes
1 For instance, Kershenbaum (this volume [Chapter 2]) is supportive of the viewpoint that
language arose from natural selection processes. Kershenbaum also raises the interesting
point of needing an energetic competitive environment to allow for communication to
arise, and has useful discussion on what modes of communication we are not to expect.
In a similar vein, Berea (this volume [Chapter 7]) discusses a wide range of factors in
the development of an alien language tracing the co-evolution of language and culture
(also touched upon in Hobaiter et al., this volume [Chapter 3]). Berea argues that this
coevolution in humans has been constant, and thus is potentially generalizable to other
intelligent species. The program proposed for identifying these patterns is complex, but
undeniably important if we are to understand the many non-linguistic factors that exert
influence on language.
2 See also Ross (this volume [Chapter 12]) for discussion around Hockett’s features and
application to a linguistic Drake Equation.
3 Berwick and Chomsky (2016: 110) discussed the “basic Property of human language” –
the ability to construct a digitally infinite array of hierarchically structured expression
with determinate interpretations at the interfaces with other organic systems. In this
view, animals do not have this.
4 Practically speaking though, as any initial contact with ETI would be via long-distance
transmission of messages, it would be difficult to assess these for language features – we
need to crack the code first. Slobodchikoff (this volume [Chapter 9]) argues that decod-
ing the message will be a herculean challenge.
5 It should be noted that the issue of Pirahã, a language which is claimed not to make use
of recursion, is far from settled, posing a serious challenge to Merge-based theories (see
Futrell et al. 2016 for latest developments).
6 Evolutionary context is important. Arbib (2016) linked the presence of mirror neurons to
human language development. Would they be key for ETI? Possibly not, following the
same line of thought.
7 Reading Roberts et al. (this volume [Chapter 15]) and Samuels and Punske (this volume
[Chapter 16]), this is the argument to account for any language arising across intelligent
species – a question to be settled perhaps in the coming decades.
8 See also Pinker and Bloom (1990) for another perspective, and Christiansen and Chater
(2016) for a recent overview of positions.
9 See also Chafe (2018) with a focus on thought-based linguistics.
10 Why don’t humans inflect every idea? For example, verbal past/non-past is common
enough, but not hostile/non-hostile, perhaps due to processing restrictions (Bickerton
1990: 56).
11 For further discussion on the integration of this theory with the domain-general ap-
proach, the reader is recommended to read Christiansen and Chater (2016) which con-
tains an excellent discussion on the various diachronic and synchronic factors that affect
the shape of language.
 The Design Features of Extraterrestrial Language 161

12 Chomsky (2017) suggested that computational efficiency trumps communicative effi-

ciency when in conflict. Certainly, a language must do with what resources it has available
to it; however, it is not so straightforward that what is the most computationally effective
will be the strategy taken forward. An example of this is found in Sperlich (2020), as it is
shown possible for English native speakers to long-distantly bind the complex reflexive
pronoun himself from an argument position over a finite clausal boundary; this is not a
“computationally effective strategy,” yet it can occur, processed pragmatically.
13 If ETI have been around for millions of years, as Ulmschneider (2006: 209) put it, “we
have to conclude that these extraterrestrial beings must be essentially God-like, with
faculties that border on omniscience and omnipotence” (think about cumulative culture;
Tomasello 1999). If this is the case for our ETI, then there is no Earthly hope of under-
standing anything approaching the singularity.
14 If ETI has surpassed social conventions, we would not expect to see these social conven-
tions affecting language use.
15 We might assume that they have semantic primitives as humans are theorized to have; 14
suggested, with a further few under investigation (Goddard and Wierzbicka 2014: 10).
16 Human syntax does not necessarily have to be hierarchical; there are other proposals on
dependency grammar (Mel’cuk 1989) and processing order (O’Grady 2005; Frank et al.
2012).
17 My thanks to a reviewer for helping clarify my thoughts here.
18 See also van der Auwera and Gast (2010) for discussion on prototype theory and proto-
agent/patient entailments, as compared to having hard linguistic categories like “subject.”
19 The same reviewer questions why an ETI language should have a fixed word order, as
apart from argument structure, languages also take into account information structure
(e.g., topic-comment); thus, variation could present.
20 See also Bach and Chao (2009) for a sketch on semantic theory and universals.
21 If ETI were to be highly automated in their language processing (but perhaps losing flex-
ibility), we might see the ETI language in the syntactic processing mode rather than the
pre-grammatic mode (Givón 2018: 164). As a further note, linguists do not seem to have
an objective definition of what “efficient” or “optimal” is – it could be a theory internal
construct (fewer steps taken), a timing issue (it is faster to process), or perhaps it requires
two neurons to fire as compared to five.
22 Of course, this differs from the argument that Merge is the necessary building block of
language that is already encoded genetically via a mutation; rather, any advanced species
could actually program the specifics of the language in terms of its lexicon, syntax, etc.
While seemingly attractive, there are disadvantages, as well. Nixon and Tomaschek (this
volume [Chapter 17]) emphasize the importance of learned communication over that with
genetically encoded communication, as the latter is far too slow to deal with rapid cultural
changes in an intelligent species – unless they have consciously stopped innovating.
23 See also Roberts et al., this volume (Chapter 15), for their comments on AI.
24 Slobodchikoff (this volume [Chapter 9]) argues that “universals” of mathematics and
geometry are not a good candidate for communication, as they may be perceived differ-
ently by ETI.
25 Harbour (this volume [Chapter 18]) points out that written human messages provide key
information about how we organize our language (e.g., English words are built around
syllables) and gives insight into our cognition. An aspect writing usually fails to encode
is that our communication is multimodal, having specialized processing for this, as well
(Holler and Levinson 2019) – a part of our cognition will be absent from the message.
This could be corrected for in METI by including multimodal elements drawn from a
conversation. See Ortner (this volume [Chapter 5]) for discussion on the multimodal
elements of language.
26 See Sperlich (forthcoming) on the advantages of using natural language messaging over
artificial language.
162 Darcy Sperlich

References
Arbib, Michael A. 2016. “Towards a computational comparative neuroprimatology: Fram-
ing the language ready brain.” Physics of Life Reviews 16: 1–54.
Bach, Emmon, and Wynn Chao. 2009. “On semantic universals and typology.” In Language
Universals, edited by Morten H. Christiansen, Christopher Collins, and Shimon Edel-
man. Oxford: Oxford University Press. Accessed August 1, 2018. doi:10.1093/acprof:
oso/9780195305432.003.0008.
Barbe, Katharina. 1995. Irony in Context. Amsterdam: John Benjamins.
Berwick, Robert C., and Noam Chomsky. 2016. Why Only Us: Language and Evolution.
Cambridge, MA: MIT Press.
Bickerton, Derick. 1990. Language and Species. Chicago, IL: University of Chicago Press.
Bybee, Joan. 2009. “Language universals and usage-based theory.” In Language Uni-
versals, edited by Morten H. Christiansen, Christopher Collins, and Shimon Edel-
man. Oxford: Oxford University Press. Accessed August 1, 2018. doi:10.1093/acprof:
oso/9780195305432.003.0002.
Chafe, Wallace. 2018. Thought-Based Linguistics: How Languages Turn Thoughts into
Sounds. Cambridge: Cambridge University Press.
Chomsky, Noam. 2017. “The galilean challenge.” Inference: International Review
of Science 3 (1). Accessed August 1, 2018. http://inference-review.com/article/
the-galilean-challenge.
Christiansen, Morten H., and Nick Chater. 2015. “The language faculty that wasn’t: A
usage-based account of natural language recursion.” Frontiers in Psychology 6: 1182.
doi:10.3389/fpsyg.2015.01182.
Christiansen, Morten H., and Nick Chater. 2016a. “The now-or-never bottleneck: A funda-
mental constraint on language.” Behavioral and Brain Sciences 39. Accessed August 1,
2018. doi:10.1017/S0140525X1500031X.
Christiansen, Morten H., and Nick Chater. 2016b. Creating Language: Integrating Evolu-
tion, Acquisition, and Processing. Cambridge, MA: MIT Press.
Coleman, John S. 2006. “Design features of language.” In Encyclopedia of Language and
Linguistics, edited by Keith Brown, 471–475. Amsterdam: Elsevier.
Cristofaro, Sonia. 2010. “Language universals and linguistic knowledge.” In The Oxford
Handbook of Linguistic Typology, edited by Jae Jung Song, 227–249. Oxford: Oxford
University Press.
Evans, Nicholas, and Stephen C. Levinson. 2009. “The myth of language universals: Lan-
guage diversity and its importance for cognitive science.” Behavioral and Brain Sciences
32: 429–492.
Frank, Stefan L., Rens Bod, and Morten H. Christiansen. 2012. “How hierarchical is lan-
guage use?” Proceedings of the Royal Society B. Accessed August 1, 2018. doi:10.1098/
rspb.2012.1741.
Futrell, R., L. Stearns, D.L. Everett, S.T. Piantadosi, and E. Gibson. 2016. “A corpus in-
vestigation of syntactic embedding in Pirahã.” PLoS One 11 (3): e0145289. https://doi.
org/10.1371/journal.pone.0145289.
Givón, Thomas. 2018. On Understanding Grammar: Revised Edition. Amsterdam: John
Benjamins.
Goddard, Cliff, and Anna Wierzbicka. 2014. Words and Meanings: Lexical Semantics across
Domains, Languages and Cultures. Oxford: Oxford University Press.
Hauser, Marc D., Noam Chomsky, and W. Tecumseh Fitch. 2002. “The faculty of language:
What is it, who has it, and how did it evolve?” Science 298 (5598): 1569–1579.
 The Design Features of Extraterrestrial Language 163

Hawkins, John. 2004. Efficiency and Complexity in Grammars. Oxford: Oxford University
Press.
Hawkins, John. 2014. Cross-Linguistic Variation and Efficiency. Oxford: Oxford University
Press.
Hockett, Charles F. 1960. “The origin of speech.” Scientific American 203: 88–111.
Hockett, Charles F., and Stuart A. Altmann. 1968. “A note on design features.” In Animal
Communication: Techniques of Study and Results of Research, edited by Thomas Se-
beok, 61–72. Bloomington, IN: Indiana University Press.
Holler, Judith, and Stephen C. Levinson. 2019. “Multimodal language processing in human
communication.” Trends in Cognitive Sciences 23 (8): 639–652.
Holmes, Janet, and Nick Wilson. 2017. An Introduction to Sociolinguistics. London: Pearson.
Hornstein, Norbert, and Cedric Boeckx. 2009. “Approaching universals from below:
I-universals in light of a minimalist program for linguistic theory.” In Language
Universals, edited by Morten H. Christiansen, Christopher Collins, and Shimon Edel-
man. Oxford: Oxford University Press. Accessed August 1, 2018. doi:10.1093/acprof:
oso/9780195305432.003.0002.
Huang, Yan. 2000. Anaphora: A Cross-Linguistic Study. Oxford: Oxford University Press.
Levinson, Stephen. 2000. Presumptive Meanings: The Theory of Generalized Conversa-
tional Implicature. Cambridge, MA: MIT Press.
Lindquist, Kristen A., Jennifer K. MacCormack, and Holly Shablack. 2015. “The role of
language in emotion: Predictions from psychological constructionism.” Frontiers in Psy-
chology 6: 444. Accessed August 1, 2018. doi:10.3389/fpsyg.2015.00444
Lupyan, Gary, and Rick Dale. 2010. “Language structure is partly determined by social
structure.” PLoS One 5 (1): e8559. Accessed August 1, 2018. doi:10.1371/journal.
pone.0008559.
Mel’cuk, Igor A. 1989. Dependency Syntax: Theory and Practice. Albany, NY: State Uni-
versity of New York Press.
Moravcsik, Edith A. 2010. “Explaining language universals.” In The Oxford Handbook of
Linguistic Typology, edited by Jae Jung Song, 69–89. Oxford: Oxford University Press.
O’Grady, William. 2005. Syntactic Carpentry: An Emergentist Approach to Syntax. Mahwah,
NJ: Lawrence Erlbaum.
Pellegrino, François, Christophe Coupé, and Egidio Marsico. 2011. “A cross-language per-
spective on speech information rate.” Language 87 (3): 539–558.
Pinker, Steven, and Paul Bloom. 1990. “Natural language and natural selection.” Behavioral
and Brain Sciences 13: 707–784.
Reboul, Anne. 2017. Cognition and Communication in the Evolution of Language. Oxford:
Oxford University Press.
Searle, John. 1980. “Minds, brains, and programs.” Behavioral and Brain Sciences 3:
417–424.
Searle, John. 2014. “What your computer can’t know.” Review of The 4th Revolution:
How the Infosphere Is Reshaping Human Reality, by Luciano Floridi and Super-
intelligence: Paths, Dangers, Strategies, by Nick Bostrom. The New York Review of
Books, October 9. Accessed August 1, 2018. www.nybooks.com/articles/2014/10/09/
what-your-computer-cant-know/.
Sperlich, Darcy. 2020. Reflexive Pronouns: A Theoretical and Experimental Synthesis.
Cham: Springer. doi:10.1007/978-3-030-63875-7
164 Darcy Sperlich

Sperlich, Darcy. Forthcoming. “Syntactic and pragmatic considerations of natural language

METI.” In Cognition and Communication in Extraterrestrial Intelligence, edited by
Douglas Vakoch. Oxford: Oxford University Press.
Tomasello, Michael. 1999. The Cultural Origins of Human Cognition. Cambridge, MA:
Harvard University Press.
Ulmschneider, Peter. 2006. Intelligent Life in the Universe: Principles and Requirements
Behind its Emergence. Berlin: Springer-Verlag.
Van der Auwera, John, and Volker Gast. 2010. “Language universals and linguistic knowl-
edge.” In The Oxford Handbook of Linguistic Typology, edited by Jae Jung Song, 166–
189. Oxford: Oxford University Press.
Wittgenstein, Ludwig. 1922. Tractatus Logico-Philosophicus. London: Routledge and
Kegan Paul.
15
UNIVERSAL GRAMMAR
Ian G. Roberts, Jeffrey Watumull, and Noam Chomsky

Introduction
Is language universal? In particular, is the grammar – the computational system
in the mind/brain – that powers human language universal? Could it be universal
in the way laws of nature are universal, such that any sufficiently intelligent sys-
tem would all but inevitably converge on it (in either its evolution or its science)?
Here we will expound on the conjecture that the answers to these questions are
affirmative: grammar – particularly human grammar – is not specific to our spe-
cies, but universal in the deepest of senses. The implications for xenolinguistics
are obviously profound: we should predict that any extraterrestrial intelligence
(ETI) – indeed any sufficiently intelligent system (e.g., artificial intelligence) – we
encounter would likely be endowed with a cognitive computational system that
runs human-style linguistic “software”, thus eliminating any principled limit to
effective communication. However, there could exist be material differences in the
“hardware” used to physically externalize linguistic information, but these would
pose mere engineering challenges rather than insoluble conceptual problems. It is
not unreasonable to suppose that the combined intelligence of humans and ETIs
could construct the necessary interface(s). In that case, effective communication
between these “universal minds” would be guaranteed.
The human mind, Descartes argued, is undoubtedly in some sense a “univer-
sal instrument” (Descartes 1637 [1995]). We cannot know with certainty what he
intended by this provocative comment, but we do know that the Cartesians would
have understood language as fundamental to any nontrivial notion of “universal-
ity” and “intelligence” because it is language that empowers humans to generate
an unbounded set of hierarchically structured expressions that can enter into (or
in fact constitute) effectively infinitely many thoughts and actions – that is, the

DOI: 10.4324/9781003352174-15
166 Ian G. Roberts, Jeffrey Watumull, and Noam Chomsky

competence of every human, but no beast or machine, to use language in crea-

tive ways appropriate to situations but not caused by them, and to formulate and
express these thoughts coherently and without bound, perhaps “incited or inclined”
to speak in particular ways by internal and external circumstances but not “com-
pelled” to do so. This linguistic competence (and especially its creative use), in
concert with other mental faculties, establishes the general intelligence necessary
for the evolutionary “great leap forward” of our species (see Chomsky 2016):

there might have been a crucial mutation in human evolution which led, in
almost no time from an evolutionary perspective, from [humans living in] caves
to [their creating knowledge of such sophistication as to enable us to imagine
and construct things as complex as, say,] spaceships. It’s a plausible speculation
that the mutation in question was whatever it is that makes our brains capable
of computing recursive syntax, since it’s the recursive syntax that really gives
language – and thought – their unlimited expressive power. It’s one small step
from syntax to spaceships, but a great leap for humans.
(Roberts 2017: 182)

A great leap for humans – and, on Earth, only humans, evidently (see Berwick and
Chomsky 2016).
Moreover, it has been assumed that the essential properties of human language
are not only unique, but logically contingent:

Let us define “universal grammar” (UG) as the system of principles, conditions,

and rules that are elements or properties of all human languages not merely by
accident but by necessity – of course, I mean biological, not logical necessity.
Thus UG can be taken as expressing “the essence of human language.”
(Chomsky 1975: 29)

There is no a priori reason to expect that human language will have such proper-
ties; Martian could be different.
(Chomsky 2000: 16)

This assumption, we submit, merits rethinking in light of progress in the Mini-

malist Program (Chomsky 1995). Recent work demonstrating the simplicity (Watu-
mull et al. 2017) and optimality (Chomsky et al. 2017) of language increases the
cogency of the following: “the basic principles of language are formulated in terms
of notions drawn from the domain of (virtual) conceptual necessity”, the domain
defined by “general considerations of conceptual naturalness that have some inde-
pendent plausibility, namely, simplicity, economy, symmetry, nonredundancy, and
the like” (Chomsky 1995: 171, 1) that render linguistic computation optimal. To the
extent that this strong minimalist thesis (SMT) is true, the essential – computational –
 Universal Grammar 167

properties of language would derive from laws of nature – language- and even
biology-independent principles that, once realized in the mind/brain, do entail par-
ticular properties as logically necessary. For instance, it is simply a fact of logic
that the simplest (optimal) form of the recursive procedure generative of syntactic
structures, Merge, has two and only two forms of application (i.e., “external” and
“internal” in the sense of combining two separate objects or two where one is
inside the other). Relatedly, given the nature of the structures Merge generates,
minimal structural distance is necessarily the simplest computation for the struc-
ture dependence of rules. And so on and so forth (see Berwick et al. 2011; Chom-
sky 2013; Watumull 2015 for additional examples).
Research in the Minimalist Program starts with the optimality conjecture and
proceeds to inquire whether and to what extent it can be sustained given the
observed complexities and variety of natural languages. If a gap is discovered, the
task is to inquire whether the data can be reinterpreted, or whether principles of
simplicity and optimal computation can be reformulated, so as to solve the puzzles
within the framework of SMT, thus generating some support, in an interesting and
unexpected domain, for Galileo’s precept that nature is simple and it is the task of
the scientist to prove it.
As we discover more and more of “the essence of human language” to be
defined by (virtual) conceptual necessity, the less and less absurd it is to question
just how contingent a phenomenon human language really is. It may well be with
language as with other phenomena studied in the natural sciences that “[b]ehind it
all is surely an idea so simple, so beautiful, that when we grasp it – in a decade, a
century, or a millennium – we will all say to each other, how could it have been
otherwise?” (Wheeler 1986: 304). In other words, there may well be some a priori
reasons to expect human language to have the properties it does; so the ETI’s lan-
guage might not be so different from human language, after all.

Simplicity, Universality, and Merge

Our conjecture is based on the notion of simplicity as originally conceived in gen-
erative linguistics. “[S]implicity, economy, compactness, etc.” were proffered in
the first work on generative grammar as criteria the grammar of a language must
satisfy:

Such considerations are in general not trivial or “merely esthetic”. It has been
recognized of philosophical systems, and it is, I think, no less true of grammati-
cal systems, that the motives behind the demand for economy are in many ways
the same as those behind the demand that there be a system at all.
(Chomsky 1951: 1, 67)

The idea is elementary but profound: if the theory is no more simple, economical,
compact, etc., than the data it is proffered to explain, it is not a theory at all; hence, the
168 Ian G. Roberts, Jeffrey Watumull, and Noam Chomsky

more compressed the theory, the more successful – i.e., the more explanatory – it is.
Incidentally, this idea is appreciated surprisingly seldom today: many computational
cognitive scientists and machine learning theorists (and hence virtually all “artificial
intelligence” [AI] labs in academia and industry) have perversely redefined a suc-
cessful theory or computer program to be one that merely approximates or classifies
unanalyzed data.1 This contrasts dramatically with the Enlightenment definition in
which data are selectively analyzed as evidence for/against conjectured explanations.
In generative grammar since approximately 1980 (principles-and-parameters
[P&P] theory), language acquisition has been explained as the process of setting
the values for the finitely many universal parameters of the initial state of the lan-
guage faculty (universal grammar, UG). The apparent complexity and diversity of
linguistic phenomena are illusory and epiphenomenal, emerging from the interac-
tion of invariant principles under varying conditions. This was a radical shift from
the early work in generative linguistics, which sought only an evaluation measure
that would select among alternative grammars – the simplest congruent with the
format encoded in UG and consistent with the primary linguistic data. But with the
P&P shift in perspective, simplicity can be reformulated. As discussed in the earli-
est work in generative linguistics, notions of simplicity assume two distinct forms:
the imprecise but profound notion of simplicity that enters into rational inquiry
generally, and the theory-internal measure of simplicity that selects among gram-
mars (i-languages). The former notion of simplicity is language-independent, but
the theory-internal notion is a component of UG, a subcomponent of the procedure
for determining the relation between experience and i-language. In early work,
the internal notion was implemented in the form of the evaluation procedure to
select among proposed grammars/i-languages consistent with the UG format for
rule systems, but the P&P approach transcends that limited, parochial conception
of simplicity: with no evaluation procedure, there is no internal notion of simplicity
in the earlier sense. There remains only the universal notion.
In P&P, grammars – i-languages – are simple, but they are so by virtue of objec-
tive principles of computational efficiency (Chomsky 2005), not by stipulation in
UG. In fact, rather than “simple”, we propose to define P&P-style acquisition as
“economical”, which, in the Leibnizian spirit, we understand to subsume simplicity:

The most economical idea, like the most economical engine, is the one that
accomplishes most by using least. Simplicity – or fuel consumption – is a dif-
ferent factor from power [i.e., generative capacity, empirical coverage, etc.] but
has to be taken equally into consideration [. . .]. The economy of a basis may be
said to be the ratio of its strength to its simplicity. But superfluous power is also
a waste. Adequacy for a given system is the only relevant factor in the power
of a basis; and where we are comparing several alternative bases for some one
system, as is normally the case, that factor is a constant. Thus in practice the
simplest basis is the most economical.
(Goodman 1943: 111)
 Universal Grammar 169

Economy, in other words, is a minimax property. In Leibniz’s words (see Roberts

and Watumull 2015): “the simplicity of the means counterbalances the richness of
the effects” so that in nature “the maximum effect [is] produced by the simplest
means”. This is the Galilean ideal (see Chomsky 2002).
The maximally economical form of P&P-style learning explicable in terms of
objective factors is the traversal of a parameter hierarchy for parameter specifica-
tion (see Roberts 2012, 2019). In such a system, the child is not enumerating and
evaluating grammars.2 Instead, the i-language matures to a steady state in a rela-
tively deterministic process of “answering questions” that emerge naturally and
necessarily in the sense that there exist “choices” in acquisition that logically must
be “made” for the system to function at all; none of the parameters need be encoded
in the genetic endowment (see Obata et al. 2015 for similar ideas). This is the ideal,
of course. Like SMT generally, how closely it can be approximated is an empirical
matter, and there remain many challenges.
Parameter specification – i.e., the P&P conception of “learning” as the specifi-
cation of values for the variables in i-language – can be schematized as a decision
tree (parameter hierarchy) which is governed by minimax economy: minimizing
formal features (feature-economy) coupled with maximizing accessible features
(input-generalization). Traversal of a hierarchy – a conditional-branching Turing
machine program – is inevitably economical in that the shortest (in binary) and
most general parameter settings are necessarily “preferred” in the sense that the
sooner the computation halts, the shorter the parameter settings. For instance, to
specify word order, a series of binary queries with answers of increasing length and
decreasing generality (microparameters) is structured thus:

Is head-final present?

NO YES

head-initial Present on all heads?

YES NO

head-final Present on [+V] heads?

YES NO

head-final in clauses only Present on ...?

FIGURE 15.1 Parameter Hierachy

170 Ian G. Roberts, Jeffrey Watumull, and Noam Chomsky

For compatibility with computability theory and Boolean logic, the parameter hier-
archy can be translated as follows:

Hierarchy: H
State T: Decision Problem
Yes: 0/1 (0 = transition to state T + 1) (1 = halt and output parameter specification
for H)
No: 0/1 (0 = transition to state T + 1) (1 = halt and output parameter specification
for H)

Hierarchy: Word Order

State 1: Is head-final present?

Yes: Output 0 (transition to State 2)
No: Output 1 (halt and output “head-initial”)

State 2: Present on all heads?

Yes: Output 1 (halt and output “head-final”)
No: Output 0 (transition to State 3)

State 3: Present on [+V] heads?

Yes: Output 1 (halt and output “head-final in clause only”)
No: Output 0 (transition to State 4)

So in P&P, the logic is not “enumerate and evaluate” with stipulative (theory-
internal) simplicity measures; it is “compute all and only what is necessary”, which
implies the language-independent reality of economy in that, as with the parameter
hierarchies, the process answers all and only the questions it needs to. It is not that
there is any explicit instruction in the genetic endowment to prefer simple answers:
it is simply otiose and meaningless to answer unasked questions (i.e., once the
parameters are set, the computation halts). ETI grammars, too, might therefore
grow along the lines of the P&P method.
Moreover, the “answers” to “questions” can be represented in binary. Indeed,
binary is a notation-independent notion necessary and sufficient to maximize com-
putation with minimal complexity (hence, it is optimal for terrestrial and extrater-
restrial computation): functions of arbitrarily many arguments can be realized by
the composition of binary (but not unary) functions – a truth of minimax logic with
“far-reaching significance for our understanding of the functional architecture of
the brain” (Gallistel and King 2010: x) – for the physical realization of intelligence
anywhere in the universe. The mathematical and computational import of binary
was rendered explicit in the theories of Turing (1936) and Shannon (1948), the
former demonstrating the necessarily digital – hence, ultimately binary – nature
of universal computation (a universal Turing machine being the most general
 Universal Grammar 171

mathematical characterization of computation); the latter formalizing information

in terms of bits (binary digits). The consilience of these ideas is our Economy
Thesis: human language is based on simple representations (i.e., bits) and strong
computations (i.e., the binary functions of Turing machines) – and the “economy of
a basis may be said to be the ratio of its strength to its simplicity” (Goodman 1943:
111, emphasis in original).
As one of the “general considerations of conceptual naturalness that have some
independent plausibility”, economy would be a factor that obtains of any optimally
“designed” (natural or artificial) computational system. In terms of universality,
if the ETI language were optimal in the sense of conforming to virtual conceptual
necessity, then it might be surprisingly similar to human language. In point of
fact, we ought not to be too surprised. It is now well established by biologists that
convergence is a common theme in any evolutionary process: “the number of evo-
lutionary end-points is limited: by no means is everything possible. [Because of
evolutionary convergence,] what is possible usually has been arrived at multiple
times, meaning that the emergence of the various biological properties is effec-
tively inevitable” (Conway Morris 2013: xii–xiii); indeed, the distinguished Cam-
bridge paleontologist Simon Conway Morris argues that human-style intelligence
was effectively inevitable, given the initial conditions of evolution on Earth. And
there is no reason a priori to assume that the principle of evolutionary convergence
is unique to the biology of a particular planet. Quite the contrary, if we accept
the rational form of inquiry in which the principle is understood abstractly in a
computational framework. The idea is that any computational system anywhere
made of anything is governed by laws of computation (see Gallistel and King
2010: 167). Given this universality of the functional, mathematical architecture
of computation, it is possible that we may need to rethink how uniquely human or
even uniquely biological our modes of mental computation really are. One inter-
esting implication is that we must rethink any presumptions that extraterrestrial
intelligence or artificial intelligence would really be all that different from human
intelligence.
So we assume that human language is a computational process that can be char-
acterized by a Turing machine (see Watumull 2015). It is possible to explore the
space of all possible Turing machines (i.e., the space of all possible computer pro-
grams), not exhaustively of course, but with sufficient breadth and depth to make
some profound discoveries. Marvin Minsky and Daniel Bobrow once enumerated
and ran some thousands of the simplest Turing machines (computer programs with
minimal numbers of rules) and discovered, intriguingly, that out of the infinity of
possible behaviors, only a surprisingly – and intriguingly – small subset emerged
(Minsky 1985). These divided into the trivial and the nontrivial. The boring pro-
grams either halted immediately or erased the input data or looped indefinitely or
engaged in some similar silliness. The remainder, however, were singularly inter-
esting: all of these programs executed an effectively identical counting function
– a primitive of elementary arithmetic. In fact, this operation reduces to a form of
172 Ian G. Roberts, Jeffrey Watumull, and Noam Chomsky

FIGURE 15.2 Sampling of the universe of possible Turing machines

Merge (see Chomsky 2008). More generally, these “A-machines” (A for arithme-
tic) prove a point:

[I]t seems inevitable that, somewhere, in a growing mind some A-machines must
come to be. Now, possibly, there are other, really different ways to count. So there
may appear, much, much later, some of what we represent as “B-machines” –
which are processes that act in ways which are similar, but not identical to, how
the A-machines behave. But, our experiment hints that even the very simplest
possible B-machine will be so much more complicated that it is unlikely
that any brain would discover one before it first found many A-machines.
(Minsky 1985: 121)

This is evidence that arithmetic, as represented in an A-machine, is an attractor

in the phase space of possible mathematical structures (Figure 15.2):

any entity who searches through the simplest processes will soon find fragments
which do not merely resemble arithmetic but are arithmetic. It is not a matter of
inventiveness or imagination, only a fact about the geography of the universe
of computation.
(Minsky 1985: 122, emphasis added)

This thesis obviously generalizes beyond arithmetic to all “simple” computations

(see Wolfram 2002 for countless examples). “Because of this, we can expect cer-
tain ‘a priori’ structures to appear, almost always, whenever a computational sys-
tem evolves by selection from a universe of possible processes” (Minsky 1985:
119). Analogously, we submit that it is not implausible that an evolutionary search
through the simplest computations will soon find something like Merge. Merge is
an operation so elementary as to be subsumed somehow in every more complex
computational procedure: take two objects X and Y already constructed and form
 Universal Grammar 173

the object Z without modifying X or Y, or imposing any additional structure on

them: thus Merge (X, Y) = {X, Y}. This simple assumption suffices to derive in
a principled (necessary) way a complex array of otherwise arbitrary (contingent)
phenomena such as the asymmetry of the conceptual-intentional and sensory-motor
interfaces (entailing the locus of surface complexity and variety), the ubiquity of
dislocation, structure dependence, minimal structural distance for anaphoric and
other construals, and the difference between what reaches the mind for semantic
interpretation and what reaches the apparatus of articulation and perception (see
Chomsky 2017).
As implied by our Economy Thesis, simplicity can be defined in algorithmic
information theory (or the theory of program-size complexity): the complexity
of a program is measured by its maximally compressed length in bits so that
the simplest program is that with the shortest description. A search of the phase
space of possible programs, whether conducted consciously (e.g., by us, ETIs,
etc.) or unconsciously (e.g., by modern computers, evolution, etc.), automatically
proceeds in size order from the shortest and increasing to programs no shorter
than their outputs (these incompressible programs are effectively lists); many
complex programs would subsume simpler programs as the real numbers sub-
sume the natural numbers. And, as demonstrated logically and empirically, “any
evolutionary process must first consider relatively simple systems, and thus dis-
cover the same, isolated, islands of efficiency” (Minsky 1985: 122). Thus it may
well be that, given the universal and invariant laws of evolution, convergence on
systems – Turing machines – virtually identical to those “discovered” in our evo-
lutionary history is inevitable.3 Hence our questioning the proposition “Martian
could be different”.
The fact that simple computations are attractors in the phase space of possi-
ble computations goes some way to explaining why language should be optimally
designed (insofar as SMT holds) in that an evolutionary search is likely to converge
on it, which leads us to consideration of the origin of language.4 The evolution of
language is mysterious (see Hauser et al. 2014), but SMT is consistent with the
limited archaeological evidence that does exist on the emergence of language, evi-
dently quite recently and suddenly in the evolutionary time frame (see Tattersall
2012). Furthermore, there is compelling evidence for SMT in the design of lan-
guage itself. For instance, it is a universal truth of natural language that the rules of
syntax/semantics are structure-dependent (see Berwick et al. 2011): hierarchy, not
linearity, is determinative in the application of rules and interpretation of expres-
sions. Thus, the most reasonable speculation today – and one that opens productive
lines of research – is that from some simple rewiring of the brain, Merge emerged,
naturally in its simplest form, providing the basis for unbounded and creative
thought – the “great leap forward” evidenced in the archaeological record and in
the remarkable differences distinguishing modern humans from their predecessors
and the rest of the animal kingdom (see Huybregts 2017; Berwick and Chomsky
2016 for in-depth discussion of these topics).
174 Ian G. Roberts, Jeffrey Watumull, and Noam Chomsky

If this conjecture can be sustained, we could answer the question why lan-
guage should be optimally designed: optimality would be expected under the
postulated conditions, with no selectional or other pressures operating, so the
emerging system should just follow the laws of nature such as minimal com-
putation and more “general considerations of conceptual naturalness that have
some independent plausibility, namely, simplicity, economy, symmetry, nonre-
dundancy, and the like” (Chomsky 1995: 171, 1) – quite like the way a snow-
flake forms. If this is correct, then, contrary to what was once presumed, there
would be a priori reasons to expect that human language will have the properties
it does; the “principles, conditions, and rules that are elements or properties of
all human languages” (Chomsky 1975: 29) would be logically necessary, deriv-
ing from laws of nature. Remarkably, it could be in language, not physics, that
we first discover, in Wheeler’s words, “an idea so simple, so beautiful, that [. . .]
we will all say to each other, how could it have been otherwise?” (Wheeler
1986: 304).
One idea in language so simple that perhaps it could not have been otherwise
is Merge. As we have discussed, it is in some sense an attractor in the phase space
of possible generative functions: its irreducibility and conceptual necessity render
it “inevitable” in the design of any computational system. In this most elementary
of forms, Merge functions as follows. Given a workspace WS of syntactic objects
{Xi, . . ., Xm}, let ∑ be the shortest sequence (X1, . . ., Xn) such that Xi is accessible
and ∑ exhausts WS. Thus, Merge(∑) = {{X1, X2}, X3, . . ., Xm}. By this formula-
tion, which conforms to the simplest, necessary principles of computability theory,
we map WS to WS′ by taking any two accessible elements X and Y in WS, and
replace X and Y with {X, Y} in WS′. It is manifest that two and only two legiti-
mate forms of Merge follow from this formulation. (1), if X1 is a term of X2 (or
conversely), then Merge replaces X1 and X2 by {X1, X2}: Merge(∑) = {{X1, X2},
X3, . . ., Xm}. (2) If neither X1 nor X2 is a term of the other, then, again, Merge(∑) =
{{X1, X2}, X3, . . ., Xm}. Call (1) Internal Merge and call (2) External Merge. The
outputs of (1) and (2) are identical, consistent with the Galilean ideal, “general
considerations of conceptual naturalness that have some independent plausibility,
namely, simplicity, economy, symmetry, nonredundancy, and the like” (Chomsky
1995: 171, 1).
The simple, legitimate form of Merge we proffer has implications for notions
of universality. This “merger as replacement” formulation is formally equivalent
to a 2-tag system of the general form given by Post (1943) (whose formulation of
rewrite rules influenced early work on generative grammar), which is interesting
because tag systems form “a class of systems with a particularly simple underlying
structure” (Wolfram 2002: 93). In a tag system, given a sequence of elements, a set
of rules removes a fixed number of elements from the beginning of the sequence
and replaces them with a fixed number of elements to the end of the sequence.5 For
instance, consider a 1-tag system with the rewrite rules (1, . . .) => (. . . , 1, 0) and
(0, . . .) => (. . . , 0, 1): If the sequence begins with a 1, remove it, and replace it
with 1, 0 at the end of the sequence; If the sequence begins with a 0, remove it, and
 Universal Grammar 175

replace it with 0, 1 at the end of the sequence. Representing 1 with black and 0 with
white, we can express the rewrite rules as follows (Figure 15.3).
And now we can see the complex patterns that emerge from applying such sim-
ple rules to simple initial conditions (Figure 15.4).

FIGURE 15.3 Merge encoded in rules of a 2-state, 3-color universal Turing machine—
provably the “smallest” universal Turing machine

FIGURE 15.4 Merge formalized as a universal 2-tag system

176 Ian G. Roberts, Jeffrey Watumull, and Noam Chomsky

The patterns that emerge with 2-tag systems become increasingly complex
(Figure 15.5).
Merge (Figure 15.6) is the simplest 2-tag system with the following rules (n.b.,
other permutations are possible).
What is most interesting is that this 2-tag system is provably equivalent to a uni-
versal Turing machine (see the proofs in Minsky 1961; Cocke and Minsky 1964;
Davis 1958; Wolfram 2002; Watumull 2015): it can compute anything that is com-
putable. Therefore, it would not be absurd to conject that, by virtue of Merge, the
human mind is a universal Turing machine. However we will not defend the con-
jecture here – we proffer it simply in the spirit of exploration:

It is [. . .] quite possible that we, as a species, have crossed a cognitive threshold.

Our capacity to express anything, through the recursive syntax and composi-
tional semantics of natural language, might have taken us into a cognitive realm
where anything, everything is possible.
(Roberts 2017: 181–182)
 Universal Grammar 177

Let us suppose that the human mind is a universal instrument, a universal

Turing machine. Immediately, we must answer questions of scope and limits. With
respect to scope, a Turing-universal mind could arguably explain and understand-
ing everything, in principle. The argument is simple: a universal Turing machine
can emulate any other Turing machine (i.e., a universal computer can run any
program); a program is a kind of theory (written to be readable/executable by a
computer); thus, a universal Turing machine can compute any theory; and thus,
assuming that everything in the universe could in principle be explained by and
understood within some theory or other (i.e., assuming no magic, miracles, etc.),
a universal Turing machine – a Turing-universal mind – could explain and under-
stand everything. QED, perhaps. It is an intriguing conclusion, and not obviously
false.
However, notwithstanding the universal logic of computation, it is obviously
necessary that there exist constraints on the mind if it is to have any scope at all.
If a Turing-universal mind is to be a universal explainer, it should not generate all
possible explanations – true and false – because that would be merely to restate the
problem of explaining Nature: deciding which in an infinite set of explanations are
the true (or best) explanations is as difficult as constructing the best explanations in
the first place. There must be “a limit on admissible hypotheses”, in the words of
Charles Sanders Peirce (Peirce 1903, see Chomsky 2006: xi). This interdependence
of scope and limits has been expounded by many creative thinkers and analyzed by
(creative) philosophers of aesthetics: the beauty of jazz emerges not by “playing
anything”, but only when the improvisation is structured, canalized; the beauty of a
poem is a function of its having to satisfy the constraints of its form, as the eminent
mathematician Stanislaw Ulam (1976: 180) observed

When I was a boy I felt that the role of rhyme in poetry was to compel one to
find the unobvious because of the necessity of finding a word which rhymes.
This forces novel associations and almost guarantees deviations from routine
chains or trains of thought. It becomes paradoxically a sort of automatic mecha-
nism of originality.

Thus from science to art, we see that the (hypothesized) infinite creativity of the
Turing-universal human mind is non-vacuous and useful and beautiful only if it
operates within constraints – constraints discoverable by any (evolved) intelli-
gence. Might this imply that, endowed with universal linguistic Turing machines
(i.e., Turing-universal minds) analogous to ours, ETIs would share our sense of
beauty? We think it likely.

Conclusion: Language and Mind Across the Universe

In the foregoing, we have argued that that the human language faculty is opti-
mally designed for maximal computational simplicity and is instantiated in the
178 Ian G. Roberts, Jeffrey Watumull, and Noam Chomsky

fundamental operation Merge, and that the likelihood is that any recognizably
intelligent entity would possess the same system. The arguments from simplicity
and optimality take us close to the actual conceptual necessity of this conclusion.
What does this mean in the context of potential contact?
Let us assume the existence of intelligent ETIs who have developed a tech-
nological civilization to at minimum a human-level technological sophistication
such that they would be capable of conceiving the existence of civilizations alien
to them, such as us, and able to contemplate sending/receiving a signal of some
form. To be precise, we assume “human-level of technological sophistication”
to include an understanding of fundamental mathematics (particularly comput-
ability theory), fundamental physics, whatever mechanisms (if any) enter into
biological (or its equivalent) evolution in their environment, etc. The considera-
tions we have raised in the foregoing mean that it is all but logically impossible
that such a level of knowledge could be attained without a “language” for gen-
erating, storing, and communicating information. Languages are based on gram-
mars: systems of primitives, principles, parameters, and procedures encoded in
cognitive mechanisms that determine the possible structural properties of lan-
guages. Modern linguistic theory proposes the existence of a species-specific
cognitive capacity, UG, that predisposes all human children (in normal environ-
ments) to acquire the language(s) to which they are exposed. The central ques-
tion for xenolinguistics is the degree to which UG, as we currently understand
it, would resemble ETI grammar. We have argued that the essential architecture
of ETI UG must be virtually identical to that of human UG. Fundamental to the
human – and ergo ETI – architecture are three axioms: (I) a “linguistic Turing
machine” – a computable function – generative of an infinite set of hierarchi-
cally structured expressions that interface (at minimum) two extra-linguistic
systems; (II) a conceptual-intentional system; (III) an externalization system for
communicating conceptual-intentional information. We have argued that human
evolution converged on (I) as a globally optimal solution to generating the con-
ceptual structures represented in (II); there is also a system for connecting them
to externalization in (III), whose details we have largely left aside here. The com-
putational optimality/efficiency of (I) derives from a recursive structure-building
operation, capable of constructing structures of unbounded complexity by itera-
tively applying to its own output. In human UG, this operation is Merge. The
overwhelming likelihood is that ETI UG would also be based on Merge. Thus,
the greatest difficulty in communicating with ETIs would be posed not by their
grammar, but in understanding their externalization system; however, we submit
that this is an engineering problem which should pose no difficulties in principle
(although its practical nature is hard to foresee).
Ultimately, this theory of UG for humans and ETI (AI) can enkindled a more
general, grander, unified theory of Life, Information, Language, and Intelligence
(see Watumull and Chomsky, “To appear”).6
 Universal Grammar 179

Notes
1 The machine learning systems (e.g., deep learning neural networks powering large lan-
guage models) so popular in the current “AI spring” are weak AI: brute-force systems
laboriously trained to “unthinkingly” associate patterns in the input data to produce out-
puts that approximate those data in a process with no resemblance to human cognition
(thus betraying Turing’s original vision for AI). These systems will never be truly intel-
ligent, and are to be contrasted with the strong – anthronoetic – AI Turing envisioned:
a program designed to attain human-level competence with a human-style typified by
syntactic generativity and semantic fluidity – to think the way a human thinks. See Cope-
land (2004) for more. Today such programs, based on generative grammars, are finally
being built at Oceanit.
2 Such an inefficient and unintelligent technique is the modus operandi of many machine
learning (weak AI) systems.
3 Indeed, we might speculate that were we to “wind the tape of life back” and play it again,
in Stephen Jay Gould’s phrasing, not only would something like Merge re-emerge, but
something like humans could well be “inevitable”, as serious biologists have suggested
(see Conway Morris 2013).
4 Convergence is a consequence of constraints. As with intelligence, evolution and de-
velopment are possible only by coupling scope with constraints. Stated generally, the
scope of any creative process is a function of its operating within limits. In the context
of evolution, for instance, Stuart Kauffman (1993: 118) observes,
Adaptive evolution is a search process – driven by mutation, recombination, and
selection – on fixed or deforming fitness landscapes. An adapting population flows
over the landscape under these forces. The structure of such landscapes, smooth or
rugged, governs both the evolvability of populations and the sustained fitness of their
members. The structure of fitness landscapes inevitably imposes limitations on adap-
tive search.
The analogy to mind is deeply nontrivial, for “intellectual activity consists mainly of
various kinds of search” (Turing 1948: 431).
5 This remove/replace formulation mirrors that of the original remove/replace formula-
tion of Merge (see Chomsky 1995), which we have revised here so as not to stipulate a
separate “remove” step. In our formulation, there is simply “replace”.
6 www.youtube.com/watch?v=gqTyg_W_yHI

References
Berwick, Robert C., and Noam Chomsky. 2016. Why Only Us. Cambridge, MA: MIT Press.
Berwick, Rober C., Paul Pietroski, Berach Yankama, and Noam Chomsky. 2011. “Poverty
of the stimulus revisited.” Cognitive Science 35: 1207–1242.
Chomsky, Noam. 1951. The Morphophonemics of Modern Hebrew. Master’s thesis. Univer-
sity of Pennsylvania, Philadelphia.
Chomsky, Noam. 1975. Reflections on Language. New York: Pantheon.
Chomsky, Noam. 1995. The Minimalist Program. Cambridge: MIT Press.
Chomsky, Noam. 2000. New Horizons in the Study of Language and Mind. Cambridge:
Cambridge University Press.
Chomsky, Noam. 2002. On Nature and Language. Cambridge: Cambridge University
Press.
Chomsky, Noam. 2005. “Three factors in language design.” Linguistic Inquiry 36: 1–22.
Chomsky, Noam. 2006. Language and Mind. Cambridge: Cambridge University Press.
180 Ian G. Roberts, Jeffrey Watumull, and Noam Chomsky

Chomsky, Noam. 2008. “On phases.” In Foundational Issues in Linguistic Theory: Essays
in Honor of Jean-Roger Vergnaud, edited by Robert Freiden, Carlos P. Otero, and Maria
Luisa Zubizarreta, 133–166. Cambridge: MIT Press.
Chomsky, Noam. 2013. “Problems of projection.” Lingua 130: 33–49.
Chomsky, Noam. 2016. What Kind of Creatures Are We? New York: Columbia University
Press.
Chomsky, Noam. 2017. “The language capacity: Architecture and evolution.” Psychonomic
Bulletin and Review 24: 200–203.
Chomsky, Noam, Ángel J. Gallego, and Dennis Ott. 2017. “Generative grammar and the
faculty of language: Insights, questions, and challenges.” Catalan Journal of Linguistics.
Cocke, John, and Marvin Minsky. 1964. “Universality of tag systems with P = 2.” Journal
of the Association for Computing Machinery 11: 15–20.
Conway Morris, Simon. 2013. Life’s Solution: Inevitable Humans in a Lonely Universe.
Cambridge: Cambridge University Press.
Copeland, Jack B., ed. 2004. The Essential Turing. Oxford: Clarendon Press.
Davis, Martin. 1958. Computability and Unsolvability. New York: McGraw-Hill Book
Company.
Descartes, Rene. 1637 [1995]. The Project Gutenberg eBook of a Discourse on Method.
Translated by John Veitch. https://www.gutenberg.org/files/59/59-h/59-h.htm
Gallistel, Charles Randy, and Adam Philip King. 2010. Memory and the Computational Brain:
Why Cognitive Science Will Revolutionize Neuroscience. New York: Wiley-Blackwell.
Goodman, Nelson. 1943. “On the simplicity of ideas.” Journal of Symbolic Logic 8: 107–121.
Hauser, Marc D., Charles Yang, Robert C. Berwick, Ian Tattersall, Michael J. Ryan, Jeffrey
Watumull, Noam Chomsky, and Richard C. Lewontin. 2014. “The mystery of language
evolution.” Frontiers in Psychology 5.
Huybregts, Riny. 2017. “Phonemic clicks and the mapping asymmetry: How language
emerged and speech developed.” Neuroscience and Biobehavioral Reviews.
Kauffman, Stuart. 1993. The Origins of Order: Self-Organization and Selection in Evolu-
tion. Oxford: Oxford University Press.
Minsky, Marvin. 1961. Recursive unsolvability of Post’s problem of “tag” and other topics
in the Theory of Turing machines. Annals of Mathematics 74: 437–455.
Minsky, Marvin. 1985. “Why intelligent aliens will be intelligible.” In Extraterrestrials:
Science and Intelligence, edited by Edward Regis, 117–128. Cambridge: MIT Press.
Obata, Miki, Samuel Epstein, and Marlyse Baptista. 2015. “Can crosslinguistically variant
grammars be formally identical? Third factor underspecification and the possible elimi-
nation of parameters of UG.” Lingua 156: 1–16.
Peirce, Charles. 1903. Collected Papers of Charles Sanders Peirce (Volumes I–VI, edited
by C. Hartshorne and P. Weiss; Volumes VII–VIII, edited by A.W. Burks). Cambridge:
Harvard University Press.
Post, Emil. 1943. “Formal reductions of the general combinatorial decision problem.” Amer-
ican Journal of Mathematics 65: 197–215.
Roberts, Ian G. 2012. “Macroparameters and minimalism: A programme for comparative
research.” In Parameter Theory and Linguistic Change, edited by Charlotte Galves, So-
nia Cyrino, Ruth Lopes, Filomena Sândalo, and Juanito Avelar, 320–335. Oxford: Ox-
ford University Press.
Roberts, Ian. 2017. The Wonders of Language or How to Make Noises and Influence People.
Cambridge: Cambridge University Press.
 Universal Grammar 181

Roberts, Ian. 2019. Parameter Hierarchies and Universal Grammar. Oxford: Oxford Uni-
versity Press.
Roberts, Ian, and Jeffrey Watumull. 2015. “Leibnizian linguistics.” In 50 Years Later:
Reflections on Chomsky’s Aspects, edited by Ángel J. Gallego & Dennis Ott, 211–222.
Cambridge: MIT Press.
Shannon, Claude E. 1948. “A mathematical theory of communication.” Bell System Techni-
cal Journal 27: 379–423, 623–656.
Tattersall, Ian. 2012. Masters of the Planet: The Search for Our Human Origins. New York:
Palgrave Macmillan.
Turing, Alan. 1936. “On computable numbers, with an application to the Entschidungsprob-
lem.” Proceedings of the London Mathematical Society 42: 230–265.
Turning, Alan. 1948. Intelligent Machinery. Reproduced in Copeland (ed.) 2004. Page num-
bers as in Copeland.
Ulam, Stanislaw. 1976. Adventures of a Mathematician. New York: Scribner’s.
Watumull, Jeffrey. 2015. The Linguistic Turing Machine. Doctoral dissertation. University
of Cambridge, Cambridge.
Watumull, Jeffrey, and Noam Chomsky. “To appear.” Life, Information, Language,
Intelligence.
Wheeler, John Archibald. 1986. “How come the Quantum?” Annals of the New York Acad-
emy of Sciences 480: 304–316.
Wolfram, Stephen. 2002. A New Kind of Science. Champagne, IL: Wolfram Media.
16
WHERE DOES UNIVERSAL GRAMMAR
FIT IN THE UNIVERSE?
Human Cognition and the Strong Minimalist
Thesis

Bridget D. Samuels and Jeffrey Punske

Introduction
Hauser et al. (2002) begin by imagining a Martian visiting Earth, noting that such
a creature would likely contrast the shared genetic code of Earth’s living creatures
with the lack of universality in their communication systems. On the other hand,
human languages seem to be constrained by certain principles, subject to para-
metric variation, known as universal grammar. In this chapter, we will explore a
situation analogous to the one in which this hypothetical Martian finds itself: how
likely would humans be to find commonalities between human language and an
extraterrestrial (ET) communication system? Thinking about this question allows
us to define the limits of “universality” in universal grammar, as well as to refine
our understanding of the interface between the essential parts of language and the
human cognitive and motor-sensory systems that interact with them.
All components of language are subject to physical and biological constraints
that have come to be known as the “third factor in language design” (Chomsky
2005: 6). To the extent that such principles are grounded in physics, any ET lan-
guage would be subject to them. However, we hypothesize that the possibility of
radically different cognitive and externalization systems could make ET communi-
cation systems very unlike Earthly ones (on the variety of communication modali-
ties on Earth, see the contributions to this volume from Kershenbaum [Chapter
2], Herzing [Chapter 4], Pepperberg [Chapter 6], and Slobodchikoff [Chapter 9]).
Our second hypothesis follows from this. Despite the fact that human languages
vary considerably in almost every respect, there is robust evidence for a universal
syntactic spine; however, the origins of such a structure have remained largely
unexplored. We hypothesize that this spine arises from third-factor principles, and
as such, may be present in some form in ET languages. This could result in some

DOI: 10.4324/9781003352174-16
 Where Does Universal Grammar Fit in the Universe? 183

deeper commonality between human and ET languages on a syntactic level, despite

the expected external differences – thus recapitulating some of the variation seen
among languages on Earth.

What Is Language?
Differing definitions of “language” may be found throughout philosophy, psychol-
ogy, cognitive science, linguistics, and other fields. Attempts to provide simple
definitions are elusive. As Chomsky has noted, the term “language” can be mean-
ingfully divided into an internal/intensional definition (his “i-language”) and an
external one (his “e-language”) (see Chomsky 1986). Generative linguistics gener-
ally confines itself to the study of i-language (i.e., one’s mental grammar) often
through e-language objects (i.e., the sentences one produces).
We want to note differences between a mental grammar and its possible outputs.
For instance, the underlying grammatical systems for a proficient English speaker
will allow for the production of a sentence of infinite length though an application
of a simple rule like the toy rule provided in (1):

1 “Mary said {X}” is a grammatical sentence, provided that X is a grammatical

sentence.

We can expand our language to two toy grammatical rules:

2 “Bill left” is a grammatical sentence.

3 “Mary said {X}” is a grammatical sentence, provided that X is a grammatical
sentence.

Allowing other names to take the place of “Mary” for the sake of clarity, we can
see that these two simple toy rules would produce a “grammar” that would allow
for an infinitely long sentence:

4 John said Sarah said Dave said . . . Mary said Bill left.

Such a sentence, while a possible output of the underlying grammatical system,

would never be found within the e-language. No infinite sentence can exist as a
produced object. In short, our (human) cognitive, grammatical capacity exceeds
our ability to produce.
Hockett (1960) famously introduced thirteen design features which distinguish
language from other communication systems (for further discussion see the chap-
ters in this volume by Kershenbaum [Chapter 2], Ross [Chapter 12], and Slobod-
chikoff [Chapter 9]): (1) use of the vocal-auditory channel; (2) rapid fading; (3)
broadcast transmission and directional reception; (4) interchangeability; (5) total
feedback; (6) specialization; (7) semanticity; (8) arbitrariness; (9) discreteness;
184 Bridget D. Samuels, Jeffrey Punske

(10) displacement; (11) productivity; (12) traditional transmission; (13) duality of

pattern.
It is important to contextualize Hockett’s work. Hockett’s goal was to develop
the tools for examining the evolution of human language – his thirteen design fea-
tures are thus based on the evolutionary development of human and non-human
primates. The idea that Hockett’s design features were meant to capture “lan-
guage” on a truly universal scale is a mischaracterization of his goals. Further,
we have the benefit of research that was unavailable to Hockett. This consists
largely of work by Stokoe (beginning in 1960) and many others, which has shown
that signed languages such as American Sign Language are as complex as spoken
languages and are thus fully formed human languages, removing (1) “use of the
vocal-auditory channel” as a true design feature (see also Ortner, this volume
[Chapter 5]).
Such observations are critical for the present discussion because the evolution-
ary path language might take with a non-human intelligence may not produce all
of Hockett’s design features. Thus, it is critical to determine which of Hockett’s
features are required for any potential communication system to be considered lan-
guage. We believe six factors meet this criteria: (4) interchangeability; (5) total
feedback; (8) arbitrariness; (10) displacement; (11) productivity; (12) traditional
transmission. We further consider one additional feature, namely discreteness.
Human languages are composed of discrete elements at various levels of descrip-
tion (though cf. Nixon and Tomaschek, this volume [Chapter 17]). Individual
sounds combine to form larger units (morphemes and words) which themselves
combine into complex phrases and sentences. Sentences then combine into a dis-
courses, and so forth. The sizes of the discrete meaningful units vary significantly
from language to language, but all human languages have discreteness. Whether or
not this is a required element of language is a question we discuss in the following
sections.
This proposed definition allows for some flexibility, so that just as human lan-
guages are not limited by modality (vocal-auditory/sign-visual), potential ET lan-
guages may also exhibit multiple modalities that are less familiar. However, the
core cognitive elements would remain the same.

The Human Language Faculty

The constellation of abilities known collectively as the linguistic competence and
performance systems are made possible by complex interactions between physi-
ological components of the human body – including but certainly not limited to the
brain – and the external world. These physiological components are often called
“the language faculty.” The components of the language faculty and their emer-
gence in our species have been the subject of much scrutiny by researchers who are
interested in how language evolved among humans. We believe this approach can
be instructive for this work, as well.
 Where Does Universal Grammar Fit in the Universe? 185

What Is Unique to Humans?

Based on the definition of language proposed in the previous section, it seems clear
that at least three components are necessary to characterize language: an abstract
computational system to generate syntactic structures, a conceptual-intentional
system to endow those structures with meaning, and a sensory-motor system to
externalize the structures produced by the former two. Hauser et al. (2002; hence-
forth HCF) refer to these three components collectively as the faculty of the lan-
guage in the broad sense, or FLB. They refer to the computational system alone
as the faculty of language in the narrow sense, or FLN, and hypothesize that only
FLN is uniquely human. That is to say, the remainder of FLB is drawn from a
number of systems that are not special to language, or even to our species. HCF
explicitly exclude “other organism-internal systems that are necessary but not suf-
ficient for language (e.g., memory, respiration, digestion, circulation, etc.)” from
FLB (Hauser et al. 2002: 1571). We will do the same here and set aside issues con-
cerning these systems, keeping in mind that they could differ significantly in ETs.
What about the FLN is unique to humans and language? This has been a con-
troversial question in the years since it was first posed, and we do not have the
space to recap the arguments here. For present purposes, we will just outline two
possibilities. First, HCF suggested that the uniqueness of FLN may lie in syntactic
recursion, sometimes known as the operation “Merge.” Merge is a set-formation
procedure that takes two syntactic objects (roughly, morphemes), X and Y, and
combines them into a set {X, Y} (see, e.g., Chomsky 2012; Boeckx 2016; Roberts
et al., this volume [Chapter 15]). Merge can apply to its own output, thus yielding
a set {Z, {X, Y}}, and so on. The recursive property of Merge thus entails that
syntactic structures are unbounded in length, as in (4). No other species in the
animal kingdom seems to have communication systems powered by such a recur-
sive engine. However, some researchers have raised the interesting possibility that
Merge may have unlocked other cognitive abilities in humans, and we will return
to this later in the chapter.
Another possibility, which does not necessarily preclude the uniqueness of
Merge, is that we humans stand alone in our ability to turn concepts into syntactic
objects – to lexicalize them. This line of reasoning has been pursued by Spelke
(2003) and Boeckx (2016). In this view, we are special in our ability to take con-
cepts from any cognitive domain and turn them into the proper “currency” for
Merge. This would involve a sort of wrapper enabling the concept to enter the
syntactic computation (Boeckx 2016: 27 suggests this is the “edge feature” identi-
fied by Chomsky 2008) plus “instructions for the syntax-external systems to ‘fetch’
(‘bind’/activate/retrieve in long-term memory) the relevant concept on the seman-
tic side or the relevant phonological matrix on the sound/sign side” (Boeckx 2016:
28). One could imagine a faculty of language that would have Merge but not free
lexicalization, such that it would be possible to create sentences of infinite length,
but only from a predefined set of concepts. Alternatively, one could conceive of a
186 Bridget D. Samuels, Jeffrey Punske

faculty of language with free lexicalization but some structure-building operation

that is more limited than Merge. In that case, a sentence might be highly con-
strained in its form and length, but could slot an infinite number of concepts into
its templates. In light of these thought experiments, it seems likely to us that both
Merge and lexicalization are crucial parts of the human language faculty, and may
indeed be unique to humans and to language.

The Strong Minimalist Thesis

Intuitively, the meaning of a linguistic object – the thought behind it – is what
feels truly important. However, the meaning of a phrase or sentence is structur-
ally compositional, so the order in which the elements undergo Merge is crucial
to the eventual meaning. {Pat, {likes, Sam}} thus means something different from
{Sam, {likes, Pat}}. Helpfully, the order in which elements are merged is reflected
in the order in which they are pronounced/signed, subject to a limited range of
language-specific variation. More generally, syntax serves as the spine that struc-
tures meaning and externalization. Chomsky (2000 et seq.) hypothesized that the
human language faculty is in fact the optimal solution to this set of demands; it
balances the conditions dictated by the conceptual-intentional (CI) and sensory-
motor (SM) systems with which it interfaces, such that language is actually usable
for externalizing thoughts. This hypothesis is known as the strong minimalist thesis
(SMT). It has been further proposed that the CI interface is of primary importance
for the language faculty; the SM interface may be ancillary (Chomsky 2008; see
also Samuels 2011). Another way to say this is that the language faculty is primar-
ily “for” thought, with its usefulness for communication as a bonus.
The SMT may seem like a bold claim. What reason do we have to believe that
the language faculty is optimal? The view is sometimes expressed that evolution
“satisfices,” or does a just-good-enough job. However, there is reason to be more
optimistic: there are numerous examples whereby natural selection has indeed
found an optimal solution or remarkably close to it (Cockell 2018 provides an
accessible overview). Computational modeling has shown that, on a number of lev-
els, animals – including both invertebrates and vertebrates (and notably primates
like us) – have the “best-of-all-possible” neural wiring (Cherniak 2012: 361). It
may not, then, be such a stretch to take the SMT as our starting point (see Samuels
2011 for further arguments).

Three Factors in Language Design

Thus far, our discussion of FLB/FLN and the SMT have focused entirely on
the human organism and primarily on the brain as the locus of cognition. These
structures are largely determined by our genetics, but we cannot ignore organism-
external contributions to language. If a child grows up hearing English, they will
learn English; if they grow up seeing American Sign Language, they will learn
 Where Does Universal Grammar Fit in the Universe? 187

that. The externalized portion of language – roughly, what sounds it uses and
in what order, and what groups of sounds map to what meanings (i.e., the lexi-
con) – must be gleaned from the environment and committed to memory. This, of
course, is a massive oversimplification of a language learner’s task (see Nixon and
Tomaschek, this volume [Chapter 17], for additional discussion), but we set this
issue aside to discuss another contribution to language: that of physical laws and
other overarching, domain-general principles of computation and data analysis,
or as Chomsky (2005) calls them, the “third factor” in language design. The third
factor constrains the genetic (first) and environmental (second) factors just men-
tioned. In the words of Cockell (2018: 239), “the laws of physics and life are the
same.” Given that the laws of physics are truly universal, examining third-factor
effects might give some insight into possible constraints in forms of extraterres-
trial language.

Syntax
Much of the work of modern linguistics has illustrated how seemingly wildly
disparate and distinct phenomena are actually the same at a fundamental level.
In this section, we explore whether these results would necessarily obtain for ET
language. We argue that the core universals of human language are constrained
by two separate sources: general human cognitive organization (the FLB) and
third-factor principles. We consider the hypothesis that potential ET languages,
governed by the same third-factor principles as languages on Earth, could arrive
at the same FLN as human language but differ in other aspects of their FLB. As
a consequence, the likelihood of interstellar interlinguistic communication seems
remote, but possible.

Language Diversity, Universality, and Human Cognition

Human languages exhibit a great degree of variation; however, this variation is
constrained. For instance, among the possible orderings of the three key elements
of a sentence,– namely the subject, verb, and object – there seems to be an over-
whelming, if not universal, tendency for the subject to proceed to the object. The
subject typically precedes the object in more than 83% of the 1,377 languages
surveyed in the World Atlas of Linguistic Structures (Dryer 2013).
The preference for subject-predicate and subject-object orders is an area
that demands an explanation from formal linguistics – especially given that
syntactic operations can obscure the underlying structures. However, the core
preference itself is arguably part of a more general cognitive preference regard-
ing the ordering and organization of events and thus is best understood as part
of the FLB.
Within generative syntax, the most common approach is to assume a universally
ordered hierarchy of syntactic structure (“the spine”) (see, for instance, Cinque
188 Bridget D. Samuels, Jeffrey Punske

1999; Wiltschko 2014). A generalized version of this is given in (5) (cf. Wiltschko
2014):

5 Generalized Universal Spine

Discourse
Domain

Grammatical
Relations/
TAM
Thematic
Domain

The spine allows for considerable linguistic variation. Various types of movement
and other operations may apply to the lexical elements in the tree. And, follow-
ing Ritter and Wiltschko (2014), the semantic substance of particular nodes may
also vary. They examine the node INFL, which they argue is best understood as a
general “anchoring” node. In a language like English, INFL anchors the sentence
to time (tense). In languages like Halkomelem, the anchoring is spatial. In lan-
guages like Blackfoot, the anchoring is based on discourse participants. We can see
the same general core functions in the same syntactic positions across the world’s
languages. How this would arise from something as simple as combining two ele-
ments together (i.e., Merge) is not obvious.
The question remains whether this “spine” lies within the domain of FLN;
i.e., is it specific to humans and to language? We hypothesize that it does not. It
seems quite plausible that the universal spine is the result of syntax conforming to
demands from the CI systems (belonging to FLB) with respect to how we interpret
scenes and events; we agree with Sperlich (this volume [Chapter 14]) that if extra-
terrestrials share our “cause-and-effect” view of the world, they might also share
our relations of subject ~ agent, object ~ undergoer/patient, and so forth.
The question of how the syntactic spine arose, and whether to situate it in uni-
versal grammar or rather in extralinguistic principles, remains open. Boeckx (2016,
sections 2.3–2.4) provides one proposal, which reduces lexical and functional cat-
egories (e.g., nouns, verbs, adjectives, prepositions, tense, etc.) to variations on two
basic categories with differing syntactic properties. When we look more broadly at
non-linguistic intelligences on Earth, we see evidence for “linguistic” features in
a non-linguistic context. One very clear example of this is Golston’s (2018) work,
which argues that φ-features (person, number, gender) are found throughout animal
 Where Does Universal Grammar Fit in the Universe? 189

cognition, though other chapters in the present volume, in particular those of Pep-
perberg (Chapter 6) and Slobodchikoff (Chapter 9), serve as cautionary reminders
that the mental worlds of other species, even on Earth, can differ tremendously
from ours. This, again, is broadly consistent with the idea being advanced here that
much of the work of generative syntax is actually an examination of the FLB in a
particular context.

Could ETs Have Human-Like Syntax?

Given that the core syntactic operation, Merge, appears to be unique to humans, it
may seem counterintuitive to argue that ETs could have Merge, as well. However,
we find this to be an intriguing possibility with some arguments in its favor. Chom-
sky (2008: 6) and others have claimed that Merge can define the basic mathemati-
cal operation of addition: the recursive algorithm known as the successor function,
which states that every number N has a unique successor, N + 1. If our arithmetic
abilities are related to Merge, this would explain why humans alone on Earth have
the property of discrete infinity both in our language and in our system of natural
numbers. While many other species can represent small numerosities precisely (up
to ~4) and compare quantities approximately (see Dehaene 1997 for an overview),
it may be that Merge integrates these two systems, enabling the precise representa-
tion of a discrete infinity of numbers (Hiraiwa 2017).
If we encounter a technologically advanced ET species – one that has built a
spacecraft or a radio transmitter, for example – it stands to reason that they have
some system for representing mathematics. Minsky (1985) makes the strong claim
that they will not only have mathematics, but they will have the same arithme-
tic system as ours, because “there are simply no easily accessible alternatives. . . .
[T]here is nothing which resembles [it] that is not either identical or vastly more
complicated.” He supports this view with an analysis of possible Turing machines
and some thought experiments. Depending on one’s view of mathematical realism,
these arguments may or may not be compelling. But if it is correct that Merge is
intrinsically linked to our mental arithmetic, and if that mental arithmetic is shared
by ETs, then there is a possibility that ETs could also use Merge to structure their
language (a conclusion that we appear to largely share with Roberts et al., this vol-
ume [Chapter 15]). This operation could potentially then interact with the FLB of
ET species, which could differ dramatically from those of humans.

Externalization Systems
It is easy to see that the externalization of human language exhibits tremendous
variation. Languages can be spoken or signed, thus making use of multiple sensory
modalities (auditory, visual, and tactile for the deaf/blind). Within these modalities,
languages use different repertoires of sounds, hand shapes, and motion patterns,
and place different restrictions (phonotactics) on the sequencing or co-occurrence
190 Bridget D. Samuels, Jeffrey Punske

of these elements. Moreover, the relationship between sound/sign and meaning is

arbitrary; there is nothing about the word dog that represents the quality of canine-
ness. There is also variation in how syntactic structures are linearized, resulting in
word-order differences, as mentioned in the previous section. In sum, the external-
ized portion of a given language must be learned. Most, if not all, of our language
externalization system is shared with or convergently evolved in other species and
thus constitutes part of FLB but not FLN (see Samuels 2011, 2015 and references
therein).
The range of variation in these aspects of human language is large but not limit-
less (see Samuels 2011 for discussion). One could imagine ETs using other means
of biological or technological transmission that they can perceive, control, and pro-
duce. Some possibilities have been explored elsewhere, including in several chap-
ters in this volume, and include various portions of the electromagnetic spectrum,
sound, and chemicals. We refrain from further speculation here.
Though the possibilities are many, third-factor principles should constrain ET
language just as they constrain ours. For example, we expect that an ET language
would have a vocabulary of discrete units that can be combined. Jackendoff (2011:
604) summarizes this argument, which dates back at least to Hockett (1960):

[T]he only way to develop a large vocabulary and still keep vocal signals distin-
guishable is to digitize them. This is a matter of signal detection theory, not of
biology, and it is also responsible for the fact that we all have digital rather than
analog computers. So we might want to say that the digital property of phonol-
ogy comes by virtue of “natural law.”

That is to say, the discreteness of linguistic units is a third-factor effect; despite

Jackendoff’s use of the word “vocal” in the preceding passage, there is no reason
to suspect that this is dependent on modality. Kershenbaum (this volume [Chapter
2]) raises the possibility that animal cognition and communicative signals may be
non-discrete, but to our knowledge, empirical support for this view is lacking.
Furthermore, the perceptual system ought to be able to tolerate some noise in
the signal; “the importance of error correction” is also discussed in a somewhat
different context in Granger et al.’s contribution to this volume (Chapter 8). In
humans and other animals, this is partly achieved through categorical perception.
This allows us to treat every token of the word cat as belonging to the same abstract
category (type), even though no two naturally produced tokens are ever exactly
alike in volume, pitch, vocal timbre, and so forth (for an accessible introduction to
speech acoustics and physiology, see Lieberman 2018). For sound, these categories
appear to have their basis in biases of our inherited mammalian auditory system
(see Samuels 2011), but they need to be learned on a language-specific basis. If
ET languages also have a learned component, we should expect them to have a
capacity akin to vocal learning (though, of course, the modality may differ), i.e.,
the ability to commit new categories to memory and to imitate them. This ability
 Where Does Universal Grammar Fit in the Universe? 191

is rare on Earth, although it evolved independently multiple times and is present in

some birds, cetaceans, pinnipeds, elephants, and bats (see Lattenkamp and Vernes
2018 for an overview).
Aerodynamics and biomechanics jointly define what sounds humans can make
(see e.g., Marchal 2009 and Nixon and Tomascheck, this volume [Chapter 17], for
an overview), and the same should be true of ETs that employ biological means
of linguistic externalization. The gas laws are highly relevant to our production
of sound, which uses a system of air cavities, pistons, and valves to manipulate
our breath. Our tongue, larynx, and chest control gas volume and pressure, which
are related by Boyle’s Law. The force needed to cause the lungs to expand within
the ribcage is subject to Hooke’s Law, and the pressure of air in the lungs must
be understood in comparison to atmospheric pressure. Atmospheric pressure will
vary from planet to planet, but the gas laws remain the same. If ETs manipulate
volumes of gases to produce language, they will apply. Cockell (2018: 62) provides
numerous examples of how physical laws affect biology in domains besides that
of linguistic production, and concludes that “the ubiquity of convergent evolution
across the biosphere and the vast number of examples show that solutions [to those
laws] are not limitless, but usually rather few.”

Conclusions: Would an ET Language Be Comprehensible?

When we examine human languages, we see they exhibit particular patterns of
their internal elements that do not follow from the SMT. Put more succinctly, word-
order patterns and many phonological properties may be better understood as a
product of the FLB rather than the FLN.
The central thesis of this chapter, as it relates to the interaction of CI systems
and language, is that the many of the core, universal properties of human lan-
guages draw more broadly from our general cognitive systems and preferences.
These preferences are part of our genetic endowment and are shared with many
other species on Earth because of our shared evolutionary paths. However, only
humans have language because of the unique contribution of our FLN (Merge and/
or lexicalization).
We have raised the possibility that human language and a potential ET language
could share some of the properties of the FLN, specifically Merge, in addition to
properties determined by third-factor principles, such as discreteness. It is very
likely that an ET message structured by these principles would be recognizable as
a language. However, given the potential differences in the FLB, especially major
potential differences in the CI systems, the ability to parse and understand such a
message is highly questionable. The question of whether ETs could communicate
about any concepts they could form – in other words, whether they could have free
lexicalization, as discussed in the third section of this chapter – is impossible to
answer. Thus, the major obstacle preventing us from understanding ET languages,
or vice versa, may ultimately not be one of exo-linguistics but one of exo-cognition.
192 Bridget D. Samuels, Jeffrey Punske

References
Boeckx, Cedric. 2016. Elementary Syntactic Structures: Prospects of a Feature-Free Syn-
tax. Cambridge: Cambridge University Press.
Cherniak, Christopher. 2012. “Neural wiring optimization.” Progress in Brain Research
195: 361–371.
Chomsky, Noam. 1986. Knowledge of Language: Its Nature, Origin and Use. New York:
Praeger.
Chomsky, Noam. 2000. “Minimalist inquiries: The framework.” In Step by Step: Essays on
Minimalist Syntax in Honor of Howard Lasnik, edited by Roger Martin, David Michaels,
and Juan Uriagereka, 89–155. Cambridge, MA: MIT Press.
Chomsky, Noam. 2005. “Three factors in language design.” Linguistic Inquiry 36(1): 1–22.
Chomsky, Noam. 2008. “On phases.” In Foundational Issues in Linguistic Theory, edited
by Robert Friedin, Carlos P. Otero, and Maria Luisa Zubizarreta, 133–166. Cambridge:
MIT Press.
Chomsky, Noam. 2012. “Foreword.” In Phases: Developing the Framework, edited by Án-
gel J. Gallego, 1–7. Berlin: Mouton de Gruyter.
Cinque, Guglielmo. 1999. Adverbs and Functional Heads: A Cross-Linguistic Perspective.
Oxford: Oxford University Press.
Cockell, Charles S. 2018. The Equations of Life: How Physics Shapes Evolution. New York:
Basic Books.
Dehaene, Stanislas. 1997. The Number Sense: How the Mind Creates Mathematics. Oxford:
Oxford University Press.
Dryer, Matthew S. 2013. “Order of subject, object and verb.” In The World Atlas of Lan-
guage Structures Online, edited by Matthew S. Dryer and Martin Haspelmath. Leipzig:
Max Planck Institute for Evolutionary Anthropology. Accessed September 24, 2018.
http://wals.info/chapter/81.
Golston, Chris. 2018. “Phi features in animal cognition.” Biolinguistics 12: 55–98. Ac-
cessed September 18, 2018. www.biolinguistics.eu/index.php/biolinguistics/article/
view/603/419.
Hauser, Marc D., Noam Chomsky, and W. Tecumseh Fitch. 2002. “The faculty of language:
What is it, who has it, and how did it evolve?” Science 298: 1569–1579.
Hiraiwa, Ken. 2017. “The faculty of language integrates the two core systems of number.”
Frontiers in Psychology 8: 351.
Hockett, Charles F. 1960. “The origin of speech.” Scientific American 203: 88–96.
Jackendoff, Ray. 2011. “What is the human language faculty? Two views.” Language 87
(3): 586–624.
Lattenkamp, Ella Z., and Sonja C. Vernes. 2018. “Vocal learning: A language-relevant trait
in need of a broad cross-species approach.” Current Opinion in Behavioral Sciences 21:
209–215.
Lieberman, Philip. 2018. “Why human speech is special.” The Scientist, July. Accessed Sep-
tember 24, 2018. www.the-scientist.com/features/why-human-speech-is-special-64351.
Marchal, Alain. 2009. From Speech Physiology to Linguistic Phonetics. Hoboken, NJ:
Wiley & Sons.
Minsky, Marvin. 1985. “Communication with alien intelligence: It may not be as difficult as
you think.” Byte Magazine 10 (4): 127–138. Accessed April 21, 2023. https://web.media.
mit.edu/~minsky/papers/AlienIntelligence.html
Ritter, Elizabeth, and Martina Wiltschko. 2014. “The composition of INFL.” Natural Lan-
guage & Linguistic Theory 32 (4): 1331–1386.
 Where Does Universal Grammar Fit in the Universe? 193

Samuels, Bridget D. 2011. Phonological Architecture: A Biolinguistic Perspective. Oxford:

Oxford University Press.
Samuels, Bridget D. 2015. “Can a bird brain do phonology?” Frontiers in Psychology 6:
1082. doi:10.3389/fpsyg.2015.01082.
Spelke, Elizabeth. 2003. “What makes us smart? Core knowledge and natural language.” In
Language and Mind: Advances in the Study of Language and Thought, edited by Dedre
Gentner and Susan Goldin-Meadow, 277–311. Cambridge, MA: MIT Press.
Wiltschko, Martina. 2014. The Universal Structure of Categories: Towards a Formal Typol-
ogy. Cambridge: Cambridge University Press.
17
LEARNING AND ADAPTATION OF
COMMUNICATION SYSTEMS IN
BIOLOGICAL LIFE FORMS
Jessie S. Nixon and Fabian Tomaschek

Introduction
This volume contemplates what possible languages or communication systems
might be used by extra-terrestrial life forms, with the aim of creating some insight
into how to tackle the task of interpreting communication signals should we ever
encounter them. Other researchers have already addressed the issue that it is
exceedingly improbable that we will ever encounter extra-terrestrial communica-
tion. We leave that issue aside and focus on the value of this exercise for better
understanding terrestrial communication systems. This thought experiment allows
us to step back and take an outside perspective on the communication systems
here on Earth. As Michaud (2011) points out, in the absence of any information
about extra-terrestrial life, one of the only means available to us for predicting their
nature and behaviour is by analogy with ourselves. We take this same approach.
In this chapter, we examine terrestrial communication signals, focusing mainly
on human language; the adaptability of speech perception, including adaptation
to environmental cues; and the cognitive and learning mechanisms that have been
proposed to support such a complex, adaptive communication system. We discuss
implications for extra-terrestrial communication and for the task of interpreting
such communication signals.

Human Communication Systems

Human communication systems come in diverse forms and are extremely complex.
Some examples of communication systems that have emerged through pressure
for communication are sign language (Nyst 2015), gestures, visual cues such as
facial expression and body language, drum languages (Rattray 1923) and whistling

DOI: 10.4324/9781003352174-17
 Learning and Adaptation of Communication Systems 195

FIGURE 17.1 a) Schematic representation of the source spectrum created by the vocal
cords. b) Filter spectrum of the vocal tract, which functions as a reso-
nance filter. c) Spectrum of the articulated vowel. Each spectral peak rep-
resents a formant (F1, F2, F3). Changing the shape of the mouth (e.g., its
diameter) changes the relative location of the formants.
Source: Figure according to Fant (1960: 19)

languages (Meyer 2004).1 This diversity demonstrates a high degree of adaptivity –

as long as the sensory information is able to be discriminated to a sufficient level of
complexity, it can be used for communication. However, by far, the most common
form of communication is spoken language.
Human speech is primarily based on an acoustic signal which is created by move-
ments of the human masticatory organs – jaw, lips and tongue (Weber et al. 1986;
Yashiro and Takada 2005). The acoustic system used by humans is often divided
into two broad categories – consonants and vowels. “Consonants” constitute partial
or complete constriction of airflow. Constriction of airflow is achieved in various
ways, including closing of the jaw, placing various sections of the tongue against
different areas of the mouth, closing the lips and sealing the oral cavity to allow
airflow through the nasal cavity. Sounds created by an unconstricted airflow are
classified as “vowels”. To produce vowels, the vocal tract functions as a resonance
filter that filters the full spectrum signal created by the vocal cords (Figure 17.1a–b).
The acoustic signal contains several spectral peaks (Figure 17.1c), where the first
peak is called the first formant (F1), the second peak the second formant (F2), and
so forth. The spectral properties in Figure 17.1c represent the vowel schwa, which is
articulated by positioning the tongue in the center of the vocal tract and is equivalent
to the first vowel in words like “about” or the second vowel in “speaker”.
Moving the tongue in the oral cavity modifies the vocal tract’s shape, changing the
characteristics of the resonance filter, which results in different formant frequencies,
thus in different vowel qualities. For example, a high fronted tongue position creates
an [i], a lower central tongue position an [a] and a high-retracted tongue position with
lip rounding a [u] (cf. Figure 17.3a. Figure 17.3b shows the vowel space on the basis
of F1 (y-axis) and F2 (x-axis) frequencies. The black triangle represents the position of
the five human standard vowels in the F1/F2 vowel space. Note that the orientation of
the triangle in Figure 17.3b is equivalent to the triangle’s orientation in Figure 17.3a.
196 Jessie S. Nixon and Fabian Tomaschek

During speech, the various different speech apparatuses move dynamically,

leading to continual changes to the resonance filter. This means that the resonance
filter not only depends on the vowel type, but also, for example, on the surrounding
consonants – so the vowel in “bat” is different to the vowel in “can”, because both
preceding and following mouth shapes affect the resonance filter (Magen 1997;
Öhman 1966). If clustered by sound category, cue values vary around a centre with
considerable overlap between clusters. In other words, rather than being discrete,
with a consistent mapping to semantics or even to sound categories, acoustic cues
are continuous and variable. In fact, there is a multitude of sources of variability in
the signal that do not help the listener understand the message (Altmann 1980; Fos-
ler-Lussier and Morgan 1999; Fuchs et al. 2008; Gay 1978; Hirata 2004; Jiao et al.
2019; Leemann et al. 2012; Tomaschek and Leemann 2018; Tomaschek et al. 2013,
2014, 2018a, 2018b; Weirich and Fuchs 2006; Winkler et al. 2006) and which fall
outside the limits of what is traditionally considered the linguistic message, such
as indexical information. For example, age perception from voice depends on lin-
guistic-cultural familiarity with the speaker, suggesting that speech reflects – and
listeners learn – not only biological changes with age, but also language-specific
socio-cultural age effects (Jiao et al. 2019).
Speakers use the speech signal to convey a message. According to information
theory, communication boils down to the reproduction of a message at one point
from a message sent at another point (Shannon 1948). There is a set of shared codes
that point to a limited set of predefined messages. “Comprehension” of a message
simply means selecting the correct message from all the possible messages that
the sender could have sent. If we consider international Morse code prosigns, this
makes sense intuitively. Probably the most famous example is - - - – – – - - -, a
series of three short, three long, then three short signals. This code signals a distress
call in a case when there is an imminent threat to life or property. It is distinct from,
for example, a general call for attention or hail – - – - – or from a request for rep-
etition - - – – - - or “correction, my preceding message was an error”: - - - - - - - -.
The messages do not break down into components that “contain meaning”.2 Com-
munication of this message from the sender to the receiver is possible because
both parties share the same code and this code allows the receiver to discriminate
between a distress call, a request for repetition, an error correction and any other
possible message. Ramscar et al. (2010) argue that language is likewise discrimina-
tive. Cues in the signal serve to reduce uncertainty, enabling the receiver to select
which of the possible messages was intended, rather than each message being built
up from component parts or units of meaning (see Ramscar et al. 2010; Wittgen-
stein 1953, for discussion of the problems around meaning and reference).
Ramscar (2019) argues that while language is discriminative, in human lan-
guage, there is no predefined code at birth which is shared by all speakers. Instead,
the language develops over a lifetime through interaction with the world and
with other users of the communication system. The code is constantly negotiated
between sender and receiver. One consequence is that no two users have identical
 Learning and Adaptation of Communication Systems 197

code, so it is only partially shared. Ramscar (2019) also argues that language is
deductive. Cues reduce uncertainty by excluding unlikely messages. Using the pre-
viously discussed Morse code prosigns as an example, a receiver encountering a
distress call can immediately rule out (or at least have low expectations of) a hail
signal on hearing the first short signal; a request for repetition becomes unlikely on
the third short signal; and a correction message becomes unlikely on hearing the
first long signal. Of course, context also plays a role, so when there has been no
prior communication, a correction message or a request for repetition are already
unlikely before any signal is encountered. Similarly, in language, the context serves
to constrain the set of possible messages. The speech signal (if effective) serves to
further reduce uncertainty by eliminating alternatives – although in real communi-
cation, typically some degree of uncertainty remains, the extent of which depends
on many factors.

What Cognitive and Learning Mechanisms Support Such

a Complex System?
Although speech is generally an effective means of communication (to an extent),
the speech signal is so complex that after decades of study, we still on an aca-
demic level understand neither what cues make up the system nor the relationship
between auditory cues and the message. Some researchers propose that speech is
made up of a small set of discrete sound units (sometimes called “phonemes”).
However, if people are asked to identify words taken from recordings of running
speech, accuracy is as low as 20%–40% by participant (Arnold et al. 2017). If
speech consisted of strings of sound units, we would not expect such difficulties
when these individual units are extracted from speech (see Nixon and Tomaschek
2023a, for a full discussion of the continuous nature of the speech signal and a
review of how models of speech perception have taken this into account). Contrast
this with written language: a single word can generally still be read with high accu-
racy when extracted from a sentence.
A large body of evidence has accumulated over the last several decades showing
that human infants and adults are highly sensitive to the distributional characteristics
of cues in speech (Clayards et al. 2008; Maye et al. 2002; McMurray et al. 2009; Nixon
and Best 2018; Nixon et al. 2016, 2018; Yoshida et al. 2010). This line of research has
been groundbreaking, because it has demonstrated that humans have learning mecha-
nisms that enable them learn speech and language through exposure and interaction.
Prior to this, many researchers believed that language was too complex to learn and
must therefore be innate. However, while these data are often interpreted as suggest-
ing that learning occurs by directly learning the statistical structure through statistical
clustering mechanisms (e.g., Maye et al. 2002), computational models have shown
that unsupervised statistical clustering alone is not sufficient to account for learning
of speech sounds (e.g., McMurray et al. 2009). This has given rise to research into
alternative mechanisms that might account for learning of speech.
198 Jessie S. Nixon and Fabian Tomaschek

Recent work has discovered a role of error-driven discriminative learning in

speech acquisition (Nixon 2018, 2020; Nixon and Tomaschek 2020, 2021). Error-
driven learning is a learning system used by many animal species on Earth (Waelti
et al. 2001; Soto and Wasserman 2010). Learning processes in animals have
been meticulously studied for more than a century, from which well-developed,
well-tested mathematical, theoretical and computational models of learning have
emerged (e.g., Rescorla and Wagner 1972; Widrow and Hoff 1960). Despite enor-
mous surface variation in biological systems, error-driven learning models have
successfully predicted learning behaviour across a broad range of species, includ-
ing humans, and under a wide variety of conditions. What these studies demon-
strate is that in a wide variety of biological systems on Earth, learning involves
using incoming sensory information, or cues, to predict important events in the
world. When cues frequently predict a certain outcome, then experiencing that
cue again in the future leads to an expectation that the outcome will also occur.
Error-driven learning models have been remarkably successful in predicting and
explaining a broad range of phenomena in human language, including language
acquisition, comprehension and production (Arnold et al. 2017; Baayen et al. 2011,
2016; Nixon 2018, 2020; Nixon and Tomaschek 2020, 2021; Tomaschek et al.
2021, 2022; see Nixon and Tomaschek 2023b, and the contributions therein for an
overview).
Further research is needed to reconcile the various learning models with respect
to human language learning; it seems likely that multiple learning mechanisms
may be involved. Another aspect that needs to be considered for human learning
is the role of higher-order cognitive processes (Nixon et al. 2022; Reber 1989),
which seem to differ between children and adults in language learning (Ramscar
et al. 2013). For example, some researchers have modelled causal inference learn-
ing with error-driven learning models (Shanks 1986; Van Hamme and Wasserman
1994), but other research has shown that human cause attribution involves a com-
plex set of processes, including reasoning (Einhorn and Hogarth 1986). Anecdo-
tally, these sorts of higher-order cognitive processes seem to characterise second
language learning in the classroom more than first language or immersive learning.

Error-Driven Learning in Non-Human Animals

Error-driven learning is not restricted to humans, of course; its discovery emerged
from animal research. Perhaps the most famous learning experiments are Pavlov’s
experiments with dogs at the turn of the 20th century. Pavlov found that if he
repeatedly presented a stimulus, such as a bell, before feeding the dogs, eventually
the dogs would start salivating in response to the stimulus even when the food was
not present. Initially, this was considered a simple reflex response and thought to
be merely due to contiguity of the stimuli: the reflex response that belonged to the
food somehow got taken over by the new stimulus (e.g., Hull 1943; Spence 1956;
reviewed in Gluck and Bower 1988). However, in the decades that followed, it
 Learning and Adaptation of Communication Systems 199

a)Light b)Tone
Weight of light to shock

Weight of light to shock

0.8

0.8
0.4

0.4
start: tone start: tone
0.0

0.0
0 200 400 600 800 1000 0 200 400 600 800 1000

Events Events

c)Light d)Tone
tone is blocked tone is blocked
Weight of light to shock

Weight of light to shock

0.8

0.8
0.4

0.4

start: tone start: tone

0.0

0.0

0 200 400 600 800 1000 0 200 400 600 800 1000

Events Events

FIGURE 17.2 Illustration of the blocking effect estimated by means of Rescorla-Wag-

ner learning equations. a–b) Light and tone a cues for shock are trained
together. c–d) Light is pre-trained, tone is blocked. Vertical dashed line
indicates the beginning of training with the tone cue in addition to the
light cue.

became clear that, rather than a simple association, the animals were using infor-
mation in the cues to learn about the world through a process of prediction and
error feedback.
In a seminal study, Kamin (1968) trained rats with two different cues – a light
and a tone – followed by an electric shock (see Figure 17.2). In the test, if he pre-
sented just one of these cues, say the tone, the rats would exhibit a fear response to
the tone – they had learned that the tone predicted the shock. A separate group of
rats had identical training and test, except that before the light + tone training, there
was a pre-training phase during which the rats were trained with only one cue: the
cue that was not tested – in this case, the light. The pre-trained rats did not respond
200 Jessie S. Nixon and Fabian Tomaschek

to the tone. Because the light already provided sufficient information to predict the
shock, the tone added no additional predictive value. There was no longer sufficient
uncertainty left to drive learning of the tone. So the tone was not learned. In other
words, learning of the tone was blocked by the already-learned light cue.
The Rescorla-Wagner equations were developed to account for this finding, as
well as a wealth of data that had accumulated over the previous decades from
research on animal learning. The level of expectation of a particular outcome given
a particular cue or cues is described as the cue-outcome connection weight. Con-
nection weights are adjusted in each learning event. A few simple principles can
be inferred from these mathematical equations. First, for any cues not present in a
given learning event, there is no change to the connection weights. For cues present
during a learning event connection weights are 1) increased to all outcomes that
occur and 2) decreased to any outcomes that do not occur, but which have been
encountered previously. What is important is that the amount of change in 1) and
2) depends on the history of learning. When an outcome is highly expected (that is,
connection weights are strong due to previous learning), and the outcome occurs,
there is little surprise and therefore little learning, so although connection weight
increases, the increase is relatively small. In contrast, if an outcome is not expected –
that is, connection weights are weak, because for example, the cues have seldom
been encountered before or the cues have been encountered before, but the out-
come seldom occurred – and yet the outcome does occur, the surprise is greater and
therefore learning is greater. The increase in connection weights is greater than in
the former example.
These simple equations make strong predictions that have been powerful tools
for understanding learning behaviour. For example, learning is non-linear and cur-
rent learning depends on the history of learning. This leads to the famous “learn-
ing curve” – the idea that learning is fast in the beginning then tapers off as things
become increasingly predictable.

How the Speech of Extra-Terrestrial Intelligence Might Sound

Without any data about extra-terrestrial life, the possibilities for the types of com-
munication systems that might exist seem virtually limitless. Rather than attempt
to catalogue the possibilities, our approach in this section is to start with what we
know about human speech on Earth, as outlined previously, and make the subtlest
of adjustments to the conditions. We assume an acoustic speech signal, produced
with a similar biological system to the human speech apparatus. What we find is
that varying just one dimension – vocal tract size or atmospheric conditions – could
result in dramatic changes to a signal that is otherwise identical to human speech,
potentially rendering it unrecognisable.
Although human speakers are generally unaware of this, the acoustic charac-
teristics of their speech signal depend on 1) the air temperature and density, which
affect the speed of sound, c, and 2) their body size, which affects the size of their
 Learning and Adaptation of Communication Systems 201

vocal tract L. These variables can be used to calculate the formant frequencies of
[@] and therefore all other vowels using equation (1).

(1) Fn = (2n − 1) × c / 4L

Where

n = the formant’s number

c = the speed of sound
L = the length of the vocal tract

For an average human vocal tract whose length is L = 17.5 cm and the average
speed of sound c = 350 m/s (with air density ~ 1.20 kg/m and 20°C), formant fre-
quencies are F1 = 500 Hz and F2 = 1500 Hz. The human vowel space in Figure
17.3 is based on these measures. When temperature on Earth drops to −2°C at sea
level, speed of sound is slower (c = 319 m/s), directly affecting formant frequen-
cies produced by the same human speaker. Looking at the differences between the
planets in our own solar system in terms of their atmospheric composition and
therefore speed of sound (NASA n.d.), it is very unlikely that the atmosphere of the
extra-terrestrial life form’s home planet will have exactly the same physical charac-
teristics as our own. For example, speed of sound on Mars ranges from 244 m/s at
ground level to 220 m/s at the top of Olympus Mons at an altitude of 24 km (NASA
n.d.). Since speed of sound will be proportional to the density and temperature of
the life form’s planet, the spectral characteristics of the acoustic signal of extra-ter-
restrial life forms will vary, too, if they communicate by means of acoustic signals.
Let us inspect what effect variation in speed of sound has on vowel acoustics. To
illustrate the effects of atmospheric properties of the life form’s home planet, we
can manipulate the c parameter in equation (1). The greater c on the home planet,
the higher the formant frequencies, significantly affecting the vowel quality of the
produced sound.
Life forms that live on a planet whose atmosphere has c = 200 m/s will produce
vowels that have frequencies lower than human vowels (grey dashed triangle in
Figure 17.3b). An [a] produced by the extra-terrestrial intelligence will thus sound
to human ears like an [o]. When the speed of sound is higher than on Earth, e.g., c =
500 m/s, the life form will produce vowels whose formants are on average higher
(grey dotted triangle in Figure 17.3b). This means that the extra-terrestrial’s [u]
will sound like a [ə] and an [o] will sound like an [a]. As a side note, similar effects
can be achieved when humans speak in a helium-saturated atmosphere (MacLean
1966).
Variation in air density and temperature on Earth results in minimal – but meas-
urable – variations of the quality of these vowels. Furthermore, atmospheric charac-
teristics on Earth can vary on a daily basis, so humans need to learn to account for
these changes. This is also the case with crickets. The frequency of the male cricket’s
202 Jessie S. Nixon and Fabian Tomaschek

FIGURE 17.3 a) Midsagittal view of a schematic representation of the mouth. 1–2 =

lips, 3–4 = teeth, 5 = tongue tip. The black triangle represents the location
of the tongue body when articulating the respective vowels. b) Effects of
vocal-tract length onto the spectral properties of the vowel space. The
black triangle indicates the human vowel space, and the grey triangles
indicate how vowels produced by longer and shorter vocal tracts are per-
ceived by human ears. Note that frequencies are inverted on both axes to
mirror the orientation of the vowel space with that in the mouth. c) Effects
of the speed of sound. The black triangle represents the human vowel
space, and grey triangles represent what the same articulatory targets
would sound like in atmospheres with faster or slower speeds of sound.

signal varies with the ambient temperature and female crickets adapt their perception
to the temperature-dependent changes (Doherty 1985). By analogy, extra-terrestrial
life forms will also require learning mechanisms sufficient for adapting to (subtle)
changes in the physical properties of their home planet’s atmosphere.
We now consider a scenario in which the extra-terrestrial’s body – and thus their
vocal tract – differs in size from humans. Even with the same vocal apparatus as
 Learning and Adaptation of Communication Systems 203

humans, differences in size can lead to a very different signal. In fact, this is what
we see in the calls of prairie dogs – rodents roughly 30 cm in height when standing –
which indicate the presence of various predators by means of alarm calls. Like
humans, they create calls by means of their masticatory apparatus, manipulating
the acoustic quality of their calls (Slobodchikoff and Placer 2006; Slobodchikoff,
this volume [Chapter 9]). While the spectral characteristics of prairie dog calls
strongly resemble those of vowels, the formants of these sounds consist of higher
frequencies.
We can illustrate the effect of body size, and therefore vocal tract size, by
manipulating the parameter c in equation (1). In contrast to speed of sound, the size
of the vocal tract has very strong effects on the acoustic characteristics of vowels if
we assume c = 350 m/s (Figure 17.3c). Extra-terrestrial life forms whose vocal tract
is significantly shorter than the average human vocal tract, e.g., 5 cm long, would
produce formant frequencies significantly higher than human formant frequencies.
F1 would fall into the range 1,500–2,000 Hz, and F2 would fall into the range of
4,500–6,000 Hz (grey dashed triangle in Figure 17.3c). These frequencies are still
within the range of human perception – but although F1 values are within the range
that humans use for speech, the F2 values are higher than the frequencies used in
any vowels in human speech. Therefore, all the speech sounds would sound strange
to human ears, independently of which vowel type (i.e., vocal tract shape) these
life forms produced. Because of the high F1 and F2 values, the signal would be
most similar to and subjectively sound most like a mix between human [a] and [e]
vowels (for which F1 are located between 500 Hz and 1,000 Hz and F2 are located
between 800 Hz and 2,500 Hz). Life forms which have longer vocal tracts than the
average human would produce vowels with lower formant frequencies (grey dotted
triangle in Figure 17.3c). Supposing a vocal tract of length 30 cm, the lower part
of the vowel space of these beings would overlap with the upper part of the human
vowel space. To produce an [a], humans lower the jaw and move the tongue a little
bit forward. If life forms with a longer vocal tract performed the same articulatory
gesture, the physical signal would be equivalent to (and would sound to humans
like) a human [o]; similarly, an [o] would sound like an [u].
Thus far, we have discussed what vowel-like sounds would sound like if body
size – and therefore the vocal tract size – were different from those of humans. These
assumptions were based on adult speakers. However, the size of the human vocal
tract changes between birth and adult life between roughly 7 cm and 17 cm. Fur-
thermore, female speakers have, on average, shorter vocal tracts than male speak-
ers, yielding higher formant frequencies. If extra-terrestrial life forms undergo a
life cycle similar to life forms on Earth, during which they grow (which is very
probable if we are dealing with biological life forms), and if there are intraspecies
variations in body size, the physical characteristics of their communication signals
will vary. This will be true not only for acoustic signals generated by wind pipes,
but also by other kinematic mechanisms, as well as by electromagnetic mecha-
nisms, i.e., light emitters.
204 Jessie S. Nixon and Fabian Tomaschek

FIGURE 17.4 a) Human audiogram. Frequencies on the x-axis. Perceptual threshold on

y-axis in dB. The grey bar indicates perceptual optimum. b) Atmospheric
opacity, indicating the absorption (y-axis) of electromagnetic frequencies
(x-axis). The grey bar represents the human visual window.
Sources: Figure 17.4a according to Heffner and Heffner (2007), adapted with permission; Figure
17.4b adapted from Atmospheric electromagnetic opacity (2022)

Earth’s atmospheric characteristics and human physiology allow humans to

produce a signal ranging between 50 Hz and 20,000 Hz (Figure 17.4a). Human
pitch ranges from roughly 70–300 Hz, vowels are encoded primarily within 250 Hz
and 2,500 Hz, and consonants such as fricatives use frequencies up to 8,000 Hz.
Inspecting the human audiogram (Figure 17.4a), the obvious conclusion arises that
human hearing capabilities probably co-evolved with our speaking capabilities,
which depend on the physical properties of our atmosphere. The optimal frequen-
cies for a human are located between 250 Hz and 8,000 Hz, exactly the frequency
range in which human acoustic communication is located. More importantly, the
increased sensitivity between 1,000 Hz and 8,000 Hz seems to compensate for the
decrease in spectral amplitude in these frequency ranges (cf. Figure 17.3a–c). We
therefore expect extra-terrestrial life forms to have adapted both production and
perception to the physical characteristics of their home planet’s atmosphere. This
conclusion is not restricted to communication in the acoustic domain.
Like the acoustic signal, electromagnetic communication is highly constrained
by the physical properties of a planet’s atmosphere, i.e., its electromagnetic opacity
(Figure 17.4b). Earth’s atmosphere, for example, provides only two windows for
electromagnetic communication (Bothais 1987; Goody and Yung 1995). The first is
located between wavelengths of 100 nm and 100 µm. Human and most animal vision
evolved to be optimal for a small part of the first window (represented by the grey bar).
 Learning and Adaptation of Communication Systems 205

The second window – between a wavelength of roughly 1 cm and 10 m – allows us

to use radio waves for communication. In the case of Earth’s atmosphere, the opacity
is defined by the ratio of water vapour, oxygen and ozone. Knowing the composition
of an exoplanet’s atmosphere would allow us to predict for which electromagnetic
frequencies potential life on that planet has evolved. Likewise, this knowledge would
allow us to predict whether we have common potential visual ranges.

Implications for the Communication Systems of Extra-Terrestrial

Life Forms
In a case wherein extra-terrestrial life forms evolve in different atmospheric condi-
tions or have different sized bodies, but have similar learning mechanisms as the
species of Earth, we can make certain predictions about the characteristics of their
communication system. First, the system will be highly adaptive under conditions
of uncertainty. Second, the system will become significantly less adaptive with
learning or as uncertainty decreases. Like Earth species, as learning approaches
asymptote, it will become more difficult for these entities to adapt to new informa-
tion, as long as event outcomes remain predictable.
Importantly, the system will downweight or unlearn cues that are not informative for
predicting outcomes. This vital mechanism is what allows the life forms to ignore the
noise in the system and learn to discriminate to a high level of sensitivity the cues that
are important for communication of messages. This same mechanism also has the effect
that cues in the signal that are not detectable with the life form’s sensory system will end
up not being discriminated. The life form will have learned to ignore the variation in the
cues, as in the simulation of perception of extra-terrestrial vowels by human listeners.
Note that events and objects in the world provide outcomes that enable dis-
criminative learning of the speech signal (a “dog” is not a “dock”; Nixon 2020).
Linguistic labels in turn provide outcomes for discrimination learning of events
and objects in the world (a dog is not a cat; Ramscar et al. 2010). Communication
is predictive about the world, and the world is also predictive about communication
events. Each aids and influences the learning of the other.
Using the values obtained above for species that produce speech in the same
way as humans, but have long, 30 cm vocal tracts, we simulate an extra-terrestrial
species learning two vowel categories. Although we expect an unimaginably large
degree of variation is possible in the types of communication signals across extra-
terrestrial species, for the sake of illustration, we consider only tiny deviations
from the characteristics of Earth’s languages.
Note that this analysis assumes that the extra-terrestrial species experiences time
in a linear fashion as on Earth. Events earlier in time can be used to predict upcom-
ing events, but not vice versa (Nixon 2018, 2020; Ramscar et al. 2010). If this is not
the case, we might expect quite different learning mechanisms, such as for example,
Hebbian (Gerstner and Kistler 2002) learning, in which learning occurs through
simple co-occurrence of stimuli, rather than prediction of outcomes from cues.
 206 Jessie S. Nixon and Fabian Tomaschek

155
Perceptual characteristics
human [i] ET [i]
human [u] ET [u]

1.00
117

0.75
Cue density

perception
0.50
78

0.25
40

0.00
1

2700.0 2037.5 1375.0 712.5 50.0

F2 continuum

FIGURE 17.5 Figure illustrates how perceptual characteristics (lines, left y-axis)
develop form different distributions of cues (bars, right y-axis) along a
given continuum (x-axis). Note that the x-axis is inverted to mirror the
vowel space. Light grey bars represent the distribution of F2 cues for
humans, with higher frequencies being interpreted as [i] and lower fre-
quencies as [u]. Dark grey bars represent distribution of F2 for extra-
terrestrial life forms.

Error-driven learning is also most expected when there is a sufficient degree of

uncertainty in the organism’s world to create the necessity of unlearning non-dis-
criminative cues or noise in the signal. Note also that this simulation addresses the
specific situation when the frequency range partially overlaps the human speech
frequency range. Different results would be expected if there were no overlap or if
the communication signal consisted of entirely different cues (such as electromag-
netic, for example, or acoustic cues not familiar to the listener).
Using the Rescorla-Wagner learning equations (Rescorla and Wagner 1972),
a computational formalisation of error-driven, discriminative learning, we simu-
lated the perceptual characteristics of the [i] and [u] vowels. This simulation is
not intended to represent actual learning of speech in the real world. It is a gross
oversimplification; we do not assume that vowel categories are learned with
vowel labels, but rather that speech co-occurs with events and the presence of
various objects and so on, as discussed previously (perhaps as a multimodal ver-
sion of the model presented in Nixon and Tomaschek 2020, 2021, for example).
Our aim with this tiny toy example is to hone in on the specific cues of interest,
in order to highlight the quite dramatic perceptual effects of a minimal change in
conditions.
The equations learn to discriminate the vowels ([i] and [u]) by formant cue
according to the covariance structure in the data. This simulation is based on vowel
 Learning and Adaptation of Communication Systems 207

categories that are produced with the same articulatory gestures as English [i] and
[u]. The first simulation models human learning and is trained on human vowels.
The F2 cue for these vowels clusters in two normal distributions with peaks located
at 800 Hz and 2,100 Hz (Figure 17.5, light grey bars). The second simulation is
trained on the values of extra-terrestrial vowels [i] and [u], which cluster around
400 Hz and 1,500 Hz (Figure 17.5, dark grey bars). In this toy example, the cues
are the spectral frequencies and there are only two possible outcomes, [i] and [u],
for each species. The lines represent the activation level for each vowel for the
humans and for the extra-terrestrials. For each different vowel outcome, activation
is high at different points along the F2 continuum depending on the training.
The dashed-dotted (for [i]) and dotted lines (for [u]) show that the human per-
ception depends on the human cue distribution. Activation decreases towards the
edges of the continuum, resulting in two perceptual categories with a shift from one
vowel to the other located roughly at 1,350 Hz. A similar pattern can be observed
for the simulated extra-terrestrial life forms. However, the activation peaks are
located in lower frequency range, due to the cues in their training. The critical
result is for the human [u] (dotted line). At the points along the F2 continuum that
activate two different vowels [i] and [u] for the extra-terrestrial life form, there is
only activation of [u] for the human listeners. Therefore, both alien vowel catego-
ries are likely to be perceived as [u] – even if perhaps not good instances of [u].
Both of these cases illustrate perception after training in the native communication
systems. If there were sufficient information available to provide error feedback
that this initial perception or prediction was sometimes incorrect, then learning
through prediction error could occur. The listeners might eventually learn to dis-
criminate the new vowels.

Implications for Understanding Extra-Terrestrial Communication

As noted previously, the chances of actually encountering such a communication
system are rather slim. We also remain sceptical of this potential signal yielding
obvious profit. In terms of human encounters with such a communication system,
this would presumably require at least obtaining access to some sort of multimodal
signal in which not only the language itself but also the surrounding events are
available. We agree with Slobodchikoff (this volume [Chapter 9]) that if we had
access to the signal alone, we would have little hope of learning much about the
communication system other than characteristics of the signal itself. One charac-
teristic of words is their linguistic context (“You shall know a word by the com-
pany it keeps”, Firth 1957: 11). Covariance structures between words are modelled
with semantic networks which learn that “cat” and “dog” are more similar than
“cat” and “penguin” due to their use in language (Landauer and Dumais 1997).
However, this only examines structure within the system. It says nothing about the
intentions of the signaller, the relation of the signal to the world – the meaning. If
the communication contained both signal and sufficient information about how it
208 Jessie S. Nixon and Fabian Tomaschek

relates to the world – for example, in some sort of multimodal or even interactive
format – this would provide a better chance of learning something about the world
of the life form and the system they use to communicate about it.3 Knox et al.
(2019) have achieved this to some extent with wild orangutans and Slobodchikoff
and Placer (2006) with prairie dogs. But although these are remarkable feats, the
understanding of the communication systems remains limited. Slobodchikoff pre-
dicts that while prairie dog speech is the most complex communication system yet
discovered (by humans), it may turn out that we have previously simply failed to
notice communication in other animals. Generally speaking, humans often seem to
have a tendency to jump to conclusions about any sort of “others”, whether they
be humans, animals or plants, often with an assumption of superiority. This may
be convenient for allowing us to treat others differently to how we would wish to
be treated ourselves. But this pattern does not seem to bode well for easily under-
standing the communication of others – including extra-terrestrial communication.

Conclusion
Human communication systems are highly adaptable. This adaptability is sup-
ported by learning mechanisms that allow listeners to learn even very subtle cues
when they are important for discrimination. Feedback from prediction error allows
for rapid adaptation to new conditions and ignoring of uninformative cues. These
learning mechanisms are not limited to humans but are pervasive throughout the
animal kingdom. We hypothesise that any complex communication system will
rely on learning in order to adapt to current conditions and communicative needs.
We have discussed some practical implications for the task of interpreting com-
munication signals, as well as preparing materials for extra-terrestrial communica-
tion. Language materials on their own probably have very little information value.
We do not consider language to be made up of sequences of abstract categories,
but instead argue that it emerges through ongoing learning from interaction with
the world and other language users, where different kinds of cues – acoustic, vis-
ual, tactile, olfactory or other sensory information – predict different world-related
events. That is, concepts arise through the experience of language in context.
Experiencing the communication signal in isolation without additional information
about their function has little hope of leading to any understanding of the language.
If the METI project wants to create a common code, rather than broadcasting ran-
dom, unrelated audio and video samples (see Harbour, this volume [Chapter 18]),
it could compile a selection of distributionally structured and contextually depend-
ent information provided in multimodal channels. However, it is doubtful that this
would be sufficient. We have demonstrated that, even when the communication
signal is almost identical – using an acoustic signal, and with identical articulation
to humans – substantial effects on perception emerge from only minor changes
in a single dimension, such body size or atmospheric conditions. As we scale up
to multiple differences and their interactions, the gap is bound to become ever
 Learning and Adaptation of Communication Systems 209

larger – and this is without considering any differences in the concepts being com-
municated. We argue that language represents and emerges from rich, multimodal
experience. If this is the case, then any reduction in richness of experience presum-
ably leads to a reduction in the quality of the linguistic representation. If we return
to Shannon’s conceptualisation of communication as reproduction of a message,
this highlights the challenge. We need to reproduce in ourselves the experience of
someone without knowing their physical environment, their objects, technology,
their culture, history, beliefs, perhaps even their sensory perception. This may be
one of the obstacles that has so far hindered us from gaining a better understanding
of animal communication. Perhaps when we have learned to converse with the var-
ious terrestrial life forms all around us, then we will be ready for extra-terrestrial
communication.

Notes
1 There also exist consciously designed technologies for communication, such as braille
and other writing systems and code languages – see Harbour, this volume (Chapter 18),
for a discussion of writing systems.
2 Though the signal can be transcribed with several alternative letter combinations, the
most common is “SOS”, which has unofficially come to be associated with the phrase
“save our souls”.
3 We leave aside the technical issue of how the technology is made usable to unfamiliar users.

References
Altmann, G. 1980. “Prolegomena to Menzerath’s law.” Glottometrika 2: 1–10.
Arnold, D., F. Tomaschek, K. Sering, M. Ramscar, and R.H. Baayen. 2017. “Words from
spontaneous conversational speech can be recognized with human-like accuracy by
an error-driven learning algorithm that discriminates between meanings straight from
smart acoustic features, bypassing the phoneme as recognition unit.” PLoS One 12 (4).
doi:10.1371/journal.pone.0174623
Atmospheric electromagnetic opacity. 2022. Wikimedia Commons. https://commons.wiki-
media.org/wiki/File:Atmospheric electromagnetic opacity.svg (Accessed 25 October,
2022).
Baayen, R.H., P. Milin, D.F. Ðurđević, P. Hendrix, and M. Marelli. 2011. “An amorphous
model for morphological processing in visual comprehension based on naive discrimina-
tive learning.” Psychological Review 118 (3): 438–481. doi:10.1037/a0023851
Baayen R.H., Shaoul C., Willits J., Ramscar M. 2016. Comprehension without segmenta-
tion: A proof of concept with naive discriminative learning. Language, Cognition and
Neuroscience 31 (1): p106ff. https://doi.org/10.1080/23273798.2015.1065336
Bothais, L. 1987. Radio Wave Propagation. New York: McGraw-Hill Book Company.
Clayards, M., M.K. Tanenhaus, R.N. Aslin, and R.A. Jacobs. 2008. “Perception of
speech reflects optimal use of probabilistic speech cues.” Cognition 108 (3): 804–809.
doi:10.1016/j.cognition.2008.04.004
Doherty, John A. 1985. “Temperature coupling and ‘Trade-Off’ phenomena in the acoustic
communicative system of the Cricket, Gryllus Bimaculatus De Geer (Gryllidae).” Jour-
nal of Experimental Biology 114 (1): 17–35. https://doi.org/10.1242/jeb.114.1.17
210 Jessie S. Nixon and Fabian Tomaschek

Einhorn, H.J., and R.M. Hogarth. 1986. “Judging probable cause.” Psychological Bulletin
99 (1).
Fant, G. 1960. Acoustic Theory of Speech Production. De Gruyter.
Firth, J. 1957. “A synopsis of linguistic theory, 1930–55.” In Studies in Linguistic Analysis.
Special Volume of the Philological Society. Oxford: Blackwell. 1–31.
Fosler-Lussier, Eric, and Nelson Morgan. 1999. “Effects of speaking rate and word fre-
quency on pronunciations in convertional speech.” Speech Communication 29 (2–4):
137–158. http://www.sciencedirect.com/science/article/pii/S0167639399000357.
Fuchs, S., R. Winkler, and P. Perrier. 2008. “Do speakers’ vocal tract geometries shape their
articulatory vowel space?” Proceedings of the 8th International Seminar on Speech Pro-
duction, ISSP’08, Strasbourg: France: 333–336.
Gay, Thomas. 1978. “Effect of speaking rate on vowel formant movements.” The Journal of
the Acoustical Society of America 63 (1): 223–230. doi:10.1121/1.381717
Gerstner, W., and Kistler, W. M. (2002). “Mathematical formulations of Hebbian learning.”
Biological cybernetics, 87 (5), 404–415.
Gluck, M.A., and G.H. Bower. 1988. “From conditioning to category learning: An adap-
tive network model.” Journal of Experimental Psychology: General 117 (3): 227.
doi:10.1037//0096-3445.117.3.227
Heffner, H.E., and R.S. Heffner. 2007. “Hearing ranges of laboratory animals.” Journal of
the American Association for Laboratory Animal Science 46 (1): 20–22.
Hirata, Yukari. 2004. “Effects of speaking rate on the vowel length distinction in Japanese.”
Journal of Phonetics 32 (4): 565–589. https://doi.org/10.1016/j.wocn.2004.02.004
Hull, C.L. 1943. Principles of Behavior. New York: Appleton-Century-Crofts.
Jiao, D., V. Watson, S.G.-J. Wong, K. Gnevsheva, and J.S. Nixon. 2019. “Age estimation
in foreign-accented speech by non-native speakers of English.” Speech Communication
106: 118–126. https://doi.org/10.1016/j.specom.2018.12.005
Kamin, Leon J. 1968. “Attention-like processes in classical conditioning.” In Miami Sym-
posium on the Prediction of Behavior, edited by M.R. Jones, 9–31. Miami: Miami Uni-
versity Press.
Knox, A., J. Markx, E. How, A. Azis, C. Hobaiter, F.J. van Veen, and H. Morrogh-Bernard.
2019. “Gesture use in communication between mothers and offspring in wild orang-
utans (Pongo pygmaeus wurmbii) from the Sabangau Peat-Swamp Forest, Borneo.” In-
ternational Journal of Primatology: 1–24. https://doi.org/10.1007/s10764-019-00095-w
Landauer, T.K., and S.T. Dumais. 1997. “A solution to Plato’s problem: The latent semantic
analysis theory of acquisition, induction, and representation of knowledge.” Psychologi-
cal Review 104 (2): 211. https://doi.org/10.1037/0033-295X.104.2.211
Leemann, Adrian, Volker Dellwo, Marie-José Kolly, and Stephan Schmid. 2012. “Rhythmic
variability in Swiss German dialects.” 6th International Conference on Speech Prosody
2012, Shanghai, May 22–May 25. 607–610. ISCA.
MacLean, D.J. 1966. “Analysis of speech in a Helium-Oxygen mixture under pres-
sure.” The Journal of the Acoustical Society of America 40 (3): 625–627. https://doi.
org/10.1121/1.1910128
McMurray, B., R.N. Aslin, and J.C. Toscano. 2009. “Statistical learning of phonetic catego-
ries: insights from a computational approach.” Developmental Science 12 (3): 369–378.
doi:10.1111/j.1467-7687.2009.00822.x
Magen, H.S. 1997. “The extent of vowel-to-vowel coarticulation in English.” Journal of
Phonetics 25: 187–205. https://doi.org/10.1006/jpho.1996.0041
Maye, J., J.F. Werker, and L. Gerken. 2002. “Infant sensitivity to distributional information can
affect phonetic discrimination.” Cognition 82 (3). doi:10.1016/s0010-0277(01)00157-3
 Learning and Adaptation of Communication Systems 211

Meyer, J. 2004. “Bioacoustics of human whistled languages: An alternative approach to

the cognitive processes of language.” Anais da Academia Brasileira de Ciências 76
(2): 406–412. doi:10.1590/S0001-37652004000200033NASA. n.d. www.grc.nasa.gov/
WWW/K-12/rocket/machu.html.
Michaud, M. 2011. Seeking contact: The relevance of human history. In Communication
with Extra-Terrestrial Intelligence, edited by D.A. Vakoch. Albany: State University of
New York Press.
NASA. n.d. https://www.grc.nasa.gov/WWW/K-12/rocket/machu.html.
Nixon, J.S. 2018. “Effective acoustic cue learning is not just statistical, it is discriminative.”
Interspeech 2018–19th Annual Conference of the International Speech Communication
Association, 1447–1451. Hyderabad, India, September.
Nixon, J.S. 2020. “Of mice and men: Speech acquisition as discriminative error-driven
learning, not just statistical tracking.” Cognition 197: 104081. https://doi.org/10.1016/j.
cognition.2019.104081
Nixon, J.S., and C.T. Best. 2018. “Acoustic cue variability affects eye movement behaviour
during non-native speech perception.” Proceedings of the 9 International Conference on
Speech Prosody, 493–497. Poznan, Poland, June.
Nixon, J.S., N. Boll-Avetisyan, T.O. Lentz, S. van Ommen, B. Keij, Ç. Çöltekin, L. Liu,
and J. van Rij. 2018. “Short-term exposure enhances perception of both between- and
within-category acoustic information.” Proceedings of the 9th International Conference
on Speech Prosody, 114–118.
Nixon, J.S., S. Poelstra, and J. van Rij. 2022. “Does error-driven learning occur in the
absence of cues? Examination of the effects of updating connection weights to absent
cues.” In Proceedings of the 44th Annual Meeting of the Cognitive Science Society, ed-
ited by J. Culbertson, A. Perfors, H. Rabagliati, and V. Ramenzoni, 2590–2597. Merced:
Open Access Publications from the University of California (UC Merced).
Nixon, J.S., and F. Tomaschek. 2020. “Learning from the acoustic signal: Error-driven learn-
ing of low-level acoustics discriminates vowel and consonant pairs.” In Proceedings of
the 42nd Annual Meeting of the Cognitive Science Society, edited by S. Denison., M.
Mack, Y. Xu, and B.C. Armstrong, 585–591. Merced: Open Access Publications from
the University of California (UC Merced).
Nixon, J.S., and F. Tomaschek. 2021. “Prediction and error in early infant speech learn-
ing: A speech acquisition model.” Cognition 212: 104697. https://doi.org/10.1016/j.
cognition.2021.104697
Nixon, J.S., and F. Tomaschek. 2023a. “Does speech comprehension require phonemes?”
Forthcoming in The Handbook of Usage-Based Linguistics, edited by M. Diaz-Campos
& S. Balasch. Oxford: Wiley Blackwell Publishing.
Nixon, J.S., and F. Tomaschek. 2023b. “Introduction to the special issue emergence of
speech and language from prediction error: error-driven language models.” Manuscript
submitted to Language, Cognition and Neuroscience.
Nixon, J.S., J. van Rij, P. Mok, R.H. Baayen, and Y. Chen. 2016. “The temporal dynamics of per-
ceptual uncertainty: Eye movement evidence from Cantonese segment and tone perception.”
Journal of Memory and Language 90: 103–125. https://doi.org/10.1016/j.jml.2016.03.005
Nyst, V. 2015. “The sign language situation in Mali.” Sign Language Studies, 15 (2): 126–
150. www.jstor.org/stable/26190976
Öhman, S.E.G. 1966. “Coarticulation in VCV utterances: Spectrographic measure-
ments.” Journal of the Acoustical Society of America 39 (151): 151–168. https://doi.
org/10.1121/1.1909864
Ramscar, M. 2019. Source Codes in Human Communication. https://psyarxiv.com/e3hps.
212 Jessie S. Nixon and Fabian Tomaschek

Ramscar, M., M. Dye, and J. Klein. 2013. “Children value informativity over logic in word
learning.” Psychological Science 24: 1017–1023.
Ramscar, M., D. Yarlett, M. Dye, K. Denny, and K. Thorpe. 2010. “The effects of feature-
label-order and their implications for symbolic learning.” Cognitive Science 34 (6): 909–
957. doi:10.1111/j.1551-6709.2009.01092.x
Rattray, R.S. 1923. “The drum language of West Africa.” African Affairs 22(87): 226–236.
https://doi.org/10.1093/oxfordjournals.afraf.a100065
Reber, A.S. 1989. “Implicit learning and tacit knowledge.” Journal of Experimental Psy-
chology: General 118 (3): 219–235.
Rescorla, R., and A. Wagner. 1972. “A theory of pavlovian conditioning: Variations in the ef-
fectiveness of reinforcement and nonreinforcement.” In Classical Conditioning II: Cur-
rent Research and Theory, edited by A.H. Black and W.F. Prokasy, 64–69. New York:
Appleton Century Crofts.
Shannon, C.E. 1948. “A mathematical theory of communication.” Bell System Technical
Journal 27 (3): 379–423.
Shanks, D.R. 1986. “Selective attribution and the judgment of causality.” Learning and
Motivation, 17 (4): 311–334.
Slobodchikoff, C.N., and J. Placer. 2006. “Acoustic structures in the alarm calls of Gun-
nison’s prairie dogs.” The Journal of the Acoustical Society of America 119 (5): 3153–
3160. https://doi.org/10.1121/1.2185489
Soto, F.A., and E.A. Wasserman. 2010. “Error-driven learning in visual categorization and
object recognition: A common-elements model.” Psychological Review 117 (2): 349.
doi:10.1037/a0018695
Spence, K.W. 1956. Behavior Theory and Conditioning. New Haven, CT: Yale University
Press.
Tomaschek, F., D. Arnold, Franziska Bröker, and R.H.R. Baayen. 2018a. “Lexical frequency
co-determines the speed-curvature relation in articulation.” Journal of Phonetics 68:
103–116. https://doi.org/10.1016/j.wocn.2018.02.003
Tomaschek, F., and A. Leemann. 2018. “The size of the tongue movement area affects the
temporal coordination of consonants and vowels – A proof of concept on investigating
speech rhythm.” JASA-EL 144 (5). https://doi.org/10.1121/1.5070139
Tomaschek, F., I. Plag, M. Ernestus, and R.H. Baayen. 2021. “Phonetic effects of mor-
phology and context: Modeling the duration of word-final S in English with naïve
discriminative learning.” Journal of Linguistics 57 (1), 1–39. https://doi.org/10.1017/
S0022226719000203
Tomaschek, F., and M. Ramscar. (2022). “Understanding the phonetic characteristics of
speech under uncertainty – implications of the representation of linguistic knowledge in
learning and processing.” Frontiers in Psychology: 1833. doi:10.3389/fpsyg.2022.754395
Tomaschek, F., B.V. Tucker, R.H. Baayen, and M. Fasiolo. 2018b. “Practice makes perfect:
The consequences of lexical proficiency for articulation.” Linguistic Vanguard 4 (s2).
doi:10.1515/lingvan-2017-0018
Tomaschek, F., B.V. Tucker, M. Wieling, and R.H. Baayen. 2014. “Vowel articulation af-
fected by word frequency.” Proceedings of the 10th ISSP, 425–428. Cologne.
Tomaschek, F., M. Wieling, D. Arnold, and R.H. Baayen. 2013. “Word frequency, vowel
length and vowel quality in speech production: An EMA study of the importance of
experience.” Proceedings of the Interspeech. Lyon.
Van Hamme, L.J., and E.A. Wasserman. 1994. “Cue competition in causality judgments:
The role of nonpresentation of compound stimulus elements.” Learning and Motivation
25 (2): 127–151.
 Learning and Adaptation of Communication Systems 213

Waelti, P., Dickinson, A., and Schultz, W. 2001. “Dopamine responses comply with
basic assumptions of formal learning theory.” Nature 412: 43–48. https://doi.
org/10.1038/35083500
Weber, F., M.W. Woolridge, and J.D. Baum. 1986. “An ultrasonographic study of the or-
ganization of sucking and swallowing by newborn infants.” Developmental Medicine &
Child Neurology 28 (1): 19–24.
Weirich, M., and S. Fuchs. 2006. “Palatal morphology can influence speaker-specific reali-
zations of phonemic contrasts.” Journal of Speech, Language, and Hearing Research 56:
1894–1908. doi:10.1044/1092-4388(2013/12-0217)
Widrow, G., and M.E. Hoff. 1960. Adaptive Switching Circuits. IRE Western Electron. Show
Convention, Convention Record Part 4, 96–104.
Winkler, R., S. Fuchs, and P. Perrier. 2006. “The relation between differences in vocal
tract geometry and articulatory control strategies in the production of French vowels:
Evidence from MRI and modeling,” 8th International Seminar on Speech Production,
ISSP’08, Strasbourg: France, 509–516.
Wittgenstein, L. (1953). Philosophical Investigations [Philosophische Untersuchungen].
Yashiro, K., and K. Takada. 2005. “Model-based analysis of jaw-movement kinematics us-
ing jerk-optimal criterion: Simulation of human chewing cycles.” Journal of Electro-
myography and Kinesiology 15 (5): 516–526. doi:10.1016/j.jelekin.2004.11.005
Yoshida, K.A., F. Pons, J. Maye, and J.F. Werker. 2010. “Distributional phonetic learning at
10 months of age.” Infancy 15 (4). doi:10.1111/j.1532-7078.2009.00024.x
18
WRITING SYSTEMS AND METI
Off-the-Shelf Encoding of Human Physiology,
Language, Cognition, and Culture

Daniel Harbour

Introduction
Imagine that we receive an alien message tomorrow and that, thanks to its mostly
mathematico-scientific form and content, it turns out to be readily decipherable.
Three questions would arise from the next day’s headline:
Intelligent Life Found
Found where? What life? What intelligence?
Our attempts to message extraterrestrial intelligence (METI) do not address
these questions equally for potential alien audiences. Origin and biology were rep-
resented in the earliest deliberate, content-dense transmissions, the Pioneer Plaques
and the Arecibo transmission of 1972‒1974. For reasons of space, these said little
about the nature of our intelligence. But these size restrictions no longer constrain
us and, though human biochemistry is easier to portray than human thought, we
should still tackle the task of representing it systematically. Silence is neither desir-
able nor – if we reflect on previous messages – possible. If ET-planned METI goes
through anything like the thought process described in this chapter, the common
SETI (Search for Extraterrestrial Intelligence) assumption that interstellar mes-
sages will read like resumés of scientocracies is likely false. Just as the form of
interstellar messages must be finely balanced – between redundancy, which makes
them noticeably artificial, and decodability – so, too, must their content be bal-
anced between the universal and the particular. The mathematico-scientific com-
mon ground between any two METI-capable civilizations is not a communicative
end in itself, but a tool by which to encode information on the kind of res cogitans
behind any message.
The human accent in human METI should be deliberate and carefully planned.
My main purpose is to argue that writing systems are excellent, multifaceted

DOI: 10.4324/9781003352174-18
 Writing Systems and METI 215

indicators of human language, cognition, and culture. Moreover, in the case of Vis-
ible Speech in particular (Bell 1867), they offer a natural bridge from the scientific
content of previous METI to future, humanized messages that take full advantage
of advances in understanding what is a reasonable size of message to transmit.
I begin with overviews of potential message size in the next section (What Can
We Send?) and past messages the section after (What Have We Sent?). Though
past messages are substantially smaller than what we could send, these overviews
nonetheless converge on a philosophy in which mathematics and science consti-
tute stepping stones to broader communication. This spotlights the risk of being
insufficiently or excessively human-centered as pitfalls to avoid in larger-message
design (as discussed in section Potential Pitfalls). Instead, a message design that
transitions gradually from science to humanities, with writing systems as part of
the bridge, offers several design advantages (as described in the section Human
Writing Systems).

What Can We Send?

The first question facing METI is what size of message to send into space. A variety
of research converges on the conclusion that we can send substantial quantities of
information in the reasonable hope that it will be salient and, if found, complete.
Taking into account galactic background radiation and the interstellar medium
of dust, neutral gas, and plasma, Harp et al. (2011) argue for a “twice-sent signal”
format, like a musical canon. Transmitters send messages in two staggered but
overlapping installments, so that any dispersion the two versions undergo is identi-
cal. The result is distortion-resistant, information-rich, and salient against galactic
background radiation. Harp et al.’s proposal contrasts with many previous SETI
efforts, which focused on sine waves. These are salient because artificial but are
also informationally “monochromatic.” How much color can our messages bear?
Reviewing the technical limits on interstellar messages, Shostak (2011) argues
for sending an encyclopedia-sized body of information as a feasible balance of
content and form. The round-trip message time just within the Milky Way (where
Dick 1996 estimates up to 100,000 civilizations may exist) is measured in (tens of)
hundreds of light years. This amounts to one-way communication from the point
of view of individual senders and their cultures and so favors sending as much as
possible at once. However, the fact that no one is likely to be listening from the
start of our transmission places a limit on how much of a long transmission we can
expect to see received. This favors frequently repeated shorter messages instead.
As a happy medium, Shostak suggests that, using 1 micron of infrared with one bit
per pulse for a total duration of ~0.1 seconds, we could ping a million star systems
once a day with the equivalent of the Encyclopedia Britannica to each.
Benford et al. (2011) reach a similar conclusion about broadcast length and
repetition by considering cost-optimization of interstellar beacons. The design of
galactic-scale beacons they arrive at consists of narrow “searchlight” beams visible
216 Daniel Harbour

for short periods in an alien observer’s sky. Consequently, the transmission strat-
egy from Earth should be a rapid scan of the galactic plane. The result would be
infrequent pulses of a few seconds from the receiver’s point of view, recurring over
periods of a month to a year.

What Have We Sent?

An encyclopedia is far larger than any message, physical or digital, yet sent from
Earth. Nonetheless, it is worth reviewing previous messages, as their careful con-
struction provides a useful guide for larger formats.
Pioneers 10 and 11, launched in 1972 and 1973, respectively, bear plaques
depicting a man and woman and our solar system, with the position of Earth high-
lighted, alongside information on the hydrogen atom and pulsars intended to serve
as measures of distance and time. The Golden Discs aboard Voyagers 1 and 2 carry
sounds and images of humans, animals, and the planet. The recordings include
a heartbeat, breath, birth, and speech. The images progress from mathematics,
physics, and astronomy to biochemistry and human biology, and then to human
activities like breastfeeding, eating, learning, building, driving, and flying. Both
therefore embody the high-level ambition of using science as a basis for commu-
nicating about humanity.
The potential for such concrete communication is very limited however. Even
traveling at 15‒17 km/s (Cook and Brown 2010), the Voyager probes are lumber-
ing along compared to the speed-of-light travel of radio and laser messages. They
are also far less repeatable. The difference is that of a messenger on horseback
versus email spam on a galactic scale.
The earliest deliberate high-powered transmission of information into deep
space was the Arecibo radio message (The Staff at the National Astronomy and
Ionosphere Center 1975). Formatted as a semiprime number of binary digits, it
unfurls into a 23 × 73 grid representing numbers, then chemical elements, then
compounds critical to life, as well as sketches of DNA, a person, the solar system
(highlighting Earth), and the Arecibo transmitter.
The Arecibo message was incorporated into the Cosmic Calls of 1999 and 2003
alongside further messages. One, the Bilingual Image Glossary (BIG), compris-
ing twelve binary drawings subtitled in Russian and English, was mostly about
humans, but also covered the sun, Earth, and our galaxy. Another was the metic-
ulously planned Dumas-Dutil message (Dumas 2003). Based on Freudenthal’s
(1960) METI-purposed language, Lincos, its twenty-three pages represent a sub-
stantially expanded implementation of the ideas and information in past messages.
Like Arecibo, it contains a range of mathematical, physical, chemical, and astro-
nomical information but adds geochemical and geological data, too, and it encodes
a modified version of the Pioneer Plaque humans (with the woman no longer pas-
sive), together with information about our physical characteristics and biochemical
composition. It concludes, like Arecibo, with information about the transmitter, but
 Writing Systems and METI 217

then adds a page of questions for alien recipients. (The ability to ask questions is a
design feature of Lincos.)
In sum, the composers of past messages have viewed the status of mathematics
and science as a universal language in a very particular light. It provides a means
of talking about humanity in scientific terms. But their messages go further. The
Cosmic Calls follow Freudenthal in asking questions, which reveals a most human
fact about us (Bromberger 1992): that we recognize, and push against, the limits
to our knowledge. BIG goes so far as to hint at the structure of human language.
Assuming that the subtitles are identified as writing systems, they are clearly dis-
tinct (each has letters the other lacks) but related (shared letters), and the word
lengths vary unpredictably (meaning, in familiar terms, that they correspond to
different languages, rather than one language written in different ways), and even
in such a short sample, it may be possible to conclude that the writing system is not
based on inherent meanings of the characters and is therefore more likely alpha-
betic than hieroglyphic.

Potential Pitfalls
Every message has a subtext. Incomplete surveys, for example, provide an inad-
vertent measure of the respondents’ conscientiousness (Hedengren 2013). If alien
curiosity is at all like ours, any message we send will be pored over from multiple
perspectives and – even before its text is exhausted – its subtext will be mined. We
should, therefore, reflect on the potential subtext carefully. Three past proposals for
long-message METI are problematic in this regard, as discussed in the following
subsections: messages that are excessively science-based, artificial languages that
are inadvertently (non)human, and messages consisting of off-the-shelf encyclope-
dias. All risk sending the wrong message, figuratively speaking.

Science Primers
It is axiomatic that METI-ready civilizations must be reasonably advanced in
mathematics and science. It follows, as per the next subsection, that METI should
use mathematics and science to devise self-decoding messages (thereby avoiding
the difficulties that Slobodchikoff, this volume [Chapter 9], discusses for under-
standing prairie dog communication). It does not follow, though, that self-decoding
messages (ours or others) should be primarily mathematico-scientific (Finney and
Bentley 2014).
Earth has enjoyed only a few centuries of high-intensity scientific research and
only a few decades of SETI. This makes us juveniles in scientific and SETI terms. It
is conceivable that there are “adult” civilizations, tens of millennia more advanced
than ours, engaged in METI. If they send us a decodable primer on their mathemat-
ics and science, it might be useful. Maybe cold fusion is a far harder problem to
crack than we imagine. Maybe adult civilizations have seen other juvenile ones
218 Daniel Harbour

self-destruct from battles over, or overuse of, nonrenewable energy sources. If so,
cold fusion construction manuals would make worthwhile messages.
However, purely mathematico-scientific communication is just as likely to be
frustrating. Mathematics is as much about why as what. Knowing the answer to
the P versus NP problem would be useful, but not as useful as knowing why the
answer holds – just as a diagram showing that cold fusion reactors are possible is
far less useful than knowing how to build one. Moreover, a purely scientific mes-
sage would leave tantalizing questions of subtext. Is the information being sent in
altruism, in arrogance, under duress, in evangelism, or in the shadow of extinction?
These questions are more acute for our own METI. A primer on (scientifically
juvenile) human mathematics and science is unlikely to educate its recipients. By
chance, it might be picked up by a civilization that had also only just attained the
capacity for interstellar communication, or by one less advanced in some areas than
we are (maybe there’s a planet of Kantians who have never explored non-Euclid-
ean geometry). But, if neither of those scenarios holds, our recipients are likely
to be more knowledgeable than us. Our message would be the epistemological
equivalent of a gift of children’s clothes several sizes too small. Its subtext might
be taken as arrogance or naïveté.

Artificial Languages With Human Accents

A second fact inherent to science primers makes them an unattractive option. Con-
sider Dumas’s sketch for an expansion of the Dumas-Dutil message, where he for-
mulates human procreation in mathematical notation (Dumas 2011: 410):

M + W ⇒ M + W + M or M + W + W

Two facts are noteworthy here. (Artificial languages are discussed further in
Granger et al., this volume [Chapter 8], and Sperlich, this volume [Chapter 14].)
On the semantic side, the biological statement uses “+” in a nonmathematical
sense. This introduces ambiguity into the artificial language. Ironically, though,
three sentences after the preceding equation, Dumas criticizes human languages
for being “too imprecise and ambiguous.” Given his oversight, the natural ques-
tion is: Is all ambiguity problematic? The current ambiguity is an instance of what
Wittgenstein (1953: 32) calls, aptly enough, “family resemblance.” This is a recur-
rent trait of human language and, apparently, tolerable even in artificial languages.
On the syntactic side, note that “+” occurs between what it connects, M + W,
rather than before, as in the Polish notation, + M W, or after. Although common
in computer science, the Polish order is unattested in natural languages (including
Polish). Instead, both connective-medial and connective-final orders are found
(with the latter restricted to head-final languages and rare even there; Haspelmath
2000; Stassen 2003; Zwart 2005). It would be rash for aliens to assume that
this connective-medial syntax was chosen because it reflects the bulk of human
 Writing Systems and METI 219

languages, but, at the same time, the alien linguist may well be unable to help but
wonder.
Rather than merely accepting such ambiguities and syntactic choices, we might
use them productively. We could adopt the same sign for both mathematical addi-
tion and biological union (or for logical conjunction) and distinguish them by their
syntax: medial for one, final for the other. This deliberate, contextually resolved
symbolic ambiguity would give a window, albeit small, into our conceptual net-
works and the division of labor between human syntax and semantics. The idea
that we can excise our humanness even from our artificial languages blinds us to an
opportunity to inform and engage.

Off-the-Shelf Encyclopedias
Heidmann (1997: 202) suggests that it would be enough send an encyclopedia,
like the Britannica or Universalis, in order to tell aliens about ourselves. He has
sound reasons: encyclopedias are off-the-shelf, expansive, transmissible, poten-
tially decodable text-image composites. However, if any one transmission is an
intergalactic message in a bottle, then his suggestion amounts to building a bespoke
interstellar bottle factory and then filling the bottles with whatever newspapers we
have on hand. (See also Wells-Jensen, this volume [Chapter 13], on the possibility
of sightless aliens.)
Encyclopedias vary so widely between countries that no current one could com-
mand universal consent – though maybe we could lessen this problem sufficiently
by sending abridged encyclopedias in, say, Arabic, Chinese, English, and Hindi,
possibly with smaller documents in less widely spoken languages from elsewhere,
such as Cherokee and Fijian.1 Whatever our choice, encyclopedias would create
for aliens many of the frustrations that we face with regard to antiquity. We know
Sappho almost only through fragmentary quotations by other authors, but this is
not what we should aim for in transmitting Shōnagon, Rumi, or Shakespeare.2 We
know how to interpret Anatolian logograms but only partly how to pronounce them
(Weeden 2014). This is, again, something we should avoid for our message (and
its scraps of Shakespeare). The issue is one of subtext. We should consider how
it reflects on us if we send aliens something curated for humans – in particular,
for humans of a particular cultural background. What kind of conversation have
we begun if our opening gambit is us talking to ourselves (cf, Hobaiter et al., this
volume [Chapter 3])?

Human Writing Systems

Language is key to our intelligence and a defining property of humanity. A true
portrait of human cognition must, therefore, move beyond the cosmic pidgin of
artificial languages and into the realm of natural language. Consistent with past
METI, mathematics is an obvious stepping stone: proofs can be presented both
220 Daniel Harbour

as chains of symbolic statements and as arguments in natural language, yielding a

Rosetta Stone from the universal language of mathematics to one or more human
languages; and mathematics lends itself to nonmathematical notions, such as aes-
thetics, which are highly human.3
But even more basic than content is the question of encoding. Obviously, we can
continue solely with binary. However, it bears no relation to how human languages
are written, let alone spoken or signed, and so veils our report on humanness with a
nonhuman overlay. Instead, I suggest that we take seriously the use of human writ-
ing systems themselves, alongside binary encoding.
Writing systems are the product of our cognitive structure, cultural evolution,
and environmental interaction. Lying at the confluence of such major forces makes
this “artifactual communication” (as Berea, this volume [Chapter 7], calls it) inter-
esting in and of itself. A simple illustration is Morin’s (2018) finding that symmetry
around the vertical axis is much more frequent than symmetry around the horizon-
tal axis (As are more common that Bs) across characters of the world’s alphabets
and syllabaries. Morin ties this to adaptation to our environment: we encounter
far more A-type symmetry (in people, animals, trees) than B-type (reflections in
lakes).
In the first subsection (The Evolution of Writing), I briefly sketch the devel-
opment of alphabets from pictograms emphasizing how crucially it hinges on
facts connected to human cognition. In Bell’s Visible Speech, the connections go
further: the structure of the characters reflects the physiology of speech produc-
tion (see subsection Visible Speech)). This means that Visible Speech could offer
a natural transition from the scientific content of past METI to a more culture-
based encyclopedia in human language (see subsection Visible Speech and Past
METI).

The Evolution of Writing

Writing has been invented ex nihilo only four (or five) times in human history,
approximately contemporaneously in Sumer and Egypt, in Mesoamerica (and
probably South America) some three millennia later, and in China about half-
way between. Despite differences of time and of the grammatical character of the
inventors’ languages, the four (partially) deciphered systems are very similar at the
abstract level.
All are centered on the pictogram; that is, a character that pictures its meaning
(e.g., a picture of tree to mean “tree”). Pictures taken literally extend readily to
ideograms, pictures taken figuratively (e.g., a picture of a tree to mean “wood”). To
disambiguate polysemy (such as “tree/wood”), signs acquired a third function: they
were divorced from their meanings and used purely for their sound. These could be
standalone uses (permitting a sign read as wood to be used for, say, the difficult-to-
depict modal “would”). Or they could be combined to create complex signs, where
one part gives the meaning and another the sound (e.g., a compound sheep-wood
 Writing Systems and METI 221

sign might mean “wool,” based on the meaning “sheep” and the sound wood). (See
Harbour 2019a for detailed Sumerian/Chinese comparison.)
The fact that four peoples with unrelated cultures and languages devised a com-
mon typology of logograms strongly suggests that the solution reflects a shared
intellect as much as a shared problem (one that might be shared with ETI; Roberts,
et al., this volume [Chapter 15]). The parallels run deeper, however, and continue
into descendant systems.
The extent to which the four writing systems used phonetic signs indepen-
dently (“wood” for “would”) versus dependently (in compounds like sheep-wood
for “wool”) correlates with their grammatical structure. Sumerian, Egyptian, and
Mayan – languages in which nouns and verbs take rich inflection (e.g., agreement,
aspect, case) – required speakers to write many syllables that lacked (obvious)
meanings and therefore did not lend themselves to meaning-based depiction. Their
writing used many phonetic signs independently. Baxter and Sagart (2014) recon-
struct Old Chinese as having a rich array of derivational morphemes, but ones
that preserve syllable boundaries and hence meaningfulness. This correlates with a
preference for compounding use of phonetic signs.
Later developments of these systems, especially by unrelated languages, indi-
cate that grammar drives their differences (Harbour 2019b). The Chinese system,
for instance, was adapted to write Japanese, Korean, Vietnamese, and Zhuang. Jap-
anese and Korean require substantial (syllable-altering) inflection, particularly for
verbs. Both developed standalone phonetic signs (and Japanese went so far as to
devise separate syllabaries, reserving Chinese characters for meaning-based use).
Strikingly, then, these languages reinvented the Sumerian-Egyptian-Mayan sys-
tem. Vietnamese and Zhuang, by contrast, have syllable-preserving morphology,
like Modern Chinese, and their speakers adapted the Chinese system by redesign-
ing the phonetic components of compound characters.
This should interest aliens in multiple ways. First, the existence of the Chi-
nese writing system in the present day informs both about the current state of the
language and the linguistic structure of its antecedents, as well as the unbroken
chain of transmission between them. The reconstruction of the language, culture,
and thought processes of a people from its writing system has been termed “script
archaeology” by Jaritz (1967). Analogous to explaining the silent k in knife by
appealing to earlier stages of the language, these patterns are pervasive and fasci-
nating in logography. It is culturally and cognitively revealing that Chinese “write”
(書) incorporates “brush” (聿) and that “rest” depicts a person by a tree (休) and
“companion” a person by a fire (伙). An encyclopedia in part in Chinese would
provide recipients with the seemingly impossible: an archaeological site transmit-
ted across space (cf, Nixon and Tomaschek’s call, this volume [Chapter 17], for
transmissible environments based on discriminative learning).
Second, as writing systems become more phonetic, the phonetic unit they gen-
erally invest in is the syllable, most frequently a vowel (V) or a consonant plus
vowel (CV). Even with modest consonant and vowel inventories, syllabaries tend
222 Daniel Harbour

to avoid representing the likes of CVC, CCV, and CCVC directly. Instead, these are
represented via special conventions, such as Akkadian ba-ag for bag (extra vowel
pre-supports the last consonant), Mayan yo-po for yop (extra vowel post-supports
the last consonant), or Linear B to-ro for tro (extra vowel post-supports the first
consonant).
The frequency of syllabaries (whether from evolution, as in Japanese, or from
reinvention, as in Cherokee) may reflect neurological hardwiring. Luo and Poeppel
(2007: 1001) propose that the syllable is “a computational primitive for the [neu-
rological] representation of spoken language,” in light of their finding that syllable
timing coincides with multiple endogenous brain rhythms. Giraud and Poeppel
(2012: 511) claim that it is “foundational in speech and language processing, ‘pack-
aging’ incoming information into units of the appropriate temporal granularity.”
Syllabaries may therefore reveal deep facts about our neurology, of potential inter-
est to alien readers if appropriately contextualized.
Third, against this background, alphabets require special comment. In alpha-
bets, primitive symbols represent a vowel or a consonant, but not both.4 Despite
the ubiquity of alphabets and their current global dominance, the conditions that
led to alphabetic writing were extremely peculiar. Vowels are crucial to meaning in
all of the world’s languages. However, the Afro-Asiatic family (including Ancient
Egyptian, Phoenician, and Hebrew) confines that dependence mostly to grammati-
cal morphemes. Vowel omission, then, amounts to morpheme underrepresentation,
disguising the difference between noun and verb (prat “a detail,” peret “to detail”),
or active and passive (peret “to detail,” porat “to be detailed”) (Harbour 2021).
This information is often recoverable from context and, when not, simply places
Afro-Asiatic languages on a par with the many languages that lack such morphol-
ogy to begin with (cf., English noun/verb detail).
Given the restricted role of vowels, when Ancient Egyptian developed stan-
dalone phonetic writing, it arrived at a vowelless syllabary – that is, a set of conso-
nants. This was adapted to become the Proto-Semitic abjad (consonantal alphabet),
which was then carried north to write various Semitic languages. Though some of
these came to sporadically indicate vowels (Hebrew did, Phoenician did not), it
was only at the point of handover to non-Afro-Asiatic languages, where grammar
did not cue vowels, that vowel writing markedly increased (as occurred in Brahmi,
Greek, Sogdian, Yiddish, etc.; Harbour 2021).
Thus, full alphabets arose from consonantal alphabets, which arose from an
extremely odd quirk of grammatical design in one of the early written languages.
Afro-Asiatic languages that adopted writing systems with vowels, such as Akka-
dian and Maltese, retained them. There was, then, a narrow window of opportunity,
when writing was young, for a consonantal alphabet to arise just in one corner of
the globe. The aversion that other grammatical types have to vowelless writing is
manifest in the emergence of syllabaries everywhere else and by the almost imme-
diate development of obligatory vowel marking once a consonantal alphabet is
adopted beyond Afro-Asiatic.
 Writing Systems and METI 223

Writing for aliens, we should therefore be eager to display our alphabet. If the
aliens are anything like us, they may not have one.

Visible Speech
No matter how we choose to represent linguistic content, we still have to avoid the
Anatolian problem. That is, we need to indicate pronunciation, on which so much
of linguistic and cultural interest depends and which is so fundamentally tied to our
physiology.
A ready-made solution for this is the Visible Speech alphabet invented by Alex-
ander Melville Bell and described in Bell (1867), when it had already been in use
for several years (an “engineered writing system” comparable to the “engineered
languages” of Sperlich, this volume [Chapter 14]). A principle application was
teaching speech to the deaf. However, it also had uptake among linguists because
it enabled “the writing of all languages in one alphabet,” as per the book’s subtitle.
It was able to register subtle differences in accents and could even represent infant
and animal calls. An Athenæum editorial (cited in full in Bell’s book) describes
how Bell would transcribe rare accents or languages known to audience members,
which his sons (one of them, Alexander Graham Bell), who had been waiting out-
side, would then read back, to the audience’s amazement.
The genius of Visible Speech is that its letters come as close as practicable
to pictograms of sound. They are schematics of the relevant actions of the vocal
tract. For the consonants, the basic graphic unit is a cup. Its orientation indicates
the part of the tongue used to produce the sound (Figure 18.1), with diacritics to
indicate articulatory refinements. For instance, the tongue tip symbol  may co-
occur with blockage of airfow,  t, with blockage plus vocal chord vibration,  d,
and with partial blockage and nasalization,  n. For vowels, the basic symbol is an
elongated version of the voicing diacritic for consonants. It is annotated at the top,
bottom, or both for high, low, and mid vowels respectively, and on the left, right,
or both for back, front, or central. All letters are formed from a set of ten radicals
(graphic primitives), with eighteen further modifiers (for, e.g., clicked consonants
and tone-bearing vowels).
As an example, consider  “universal” as it appears on the front
cover of Bell (1867). The first symbol,  y, indicates the body of the tongue (cup
orientation) and vocal chord vibration (bar). Compare that to the third last symbol,
 s, which engages the same articulator (cup orientation) but without vocal chord
vibration (no bar; cf,  z) and with air channeled through a convex groove (curls
either side of aperture). Similarly, compare the third and final symbols,  n and 
l. These both indicate the tongue tip (cup orientation) and vocal chords (bar). They
differ in that air is diverted via the nose for  n (nose-like squiggle) and around the
tongue for  l (indentation in cup middle). The vowels  u and  i indicate from
the high position of their curves that they are produced at the top of the mouth,
at the back and front respectively hence their orientation; they also differ in that
224 Daniel Harbour

FIGURE 18.1 The vocal tract and the basis of orientation of the cup symbol for con-
sonants –  tongue root,  tongue body,  tongue tip,  lips – and the
diacritics for nasalization and vocal chord vibration, based on nose and
glottis shape.
Source: Based on Bell 1872.

u is round (barred) and tense (bunched), whereas i is unrounded (no bar) and lax
(unfurled).  i contrasts usefully with . It, too, is also unrounded and lax, but, by
having these indications at the bottom, it shows that it is a low vowel (that of at, a
sound rarely encountered in the suffix -al in modern British accents).
Historically, Visible Speech lost ground to the International Phonetic Alphabet
(IPA) of 1888, despite prominent advocates like Henry Sweet, who would later
become president of the Philological Society. The system called for new types to
be cast, making it more costly than the IPA, which extended the Latin alphabet.
However, the viability of Bell’s ideas is underlined by the success of the Korean
writing system, Hangul (Kim-Renaud 1998). Invented by Sejong the Great in the
 Writing Systems and METI 225

15th century, it also draws on articulator shape for its character design (though not
as completely as does Visible Speech) and – despite bureaucratic resistance – it
displaced Chinese to foster an indigenous literacy that continues to this day.
Clearly, there is more to human speech sounds than Bell’s schematic characters
show. This is in part an advantage (cf, Pepperberg, this volume [Chapter 6]): writ-
ing highlights the salient and dampens the noise in spoken (and signed) language.
Ultimately, spectrograms would need to be included for a subset of sounds to allow
for the rest to be “deduced.” Once in place, though, Visible Speech could transcribe
words of a multilingual encyclopedia in which more substantive content would
be presented. This would circumvent the opacity of English spelling, the miss-
ing vowels of Arabic, and the unpredictability of Chinese characters, and it would
facilitate fuller appreciation of the languages and materials sent.

Visible Speech and Past METI

The continuity of Visible Speech with the philosophy and content of past METI is
noteworthy.
First, the aim of being self-decoding is shared. Just as the Arecibo and Dutil-
Dumas messages are intended to be self-decoding, so Bell subtitled his invention
“self-interpreting physiological letters.”
Second, the presentation of Bell’s system would be a natural extension of the
Dutil-Dumas message. Their pages 15‒16 present basic information about humans,
including a version of the Pioneer Plaque diagram. A longer self-decoding message
could include internal anatomy, especially lungs and articulatory tract. These could
then lead to images like Figure 18.1 that lay out the articulatory basis of human
speech and Bell’s symbols. In a similar vein, the Voyager Golden Discs introduce
breathing and then greetings from humans. Lomberg’s notes (cited in Lemarchand
and Lomberg 2011: 379) notes on the recording state:

I hope that some of the UN greetings will be recorded so that an intake of breath
before syllables could be heard. This would link breathing with speech, and
perhaps give a clue as to the respiratory nature of speech.

Third, besides their physiological concreteness, Bell’s symbols capture an

important cognitive fact. A foundational discovery of linguistics and especially
phonology, the study of sound systems, is that speech sounds like t, d, and n are not
primitive but are made up of smaller units, “features,” that constitute instructions
to specific articulators (Jakobson et al. 1958 [1971]). These binary oppositions are
directly reminiscent of the binary encoding used in METI itself.
Nonetheless, Bell’s letters present some challenges in the context of METI. Their
rotational symmetry is uncharacteristic of human writing systems (Morin 2018),
though not unprecedented (Canadian syllabics encode vowels by rotating conso-
nants). Also, their roundness lends itself poorly to pixelation (unlike Hangul letters).
226 Daniel Harbour

However, the most significant issue is similarity of characters. This arises from
the use of a small set of primitives to encode all sounds and from the principle of
reflecting similarity of sound by similarity of grapheme. Thus, sounds that differ
in only one articulatory feature, like k/g, are represented by characters that differ
equally minimally, /. The Dumas-Dutil message pursues the opposite strategy.
It is intended to be maximally resistant to information loss, even in a noisy recep-
tion, and none of its characters are rotations or reflections of each other (making it
highly nonhuman; Morin 2018).
These problems are all superable. The Encyclopedia Britannica contains images
of writing systems. So, Heidmann’s (1997) proposal involves sending images of
written characters. Adopting this systematically, we could send three increasingly
compact formats: full images, pixelated characters, and binary encodings. The tran-
sition from pixelation to binary can be accomplished while also breaking rotational
symmetry. In a twin-prime pn × pn + 1 grid, a pn × pn pixelation can be bookended
between a repeated p-length binary encoding of that character.5 Once enough sam-
ples of characters, words, and their phonetic transcriptions have been provided,
binary can be used for the bulk of the remainder of the message. By progressing
from human physiology to vocalization, writing, and language, an Encyclopedia
Galactica would smoothly transition from the universals of mathematics and sci-
ence to cultural wealth that is uniquely human.

Conclusion: The I in METI

We are new to METI, and our mathematics and science are unlikely to advance
other METI-engaged civilizations. However, our knowledge of humanity is unique.
Our messages should strive to express this not simply in terms of our biology, but
also in terms of our cognition. Our METI should emphasize our I. Writing systems
provide a wealth of information about our physiology, cognition, languages, and
cultures. With these facilitating the step from the universal truths of mathematics
and science to the richness of human civilization, we can turn to the much bigger
question of what the bulk of a METI-bespoke encyclopedia should deliver.

Notes
1 These languages are not picked randomly. With the exception of English and Hindi,
which are distantly related, all come from different linguistic families and differ substan-
tially in phonology, morphology, and syntax, and, in accord with the following section,
they illustrate diverse writing systems: logograms (Chinese), syllabaries (Cherokee), ab-
jads (Arabic), aksharas (Hindi), and alphabets (English, Fijian). Many other choices are
possible: Hebrew script might encode more easily than Arabic, but would mean the loss
of pharyngeals from the sample of sounds; Georgian presents a beautiful script along-
side one of the most complex morphologies of any national language; one might include
a sign language (“the body as a resource for meaning making”, Ortner, this volume
[Chapter 5]), both vis-à-vis Bell’s script in the following section, and because visual
communication might be more common among other species (pace Kershenbaum, this
volume [Chapter 2]). As with so much of METI, discussion is needed.
 Writing Systems and METI 227

2 In particular, Heidmann’s (1997) claim that “The alphabetical coding can be deciphered
using just a few pages, as well as the grammatical structures” (202) looks naïve. All suc-
cessful decipherments of ancient languages involved text-external knowledge, chiefly of
related languages. Ancient records are more fragmentary than an encyclopedia, meaning
that lessons for METI drawn from decipherment of lost languages here on Earth (Finney
and Bentley 2014; Saint-Gelais 2014) may be more pessimistic than necessary. How-
ever, Heidmann’s optimism is far from founded. Maybe computer scientists could prove
their case on the Sumerian corpus, as many basic facts about the language remain under
debate and a large number of texts have yet to be processed.
3 Aesthetics in mathematics is multifaceted. It covers beauty, simplicity, and surprise,
and applies both to results and to proofs, in different ways. Discoveries can be sensorily
aesthetic, like the Golden Ratio (Lemarchand and Lomberg 2011), or surprising, like
Morley’s theorem (the points of intersection of adjacent trisectors of adjacent angles of
a triangle form an equilateral triangle – crudely, even misshapen triangles have shapely
souls); they can also be conceptually aesthetic, like eiπ + 1 = 0, which ties together
three basic operations and five fundamental mathematical constants in a single equation.
Proofs can be aesthetic for their simplicity, like Euclid’s demonstration of the infinitude
of primes, or for the concepts they draw together, like Wiles’ proof of Fermat’s Last
Theorem, which exploits a connection to elliptical curves.
4 I construe “alphabets” broadly to include abjads (e.g., Arabic, Hebrew) and aksharas
(e.g., Amharic, Hindi), and take inherent vowels in aksharas to represent reading con-
ventions, not consonant-vowel combinations.
5 A 29 × 31 grid furnishes more than half a billion binary encodings (229). This can be
increased if other cells are unused across enough pixelations.

References
Baxter, William, and Laurent Sagart. 2014. Old Chinese: A New Reconstruction. Oxford:
Oxford University Press.
Bell, Alexander Melville. 1867. Visible Speech: The Science of Universal Alphabetics; or
Self-Interpreting Physiological Letters for the Writing of All Languages in One Alphabet.
London: Simkin, Marshall & Co.
Bell, Alexander Graham. 1872. On the Nature and Uses of Visible Speech. Boston: Rand,
Avery & Co.
Benford, James, Dominic Benford, and Gregory Benford. 2011. “Building and searching
for cost-optimized interstellar beacons.” In Communication with Extraterrestrial Intel-
ligence, edited by Douglas A. Vakoch, 279‒306. Albany, NY: State University of New
York Press.
Bromberger, Sylvain. 1992. On What We Know We Don’t Know: Explanation, Theory, Lin-
guistics, and How Questions Shape Them. Chicago, IL: University of Chicago Press.
Cook, Jia-Rui, and Dwayne Brown. 2010. NASA Probe Sees Solar Wind Decline. https://
voyager.jpl.nasa.gov/news/details.php?article_id=20.
Dick, Steven J. 1996. The Biological Universe: The Twentieth Century Extraterrestrial Life
Debate and the Limits of Science. Cambridge: Cambridge University Press.
Dumas, Stéphane. 2003. The 1999 and 2003 Messages Explained. www.plover.com/misc/
Dumas-Dutil/messages.pdf.
Dumas, Stéphane. 2011. “Proposal for an interstellar Rosetta Stone.” In Communication
with Extraterrestrial Intelligence, edited by Douglas A. Vakoch, 403‒412. Albany, NY:
State University of New York Press.
Finney, Ben, and Jerry Bentley. 2014. “A tale of two analogues: Learning at a distance
from the Ancient Greeks and Maya and the problem of deciphering extraterrestrial radio
228 Daniel Harbour

transmissions.” In Archaeology, Anthropology, and Interstellar Communication, ed-

ited by Douglas A. Vakoch, 65‒78. Washington, DC: National Aeronautics and Space
Administration.
Freudenthal, Hans. 1960. Lincos: Design of a Language for Cosmic Intercourse. Amster-
dam: North Holland.
Giraud, Anne-Lise, and David Poeppel. 2012. “Cortical oscillations and speech processing:
Emerging computational principles and operations.” Nature Neuroscience 15: 511‒517.
Harbour, Daniel. 2019a. Convergent Cultural Evolution and the Inventory of Sumerian and
Chinese Signs. Ms, Queen Mary University of London.
Harbour, Daniel. 2019b. The Limits of Logography: Phonology-Morphology Mismatches
Drive Writing System Evolution. Ms, Queen Mary University of London.
Harbour, Daniel. 2021. “Grammar drives writing system evolution: Lessons from the birth
of vowels.” In Grapholinguistics in the 21st Century 2020, edited by Yannis Haralam-
bous, 201‒221. Brest: Fluxus Editions.
Harp, Gerald R., Robert F. Ackermann, Samantha K. Blair, Jack Arbunich, Peter R. Backus,
Jill C. Tarter, and the ATA team. 2011. “A new class of SETI beacons that contain in-
formation.” In Communication with Extraterrestrial Intelligence, edited by Douglas A.
Vakoch, 47‒70. Albany, NY: State University of New York Press.
Haspelmath, Martin. 2000. “Coordination.” In Language Typology and Syntactic Descrip-
tion, Volume II: Complex Constructions, edited by Timothy Shopen, 1‒51. Cambridge:
Cambridge University Press.
Hedengren, David. 2013. Three Microeconomic Applications Using Administrative Records.
PhD thesis, George Mason University, Fairfax, VA.
Heidmann, Jean. 1997. Extraterrestrial Intelligence. Cambridge: Cambridge University
Press, Storm Dunlop (tr.). Hobaiter.
Jakobson, Roman, S. Karcevsky, and Nikolai S. Trubetzkoy. 1958 [1971]. “Quelles sont les
méthodes les mieux appropriés à un exposé complet et pratique d’une langue quelcon-
que?” In Selected Writings, Volume II: Word and Language, edited by Roman Jakobson,
3‒6. Hague: Mouton.
Jaritz, Kurt. 1967. Schriftarchäologie der Altmesopotamischen Kultur: Eine grammatolo-
gische Untersuchung zur Entstehung des ältesten Biderschriftsystems. Graz: Akademis-
che Druck.
Kim-Renaud, Young-Key. 1998. The Korean Alphabet: Its History and Structure. Honolulu,
HI: University of Hawai‘i Press.
Lemarchand, Guillermo A., and Jon Lomberg. 2011. “Communication among interstellar
intelligent species: A search for universal cognitive maps.” In Communication with Ex-
traterrestrial Intelligence, edited by Douglas A. Vakoch, 371‒398. Albany, NY: State
University of New York Press.
Luo, Huan, and David Poeppel. 2007. “Phase patterns of neuronal responses reliably dis-
criminate speech in human auditory cortex.” Neuron 54: 1001‒1010.
Morin, Olivier. 2018. Spontaneous emergence of legibility in writing systems: The case of
orientation anisotropy. Cognitive Science 42: 664‒677.
Saint-Gelais, Richard. 2014. “Beyond linear B: The metasemiotic challenge of communica-
tion with extraterrestrial intelligence.” In Archaeology, Anthropology, and Interstellar
Communication, edited by Douglas A. Vakoch, 79‒94. Washington, DC: National Aero-
nautics and Space Administration.
 Writing Systems and METI 229

Shostak, Seth. 2011. Limits on interstellar messages. In Communication with Extraterres-

trial Intelligence, edited by Douglas A. Vakoch, 357‒370. Albany, NY: State University
of New York Press.
The Staff at the National Astronomy and Ionosphere Center. 1975. “The Arecibo message of
November, 1974.” Icarus 26: 462‒466.
Stassen, Leon. 2003. “Some universal characteristics of noun phrase conjunction.” In Noun
Phrase Structure in the Languages of Europe, edited by Franz Plank, 761‒817. Berlin:
Mouton de Gruyter.
Weeden, Mark. 2014. “Anatolian hieroglyphs: Logogram vs ideogram.” In Visualizing
Knowledge and Creating Meaning in Ancient Writing Systems, edited by Shai Gordin,
81‒100. Berlin: PeWe-Verlag.
Wittgenstein, Ludwig. 1953. Philosophische Untersuchungen: Philosophical Investiga-
tions. Translated by G.E.M. Anscombe. New York: Macmillan.
Zwart, Jan-Wouter. 2005. “Some notes on coordination in head-final languages.” Linguistics
in the Netherlands 22: 231‒242.
INDEX

φ‑features 188 birdsong 5, 23, 27, 52–54

black‑capped chickadee 53
abjad 222, 226, 227 Blackfoot 188
acoustics (communication) 5–6, 9–12, 22, Bloomfield, Leonard 111
34–35, 81, 90–91, 101, 103, 190, Bobrow, Daniel 171
201–207 Boyle’s Law 191
Afro‑Asiatic (language family) 222–223 Brahmi 222
Akkadian 222
akshara 226, 227 calliphora vicina 4
alarm (calls or signals) 4, 6, 19, 22, 33–35, Cambias, James 139
52–53, 55, 89–91, 203 Canada (village) 116
alphabet 217, 220, 222–226, 227 Cartier, Jacques 116
Althusser, Louis 78 Chambers, Becky 139
American Sign Language 93, 184, 186 Champollion, Jean‑Francois 89
Amharic 227 Cherokee 219, 222, 226
Anatolian 219, 223 chimpanzee 18, 19, 21–25, 36, 52
aphasia 39 Chinese 128, 219, 221, 225, 226;
Arabic 219, 225, 226, 227 Old 221
Arecibo message 214, 216–217, 225 Chinese room thought experiment 158
Arrival 51, 99–100, 135 Close Encounters of the Third Kind 51
artificial intelligence 35, 64, 134, 135, 158, color terms 55, 97, 115, 144
165, 168, 171 Common European Framework of
associative information packets (AIPs) Reference (CEFR) 81–82
80–81 Communicating with Extraterrestrial
Athenæum 223 Intelligence (CETI) 41–47
communication 1, 5–10, 9, 11–12, 18–19,
Bardi 104 34, 42, 52, 54, 68, 128, 184, 197,
bat 12, 52, 68, 191; linguistic representation 204, 226; acoustic/auditory 5–6,
of 196 9, 11–12, 18–19, 34, 42, 52, 54,
Bell, Alexander Graham 223 68, 128, 184, 189, 197, 204, 206;
Bell, Alexander Melville 223–226 avian 5, 23, 27, 43, 52–54. 91–92;
Bilingual Image Glossary (BIG) 216–217 cetacean 33, 34, 52, 66; chemical/
 Index 231

olfactory 6–7, 34, 43, 68, 70, 71, evolution 3–4, 7–12, 22, 129, 132, 138,
79, 128–129, 190; cross‑species 19, 152–159, 160, 166, 171–178, 179,
33–34, 56–58; electrical 7–9, 12, 184, 186, 191; communication
22, 34; (electro)magnetic 9–10, 22, 62–72; cultural 42, 220
34, 128–130, 190, 203–204, 206; externalization 165, 178, 182, 189–191
prairie dog 35, 52, 89–92, 203, 208,
217; primate (non‑human) 20, 21, Faculty of Language Broad (FLB)
23, 25–28, 35, 52, 54; visual 5–6, 185–191
11, 18, 22, 34, 52, 68–69, 71, 128, Faculty of Language Narrow (FLN)
142, 184, 189, 226 185–191
compositionality 113, 119, 130–131, 135, Fermat’s Last Theorem 227
176, 186 Fibonacci sequences 80
Contact 51 Fijian 219, 226
Cooperative Principle 16, 126, 132 flycatcher 52
Core vocabulary list (CVL) 81–82; see also
Swadesh list Galileo Galilei 167–169
Cosmic Calls 216–217 Gavagai problem 81, 116
coyote 89–90 Georgian 226
cricket 201–202 German 63–64
Crimean Gothic 100 Golden Discs (Voyager Probes) 216, 225
cuneiform 65, 67 Golden Ratio 227
Greek, Ancient 89
The Darkling Sea 139 Gricean Pragmatics 7, 15–28, 43, 85, 120,
deixis 145–147 126–127
Descartes, René 165–166 Grice, Paul 7, 15–28, 43, 85, 120, 126–127
directionality 115, 144–145, 147
dog (domestic) 7, 33, 44, 45, 90, 198; Halkomelem 188
linguistic representation of 88, 116, Hangul 224–225
190, 205, 207 Hebbian learning 205
dolphin 10–11, 33, 35–36, 43, 52; Atlantic Hebrew 222, 226, 227
spotted 33; bottlenose 10, 33, 36, Hindi 219, 226, 227
43; killer whale/orca 33 Hockett, Charles 6, 10, 88–89, 97, 123–
Drake, Frank 123–135; Equation 124, 126, 128–129, 134, 153–154,
123–135, 160 160, 183–184, 190; Design
Dutil‑Dumas messages 225 Features 6, 10, 88–89, 123–124,
126, 128–129, 153–154, 160,
Economy Thesis 171–173 183–184
Egyptian 89, 221, 222; Ancient 222; Hooke’s Law 191
Demotic 89 Hopi 114–115
ELAN 103 Human‑animal interaction (HAI) 43–44
e‑language 112, 183 Hummingbird 52
electric fish 8–9, 12; Gymnotiformes 8;
Mormyriformes 8 IBM 75
elephant 52, 191 i‑language 112, 168–169, 183
Emergentist approaches 152, 159 Interactional Linguistics 40–47, 48
Encyclopedia Britannica 215, 219, 226 International Phonetic Alphabet (IPA)
English 62–65, 82, 83, 88, 100, 224–225
114–117, 145, 157, 158, 161, Iroquoian (language family) 116
183, 186, 188, 207, 216, 219,
222, 225, 226 Jacobellis v Ohio 111
ergativity 104 Japanese 145, 221–222
Euclid 218, 227 Japanese great tit 53
evidentiality 146–147 jumping spider 10
232 Index

kangaroo rat 10 Pioneer Plaques 214, 216, 225

Kant, Immanuel 218 Pirahã 160
Kanzi 93 prairie dog 35, 52, 89–92, 203, 208, 217
Koko 93 principles and parameters (P&P)
Korean 222, 224–225 168–170
Kriol 100 pronouns 115, 147
Psamtik (Pharaoh) 138
Laser Interferometer Gravitational‑Wave Ptolemy V 89
Observatory 52
Latin 62 rat 7, 10
Leibniz, Gottfried Wilhelm 168–169 recursion 18, 76, 91, 98, 153, 160,
Lincos 216–217 166–167, 176, 178, 185, 189
Linear A 89, 158 red‑tailed hawks 89–90
Linear B 222 redwing blackbird 53
locked‑in syndrome 39–40 Rescorla‑Wagner equations 200
logogram 219, 221, 226 Rongorongo 89
Rosetta Stone 89, 107
Malinowski, Bronislaw 99 Rumi 219
Maltese 222 Russell, Mary Doria 139
Marsh wren 54 Russian 216
Maxim (Gricean) 16–17, 27, 120
Mayan 221–222 Sapir, Edward 114
Menzerath’s Law 18 Sappho 219
merge 152–153, 160, 161, 167–178, 179, Search for Extraterrestrial Intelligence
185–191 (SETI) 75, 123, 126–127, 158,
Messaging Extraterrestrial Intelligence 214–217
(METI) 106–107, 123–127, 158, Sejong the Great 224–225
161, 208, 214–227 semantics 36, 80, 88, 90, 110–120, 196,
Milky Way 125, 215 219; type‑theoretic 117
Minimalist Program 128, 166–167 sensory systems 4, 11, 33–34, 182
Minsky, Marvin 123, 125, 139, 143, Shakespeare, William 219
171–176, 189 Shannon boundary 64
Model/Rival (M/R) technique 56–58 shark 7
Morley’s theorem 227 Shōnagon, Sei 219
morphology (linguistics) 10–11, 88–89, Snellen test 112
104, 110, 114, 119, 128, 130, 154, Social modeling theory 55–58
184–185, 221–222 Sogidan 222
morphology (physical) 41, 78 The Sparrow 139
sparrow 54
Olmec Script 89 speech acts 118–120
Olympus Mons 201 squirrel, ground 35
Optimality Theory 82–83 Stewart, Potter 111
orangutan 208 Strong Minimalist Thesis (SMT) 166–167,
182–191
parrot 34, 36, 54–59, 129 Sumerian 221, 227
Peirce, Charles Sanders 177 Swadesh list 81–82, 115–116
Philological Society 224 Sweet, Henry 224
Phoenician 222 syllabary 220–222, 226
phoneme 83, 88–91, 111, 127–129, 197 syllable 18, 53, 82, 161, 220–222
phonology 127–128, 154, 156, 190, syntax 36, 44, 91, 107, 113, 127–128,
225–226 154–156, 161, 166, 173, 176,
pinnipeds 191 185–189, 218–219, 226
 Index 233

Tagalog 131 Voyager probes 216, 225

third factor conditions 97, 139–143, Voynich Manuscript 97–98
182–191
truth 27, 110–113, 117, 127, 132 warbler 53
Turing machine 169–178, 179, 189 Washoe 93
Wayfarer (book series) 139
Ulam, Stanislaw 177 Wiles’ proof 227
Universal Grammar (UG) 97, 165–179, wolf 10–11, 19
182–191 World Atlas of Linguistic Structures 187

Vakoch, Douglas A. 76 Yiddish 222

vervet monkey 19, 35
Vietnamese 221 zebra finch 53
Vinca 89 Zhuang 221
Visible Speech 223–226 Zipf’s Law 18, 70, 92