This page is an attempt at describing a metamodel of language origins. It is intended to be a topic for discussion and not a definitive map of how language evolved. The route to language given here illustrates how a comprehensive theory of language origins could be structured, not how language actually originated.


The approach used in this model is cognitivist, so the proposal makes certain basic assumptions. All of these assumptions are disputed, so it is important to make them explicit:


The model shows a series of capacities, given in a rough order and with dependencies indicated. A timescale is not indicated, but the capacities are colour-coded by probable species, as follows:

These species allocations are arbitrary and open to dispute.


This model should be viewed as a hypothesis (or, to be more accurate, a hypothesis about a hypothesis), and criticisms and suggestions for improvement are most welcome. If this project is of interest, or if you have any thoughts about it, you can contact me at martin.edwardes@btopenworld.com to discuss it further.


Complex sound as a costly signal

(Steven Mithen, 2005)

Humans seem to value aspects of sound which are not directly linguistic, e.g. musicality and rhythm.

Deceptive Mimicry

(Chris Knight & Jerome Lewis, 2014)

The capacity to imitate animal calls gives a hunting advantage, but it also allows a mapping between sounds and objects. The sound that “calls” the animal also represents the animal.

Reverse Dominance

(Christopher Boehm, 1999)

Humans suppress alpha behaviours, and we value, and promote, co-operation to suppress dominance.

Tool-making

(Marek Kohn, 1999)

Tool-making requires visualisation: you have to imagine the completed tool as you are creating it.

Communicative Cooperation

(Jean Louis Dessalles, 2007)

This involves willingness to offer information and willingness to accept information.

What is being trusted is the message-maker, not necessarily the message.

Teaching and learning of skills

(Boyd, Richerson & Henrich, 2011)

Skills do not disappear when individuals die.

individuals must be willing to spend time teaching and learning skills.

PHYSICAL CULTURE

Sharing of Models

(Robin I M Dunbar, 1996)

At some stage we became willing and able to share our internal social models.

Protolanguage 1 is extended, swiftly leading to the following three stages.

Affective Teaching and Learning

(Mufwene, 2013)

Sharing social models allows the sharers to enhance their own social knowledge.

SOCIAL CULTURE

Simple Arithmetic

(Livingstone et al, 2014)

Rhesus macaques can perform simple arithmetic using symbols for numbers. Arithmetic involves nesting values inside other values.

NESTED FUNCTIONALITY

Awareness of Receiver

Sent models can contain models of the receiver.

The knowledge the receiver has of themself is different to their knowledge of others.

Awareness of Self

Received models of others can contain models of me.

To incorporate these models I must be able to produce a model of self-as-other

AWARENESS OF SELF

Recursive Social Cognition

Received 3-argument models may be deceptions by C, or by D, the new sender; so it is useful to tag them with their source (A-rel-B-by-C-by-D)

RECURSION

Recognition of Receiver

It is useful to the sender to recognise when the receiver is modelled in utterance, because the receiver knows about their own relationships.

“You” is a privileged “they”.

SECOND PERSON

Awareness of Absence

Sent models refer to individuals not by direct deixis but by agreed referential names.

So there is no need for the referred-to individuals to be present at the time of reference.

ABSENT REFERENCE

Egalitarian Awareness

Modelling self-as-other creates awareness of a disinterested self-as-other model, not an intimate real and selfish self.

DISPASSIONATE SELF

AWARENESS OF SELFNESS

Awareness of Non-Current Events

Absent reference means models do not need to be current. For deniability, it becomes useful to place the models being offered in relation to the present.

TEMPORALITY OF PAST

Complex Language 1


COMPLEX LANGUAGE 1

Routes to Language


Political Singularity

(Jean Louis Dessalles, 2014)

Tools for killing animals can also be used to kill rivals. Alpha behaviours can be penalised by alliances or even individual action.

Phonology

(Peter F MacNeilage, 2008)

The capacity to produce complex streams of sounds is a sign of cognitive complexity, and may even be an indicator of other types of dexterity.

Protolanguage 2

Social models need a two-argument grammar.

If it is mono-directional (A likes B says nothing about B’s view of A) it will also require a syntax.

A-RELATIONSHIP-B GRAMMAR

Identifying Opinion

Received models of others may be deceptive or mistaken, so it is useful to tag them by source

GRAMMATICAL HIERARCHY

A-RELATIONSHIP-B-BY-C COGNITION

Social Cognition

(Robin I M Dunbar, 2004)

Modelling emotional relationships between others as abstracts; Modelling abstract entities stripped of my own emotions.

A-RELATIONSHIP-B COGNITION

ABSTRACTION

THIRD PERSON

Machiavellian Intelligence

(Richard W Byrne, 1995)

Using others as tools allows me to recruit the muscles of others to do work for me. I need to track my relationships with others.

RELATIONSHIP+A COGNITION

Tool Use

(Christophe & Hedwige Boesch, 1990)

Tools are indirect objects, means to ends rather than ends themselves.

Linear Social Cognition

(Dorothy Cheney & Robert Seyfarth, 2007)

Baboons model a linear hierarchy of individual dominance, and the dominance hierarchy between families.

A+RELATIONSHIP+B COGNITION

Joint Attention & Deixis

(Michael Tomasello, 2008)

Being able to pay attention to the same thing allows people to focus attention outward.

SEGMENTATION            DIFFERENTIATION

SEMIOTICS

Joint ventures

(Mark Pagel, 2012)

Joint ventures let individuals cooperate for mutual gain. Individuals need to model their volatile relationships with others in detail.

RELATIONSHIP-A COGNITION

Merge

(M Hauser, N Chomsky & W T Fitch, 2002)

“Infinite employment of finite means”

(Wilhelm von Humboldt)

Projection of Self

Modelling self-as-other creates a self which can be projected – a first person to be communicated.

FIRST PERSON

Awareness of Own Reputation

Sharing 3-argument forms allows deniability. C’s reputation, not mine, is damaged by wrong info.

MODALITY

A-RELATIONSHIP-B-BY-C GRAMMAR

Awareness of Future Modality

By combining past temporality and modality, representation of models which have not yet happened becomes possible

TEMPORALITY OF FUTURE

Negotiation to Meaning

(Deutscher, 2005; Heine & Kuteva, 2007)

The sharing of simple grammatical utterances begins a process leading to negotiation to meaning through utterance form as well as lexis.

GRAMMATICALIZATION

Protolanguage 1

(Derek Bickerton, 2009)

Naming for attention; Manding, Imperatives; Stating, Declaratives; Coordinating, Interrogatives; Agreement and Negation; Simple semanticity (concrete terms)

RELATIONSHIP-A GRAMMAR

QUALIFIER-A GRAMMAR

Semanticity

(James R Hurford, 2007)

Mapping meanings to sound allows conventional naming and a new way of exchanging ideas.

It also allows negotiation to lexical meaning.

LEXIS

Vigilant Sharing

(David Erdal & Andrew Whiten, 1994)

Humans are aware of equity, and we value, and pay attention to, sharing. We are willing to altruistically punish unfair sharers.

Bipedality

(Bramble & Carrier, 1983)

Bipedality allowed greater control over breathing, as well as freeing the hands for tool use


Introduction            Bulletins            Members            Books            Project: Routes to Language

Ostensive-Inferential Communication

(Thom Scott-Phillips, 2015)

The meaning in signalling is no longer just the meaning of the signal. The intentions of the receiver are accommodated in the signal by the sender, and the intentions of the sender are read into the signal by the receiver.