How did languages evolve

• Did language evolve?

• Or
• Did it appear suddenly?
• Not an easy one to answer. Often speculations overtake actual
evidence
• So much so that, Société de Linguistique de Paris in 1866
banned inquiry into the origin of languages
• However, it is 2023 and in the last 3 or 4 decades, a lot of new
claims and counter-claims have been made on this.
• Some of these claims will be discussed now.
Evolution
• “nothing in biology makes sense except in the light of evolution”
(Dobzhansky, 1973)
• Language too makes sense in the light of evolution
• However, unlike other domains, language evolution is a little tough to
track since it leaves no fossil
• Written language arguably evolved much later than spoken and
signed languages.
Approaches to study the evolution of language
• comparative animal behavior,
• paleontology and archaeology,
• molecular biology, and
• mathematical modeling
• psychology
• linguistics
• biology
• archaeology and
• complex systems science
Why is language so special?
• Language is considered uniquely human

• Hockett (1960) suggested design features that make human language

different from other communication systems.

• arbitrariness, generativity, displacement, and reflexiveness.

• arbitrary in that the signs are symbolic; word forms in language do not
need to have any specific connection to their meanings.

• generative, that is, language systems can be used to generate new

sentences and words.

• Human language also allows for displacement: we can refer to things

and events distant in space and time

• Finally, language can be reflexive: it allows us to not only talk about

ourselves, but about language itself
agreements
• Scientists generally agree that languages did evolve.

• This is a unique system of communication, different from all other

animal communications

• Languages developed somewhere between 2 million and 2,00,000

years
 protolanguage
• If language did evolve, what were the earlier forms?

• The earlier forms need to be simpler than the fully evolved ones

• This intermediate stage prior to full-blown human language known as protolanguage

• Protolanguage could have been entirely vocal, entirely gestural or a mixture of vocal
and gestural signals
• Studies in the field of historical linguistics and language change show
that languages do change over time

• Hence, it is highly likely that language at the time of its origin, was
different from how it is now.

• However, there is a lot of disagreements among scholars about the

origin of language.
• The debates surround 2 main aspects: biological and cultural
evolution (of language)
Biological markers
FoxP2
• Research on FoxP2 gene started in a strange way. A family had a
number of SLI patients and it seemed to be hereditary.
• This was the first time, language was connected to a gene.
• Mutated version of the gene was found to be responsible for SLI
• This discovery led to the belief that language evolution could have
some genetic underpinning
• Even if this is the case, language capacity would be highly
‘polygenetic’, complex interplay of many genes.
Vocal tract
• This is another biological/physiological adaptation thought to be
responsible for evolution of language among humans

• Human vocal tract is different from that of our closest cousins among
higher primates.
• This puts us at more risk of chocking on food than them.
• In terms of natural selection, there must have been some adaptive
advantages of this change to over ride the disadvantage, namely,
language.
• Understanding human language evolution often involves comparing
us with other primates.
• This results in two types of comparison
• Analogous and homologous
• analogous structures (those similar to other animals, but which likely
developed independently under different selective pressures) and
homologous structures (those shared with our closest animal relatives
via shared ancestry).
analogous structures
• Analogous structures have the potential to reveal what selective
pressures led us to develop language.
• Comparing birdsong with human language structures
• Drawing parallel to birdsong, language evolution has been connected to:
 sexual selection
 motivated and reinforced via specific oxytocin pathways and so on.

• Some other animals also have surprising parallels to human language

 Bee dance direction, distance, and attractiveness of pollen sources
 ‘name calls’ – clicks and whistles they use to refer to specific
individuals

 These animals show displacement and reflexiveness to some degree

like human language
homologous structures
• Studying homologies of language involve communicative and
cognitive behavior of our closest relatives in the animal world
• This can lead us to know what aspect of language was present in our
last common ancestor
• This helps us understand what it is about human language that is
special and what happened during evolution that resulted in us
speaking, while our relatives did not.
• communicative behavior of other great apes , alarm calls in monkeys etc are studies for
this purpose.
Brain structures
• Brain changes over time:
• human brain increased in size and complexity when homo sapiens
became distinguished from other species around 2 million and
3,00,000 years ago.
• Early hominids did have structure corresponding to Broca’s area,
however over time this area went through changes gradually enabling
humans better control over articulatory apparatus.
• Scientists believe that Neanderthal did not have adequate control of
tongue and hence would not have been able to speak as effortlessly
as we do.
• brain imaging studies that show great apes and humans share
similarity of brain regions, BA44.

• This region is enlarged in both higher primates and humans. I

• n apes this large BA44 corresponds to sophisticated gesturers and for

humans this is responsible for language use
Cultural evolution of language
• While biological investigation focuses on aspects of language at
individual level. So e.g. structure of vocal tract, brain areas etc.
• The cultural studies on language evolution, on the other hand, looks
at interaction and learning.

• Cultural accounts of language evolution often depend on computational

models, psychological experiments, and linguistic data, but also draw to some extent on
developmental and comparative data.
• cultural interaction and learning biases are important points when looking at
computational modelling of language evolution.
Learnability or learning bias
• In order to survive, a language must be learnable and useful.
• These two factors combined, give rise to language that are structured.
• For example, learning a simple rule of plurals is better than rote
learning of all the plural forms in English.
• Also, imagine all plural in English were irregular (goose-geese; ox-
oxen; mouse-mice kinds)
• The rule governed structure thus helps in easy learning, expressibility
and transfer of the skill through generations
• evidence for these accounts comes from agent-based modeling and more traditional
psychological experiments, particularly in artificial language learning
Naming game
• A pioneering model examining the role of interaction in language evolution is known as
the Naming Game
• The Naming Game focuses on a population of interacting artificial agents tasked with
labelling a particular meaning
• agents begin by generating random strings for the meaning.
• communicative success within the population is initially low (that is, agents do not
agree on ‘which label the meaning should have’)
• a simple bias to discard all previous labels when they meet an agent which shares one
of their labels leads to high communicative success over time
• this simple bias allows agents to converge on a shared conventional label for a meaning
• This finding emphasizes the role communicative function in language evolution: the
pressure to communicate successfully, and the process of communication itself, can
shape language over time.
• Another social learning perspective focuses on coordination in learning instead of
interaction.
• Language learners do not receive disembodied input, rather, they learn from the output
of other language users.

• The iterated learning model begins with agents learning a subset of an initially random
artificial vocabulary
• A subset of the output of this first “generation” of agents then becomes the input of the
next generation
• Since agents receive only a subset of the input - but are tested the entire vocabulary -
they generalize patterns to unseen items (much like knowing blog-blogs leads to selfie-
selfies) and structure begins to accumulate within the vocabulary over successive
generations of learners.
• The iterated learning model has also been implemented experimentally: participants
were tasked with learning initially random pairings of nonsense words and meanings,
and their output was transmitted as the learning input for the next participant (
• results showed that initially random languages become structured, that is, words which
shared aspects of forms referred to similar meanings
• supporting the notion that weak individual biases are amplified by transmission and
shape the structure of language
• This line of research highlights the key role of cultural transmission in language
evolution. Each individual agent or participant only has a small bias towards structure,
but this bias is amplified by transmission, creating a highly structured language.
• These approaches emphasize the key role that cultural processes can play in creating
linguistic meaning and structure
• This kind of approach has also been extended to primates.
• Instead of a vocabulary, this study tasked baboons with learning simple random patterns
on a grid with 16 cells (Claidière et al., 2014).
• Like in the model and experiment outlined above, the baboons then reproduce the
simple patterns in a testing phase.
• Their test output then becomes the input for another baboon, creating a ‘chain’ of
learning which includes several individuals.
• At the end of these chains, the patterns are no longer random, but exhibit structure –
specifically, they resemble ‘tetronimoes’ – the shapes from the classic game Tetris,
which involve 4 consecutive filled cells
• Although biological and cultural perspectives are often characterized as alternatives, the
emerging view in language evolution is that biological and cultural factors interact with
individual learners to shape language
• From the biological perspective, there is broad agreement that at the very least some
biological adaptation primes us for language since no other species that we know of has
all of the key components of language in a single system
• On the other hand, language requires some cultural input in development, as evidenced
by unfortunate cases of language deprived children who have persistent difficultly with
structure in language in particular
Source: Kirby (2002).
• Any biological underpinnings of language necessarily evolve on the relatively slow
genetic timescale, on the order of thousands of years at a minimum.

• The cultural or glossogenetic timescale on the other hand, changes rapidly; for example,
new words emerge and others die within mere years.

• The individual or ontogenetic timescale represents language development – while

human children are undoubtedly primed to learn language biologically, their linguistic
input comes from the cultural realm, and their output in turn feeds back into the cultural
timescale.

• The cumulative language behavior of a population of individuals has the potential to

affect the input of new learners, and thus shape selective pressures at work on the
genetic scale, resulting in complex and ongoing co-evolution of genes and culture
Some of the arguments
• So far we have seen the fundamental ideas that have been put
forward to understand evolution of language in humans
• Now let’s try to follow some of the major currents of academic
arguments in this regard.
• The arguments and counter arguments have come from fields as
diverse as Linguistics, anthropology, psychology, evolutionary biology,
primatology, paleontology, cognitive science, neuroscience to name a
few.
• The main split is concerning whether language is primarily a tool for
communication or for structuring thought.
• . Also, scholars differ on whether there was an ‘evolutionary leap’
leading to discontinuity between humans and other primates
 Or there was evolutionary continuity and gradual change
• Let’s start with Chomsky and Hauser group.
• They mainly looked at biological aspect of the evolution, and not
cultural aspects
Their main claims

• Richness of ideas is accompanied by a poverty of evidence, with essentially no

explanation of how and why our linguistic computations and representations
evolved.
• (1)studies of non human animals provide virtually no relevant parallels to
human linguistic communication, and none to the underlying biological
capacity;
• (2)the fossil and archaeological evidence doesnot inform our understanding of
the computations and representations of our earliest ancestors, leaving details
of origins and selective pressure unresolved;
• (3)our understanding of the genetics of language is so impoverished that there
is little hope of connecting genes to linguistic processes anytime soon;
• (4)all modelling attempts have made unfounded assumptions, and have
provided no empirical tests,thus leaving any insights into language’s origins
unverifiable.
• the core competence for language is a biological capacity shared by
all humans and distinguished by the central feature of discrete infinity
—the capacity for unbounded composition of various linguistic objects
into complex structures.

• Hence the focus of linguistic speculation on evolution of language has

been on syntax.

• Chomsky propagated the idea of Universal Grammar, which parimarily

means that all humans are bioprogrammed to learn language.
• These distinctions were mapped, respectively, to the Faculty of
Language in the Broad (FLB) sense and the Faculty of Language in the
Narrow (FLN) sense. FLB designates processes that are shared with
other animals, and thus, are involved in language and other sensory-
motor and conceptual-intentional processes.
• FLN, in contrast, describes processes that are uniquely human and
unique to language
Gestural theory of language evolution

• Hewes ( 1973)
• The gestural theory of language evolution, our ancestors
communicated primarily through a language composed of arm and
hand gestures
• He also stressed that gestures have not ben completely replaced by
language, as gestures are a part of our communications still
• different aspects of modern culture are rooted in gesture (e.g.,
painting) and vocal communication (e.g., poetry) respectively
• Evidence from primates:
 They cannot be taught to speak though they use complex gestures
 Their vocalizations are involuntary emotional reactions and used even when
no one is there to hear (meaning communications is not the goal)
Gap in Hewes’ theory
• He does not clarify how gestures convert into language
• What are the processes involved
Corballis
• A number of subsequent researchers have built upon Hewes’s theory
• Corballis is one among them and very well known
• One key point put forward by him is that skilled motor movements
and language are typically both supported by the Broca’s area.
• He finds a ‘brain basis’ for connecting gestures with vocal language.
• Even today, brain imaging studies reveal a close connection with
motor movement and speaking in humans.
• In non-human primates, this area is devoted to complex gestures.
Mirror neurons
• Also the finding of mirror neurons helped strengthen his argument.

• Mirror neurons are those neurons that get activated when one is doing
an action or watching others doing the same thing ( playing the piano
e.g.).

• These are spread across the brain, most notably they are seen in BS
44 and 45 also.

• Hence these neurons would have provided early ‘speakers’ the

feedback they needed about others’ intensions and perspectives.
• Another important parallel he point out is that of sign language, which
is as complex, grammatically, as spoken language and hence are
parallel in many ways.
• Corballis however does not say that gestures ‘turned into’ language.
• Rather he considers speaking as another kind of gesture, a facial one.

• He also takes help from motor theory of speech perception

(Lieberman)
 Our understanding of language is derived from how words are produced
rather than the specific sounds produced
• gradual transition from manual gestures to increasing use of gestures
of the face and mouth could have co-occurred with the increasing
involvement of the hands in tool manufacture and use.

• Researchers have also used studies from the field of developmental

psychology to support the gestural theory.

• Infants typically learn to express concepts first through gestures (e.g.,

by pointing at a cat) before later expressing them through spoken
words
Philip Lieberman (Eve Spoke, 1998)

• His primary focus is the neural bases of language evolution

• According to him, language does not only serve the purpose to
communicate, but it is also a part of thought.
• The reason is that language is inextricably connected to motor
control, attention and other aspects of human behaviour.
• Chimpanzees have latent linguistic abilities. Neanderthals and other
archaic hominins possessed some form of language. Similar parallel
can be found for all human behavior.

• So the same neural structures were present/are still present in many

other analogous and homologous species.
However, there is a catch
• Same neural structure ≠ same functions

• Neural structures that date back to species that lived millions of years
ago and served other ends have been recycled to regulate speech,
lexical access, syntax and cognition.

• His theory is based on the argument that existing neural structures

were ‘repurposed’ to give humans the ability to speak.

• For example, The cerebellum and hippocampus are archaic structures.

Today they play role in cognition (memory).
Other examples

• The subcortical basal ganglia appear in early anurans (tailless

amphibians)
• The same structure in today’s humans
 regulate motor control, including speech
 support neural circuits linking prefrontal cortex, posterior
cortical regions and other subcortical structures
 These circuits are active in linguistic tasks such as lexical
access, comprehending distinctions in meaning conferred by
syntax and the range of higher cognitive tasks involving
executive control
 The basal ganglia play a critical role in conferring cognitive
flexibility.
Also,
• The cingulate cortex were found in Therapsids (archaic
mammal ancestors) transitional mammal-like reptiles .

• This cortical part were and still are involved in mammalian mother-
infant interactions

• It also controls laryngeal phonation

• is active in directing attention in cognitive tasks.

Natural selection and abrupt changes

• Lieberman turn to Darwin for support of this theory.

• Darwin stressed on the gradual nature of natural selection.

• However, he also knew that there were abrupt changes during

evolution, like, change from aquatic to terrestrial life.

• His solution for this type of evolution was recycling --- adapting an
organ for a new purpose.
• Current studies confirm that recycling is a general process and this
includes the cerebellum and basal ganglia which initially appear to
have been adapted for motor tasks functioning in cognitive tasks,
including language (e.g., Thompson, 19 ; Maraden and Obeso, 1994;
Lieberman, 2000, 2006, 2013; Monchi et al. 2001,2006.2007).
• He strongly disagrees with Chomsky that human language suddenly
appeared somewhere between 70,000 to 100,000
• Chomsky’s idea of principles and parameters that is genetically
transmitted and thus helps children acquire language is also
contested.
• Genetic variation is characteristic of all living organisms, including
humans. Thus Universal Grammar is not compatible with this
variation.
• Chomsky came up with the idea of Merge to counter this.
• Chomsky’s Merge operation yields the defining characteristic of all
human languages - complex sentences with embedded clauses.
• Merge which accounts for the acquisition of language by children and
its sudden evolution “takes exactly two (syntactic) elements and puts
them together” (Bolhuis et al. 2014).
• This too makes no sense, as different languages put different things
together.
• In English it could be determiner-noun: the cow
• Chinese does not use determiner.
• This means all the variations in all possible human languages would
have to be innately specified
• That too suddenly at about 50,000, 70,000 or 100,000 years ago.

• This whole claim, as per Lieberman, belongs to ‘the realm of fantasy’

• So, he concludes that human language evolution is ‘natural selection
acting on heritable biological variation’.

• Natural selection acting on genetic and epigenetic events in the last

500,000 years enhanced human capabilities.

• He also puts emphasis on ‘epigenetic changes arising from

environmental effects, such as disease susceptibility of descendants’

 For example, lactose tolerance among adults developing in communities that

domesticated sheep, cattle and goat etc.

 Similarly, lactose-free food-stuff on supermarket shelf indicates tht many of

today’s population had ancestors who lived in communities where milk did
not form part of adult diet. Hence the intolerance.
Culturally based theories
Daniel Everett
(Don’t sleep there are snakes; Language: the cultural
tool; How languages began)
• In the beginning there was word (John 1.1)
• No, it wasn’t (Dan Everett)
• Everett is an anthropological linguist, known for his work on the Piraha
language.
• language emerged as an instrumental product of certain cognitive
capacities of the human brain, to meet some of the requirements of
being the kind of social animal that we are.

• for strengthening mental skills (think reasoning, mathematics,

symbolic logic)
• language is a cultural tool—a tool for shaping the way members of
human communities behave, think, and know.
Everett rejects that language is innate and says that there is a combination of
human capacities that make language possible.

He divides these capacities into three types: the physical, the cerebral, and the
cognitive, which he refers to as platforms.

Each platform is made up of a set of what he regards as basic capacities, specific

to humans but not to language.

Briefly the physical platform is the human body. The cerebral and cognitive
platforms are actually subplatforms of the brain.
• Everett focuses on language as a ‘learned’ skill rather than an inherited
one.

• And, also stresses on the cultural aspects of it, by taking

conceptualization as the starting point.

• Everett offers many examples of how culture affects language; the

gamut runs from vocabulary to idioms to grammatical categories to
syntactic constructions.

His main theory, as explained in many of his books is that language and
culture are inseparable and hence would have helped in the development
of each other.
Simon Kirby (professor, artist and DJ)
• AS per Kirby, ‘complex interactions between individual learning,
cultural transmission and biological evolution in human populations.’

• He agrees that language is, at least in part, a learned behavior.

• Differences/variations in languages have no obvious correlations with

genetic differences in their speakers.

• Hence these variations are the marker of the aspect of language that
is ‘learned’.
• Structure and iterated learning’ are the points of departure for this
thesis.

• Language is unique because of iterated learning, learning from

another person who in turn learnt from yet another.
 Sort of a Chinese whisper game.

• This is how language is transmitted.

• We inherit language through 2 pathways, so to say

 First, mechanisms to learn, genetically
 Language themselves, culturally
His experiment

• A computer generated string of syllables and other formats of artificial

‘language’
• Participants learn it
• Then they are tested on ‘what did you learn?’
• Output of participant 1 as input to participant 2
• Over time and participants, the end product changes
• Similarly, novel picture (wong) –novel name pair learning
• Initially people make too many mistakes, but over a period, they get it
• The mistakes made by the earlier participants seem to make learning
easier for the later participants ( as they create some pattern to it, may
be?)
Kirby

Image copyright Simon Kirby

• The most important part of the experiment is that the stimulus has no
structure initially.

• But through the Chinese whisper, eventually a pattern seems to

emerge and the later participants get better at remembering them

• This points to the fact that learnability increases through iterated

learning.

• So language gets a structure through cultural evolution

Another type of experiment of Kirby: simulation

• Is biology not important, then?

• What happens when biological and cultural evolution happens at the
same time?
• Computer simulation (modelled on Sony curio robot)
• Simulated sets of genes determine innate learning preference
• Simulated computer learners
• Communicating and learning with each other in a computer model
• They pass on their genetic material (analogue to human genes)
• Universal features emerge culturally and natural selection of good
communicators
• Start to get this property culturally transmitted
• Get genes make you learn better
• Culture to genes: nativization

• ONLY, THIS DOES NOT HAPPEN

 Because culture shields genetic changes from natural selection
domestication
• Siberian fox, bred for lack of aggression, learn social cues
• Domesticated Bengalese finch learnt complex songs as opposed to
their wild counter part
• Domestication lift selection pressure
• So, impact on biology of culture---culture decides how we adapt----this
is a new kind of evolution
 To sum up
• neurological systems for the acquisition of language, the
representation of linguistic knowledge, and the rapid on-line
processing of language, as well as physical apparatus for the
production of speech.
• However, he says the following:
• “The existence of culture radically changes the way biological
evolution works in our species. Strong innate universals are unlikely”
• Translation: Chomsky is wrong.
Another theory
• A theory often opposed to gestural theory is that of vocal theory.
• This theory says that human language developed from simple
vocalizations, such as found among other animals.
• vocal theories have a far deeper grounding in historical models of
language, going back to the ancient Greeks
• Vocal expression of emotion, of which both music and language are
considered parts, is at the root of this theory
• But contemporary theories do not consider ‘emotion’ as an important
factor in the evolution of language

• At least not among the central factors

• The focus is rather on:
• complex phonemic repertoires, the nature of syllable structure, vocal
learning, descent of the human larynx etc.
• Jespersen (1922) claimed that our ancestors “sang out their feelings
long before they were able to speak their thoughts”
• Brown however, puts emphasis on ‘phylogenetic issues of cognitive
structure, rather than on Darwinian issues of adaptiveness and
selection mechanisms’
• Theories of language evolution based on two standpoints:
• Language as a mode of communication = dialogic view(focused on
social interaction
• Language as a cognitive tool (for thought) = monologic view(based on
internal thought)
• If we consider the dialogic view, vocal prosody forms an integral part
of it.

• Emotional meaning, according to rules of expression

• That brings us to prosody:
• expressive melodic and rhythmic features of an utterance that convey
information about emotion, intention, attentional focus, and
communicative stance
• In an utterance, combining meaningless symbols, prosody adds
meaning
• E.g. bee!
• By virtue of the prosody involved here, this is a complete sentence.
• Thus ‘emotional semantics’ is portrayed as important in this theory
Prosodic scaffold
• Holistic intonational formula
• Rhythm
• Stress groups
• Heterometric rhythm
• Affective prosody
• Polyphonic texture
Only later
• syntax-based phrase generation emerged, from its origin, as the filling
out of a prosodic scaffold during speech production process
• two-step evolution :
• first an involuntary but ritualized system of affective prosody,
• followed by a learning-based system of intonational prosody grounded
in phonemic combinatoriality.
• From there, language and music branched out as separate, functions
through the emergence of lexicality and tonality,
• After their separation, language and music are perennially reunited in
songs with words.
Recent findings

• Recent studies show a surprising parallel between language and

birdsong
• Comparison between juvenile zebra finches (Taeniopygia guttata) and
infants, such parallels emerge:
• two learning strategies:
• serial repetition (ZF)/analytic (Human infant): repeating syllables and
producing accurate output
• motif strategy (ZF) /gestalt (Human infant): repeating the entire
stretch of utterance and producing incorrect output
common neurobiological
substrate
• Human language and birdsong share a common neural substrate
• Regions in the forebrain controlling vocal production have been
identified in humans as well as three independent lineages of
songbirds, showing convergent specializations
• However, in birds that do not sing (chicken) and primates that do not
have any language (macaques) no direct projection connects the
vocal motor cortex to brainstem vocal motor neurons.
Debt to Darwin

• Darwin (1871) assumed that the earliest musical protolanguage did

not contain any propositional meaning (truth condition) like birdsong.

• Birds sing to convey intention, typically for mate selection.

• Given the lack of meaning, this musical protolanguage by itself could

not have developed into human language.

• Darwin suggested that our ancestors began to interweave gestures

and sound imitations of other animals as precursors to words in order
to insert meaning into the musical sequences
Books to read