The Origin and Evolution of Human Language
Prof. Suzanne Kemmer
Books, Websites, and other Sources
Complex behaviors are generally assumed to develop incrementally during the course of the evolution of an organism, especially when they involve learning. Such behaviors are too complex to develop simply by means of single a simple mutation and subsequent selection for a single resulting gene (as Chomsky has claimed but neuroscientists, and now other scientists using empirical evidence of various sorts, reject).
In the case of human language, the behavior is so complex that there is a whole range of skills and capacities involved, from articulatory abilities (e.g. fine, coordinated motor routines for the pharynx, larynx and articulators), to social abilities to negotiate turn-taking and sharing of attention to external object, to conceptual abilities to categorize the world in useful ways and draw inferences about cause and effect which allow for correct predictions about what will happen if you do x,y,z.
We can usefully distinguish three dimensions of development for the various skills and capacities that appeared over the course of hominid evolution and ultimately formed the necessary underpinnings for human language. These types of development are: Neuro-cognitive, social, and physical.
Note that "development" here means evolutionary development, i.e change over time in an evolving branch of the primate family. Psychologists most often use the term "development" and "developmental" to refer to changes in an organism from birth to adulthood. Evolutionary development is called phylogenetic development and development of organisms in their lifetime is called ontogenetic development. This summary is about phylogenetic development, although ontogeny is also important in the history of species.
As a caveat, it must be remembered that the three dimensions are not strictly divisible from one another. For one thing, all of the social and physical developments that were needed for language must have some neurological and cognitive correlates; we can't have the social and physical parts without the brain structures to support them. For example, it would not be possible for a modified body part such as the modern larynx, or a new set of tongue muscles, to develop without a part of the cortex also developing, or even coming to be devoted to controlling the body part.
Conversely, it does seem logically and practically possible for complex organisms to develop neuro-cognitive skills that do not necessarily have social or physical correlates. This kind of situation is illustrated by some of the cognitive capacities discovered in great apes that are apparently 'unused' in the wild and do not have social functions or adapted physical structures to go with them, e.g. learning some use of symbols in a laboratory context.
I will use the term physical below to refer to physical structures OUTSIDE the brain, since the heading neuro-cognitive necessarily includes not only cognitive/representational, but also physical brain structures.
Questions that are implied by these developments include: at what point in the hominid line did each capacity or skill begin to develop, and at what point did it reach something close to or identical to its modern human correlate? What was the sequence of development of each skill? How does it relate to other skills in the list? Are there some skills that inexorably lead to others, or provide the potential for other skills to "piggyback" on?
More questions: Which pre-human species, if any, show evidence of having had certain of these capacities, and if so, which? Which are the most recent developments? Are there any developments we would want to place in "modern" history, i.e. in the period AFTER the first Homo sapiens sapiens appeared? How did Homo sapiens neanderthalensis compare to Homo sapiens sapiens in respect to these capacities and associated behaviors?
Cortical structures for controlling muscles during extended periods of comfortable bipedal gait (evolutionarily early, pre-Australopithecine; see under physical development below). Some consider the motor sequencing involved in bipedal gait to be a precursor to later sequencing capacities relevant to language (Lieberman); some link bipedal position and gait to unleashing a capacity for a wider range of communicative gesture (Corballis).
Cortical structures for controlling muscles in throwing, aiming. Monkeys have mirror neurons linking visual and motor cortex in an area of the brain that is physically homologous to Broca's area. In humans, Broca's area plays a role in language, gesture, and the motor control processes connected with these (e.g. articulation, and the visual influences on perception of linguistic sounds, cf. the interesting McGurk effect phenomena.) It is just beginning to be recognized that a simple analogue of Broca's area in monkeys does exist and is doing work linking vision and motor control.
Research on great apes on this topic is not so advanced because it's much harder to do intrusive brain research on apes. (For one thing they are much bigger and stronger than monkeys!) Mirror neuron research on humans cannot be done intrusively either, and so research in this area is more inferential. Human neural cells are orders of magnitude more numerous and organized into far more complex hierarchical connections than monkey neurons. However, neuroscientists like Michael Arbib and Marco Iacoboni are studying the human motor-visual system and modeling it, and are suggesting testable hypotheses for establishing the precise neural mechanisms underlying observed behavioral links in humans between vision and viewed motor action.
Control of vocalization in modern humans is largely carried out in the cortex; in great apes it is subcortical. Therefore at some point a change must have occurred to link vocalization to the cortex and create the modern Broca's area. Corballis links gesture and speech evolutionarily.
Extremely high degree of cortical plasticity, particularly in neonates. Automatic self-repair and reconfiguration of the cortex when damaged, especially early in lifespan.
Extreme brain lateralization (asymmetry). Some degree of lateralization, in particular left-hemispheric dominance for vocalization, is evolutionarily quite old. But large amounts of the language cortex of modern humans are on one side of the brain (usually left), so brain lateralization relates to language development. This level of lateralization must be more recent than divergence point of human line from great apes: apes don't show much "handedness" in the sense of likelihood of preferring one hand over the other for manual manipulations and "proto-gestures". Monkeys have even more brain symmetry than apes, although lateralization is present as stated above.
Neurological substrate for categorical perception: identification of sharp boundaries in perceptual stimuli is apparently an old feature of the mammalian brain (until recently thought erroneously to give evidence for innate human language capacity, until found in other mammals, like chinchillas). Was there any adaptive development in this feature in the primate family tree?
Recognition of group-member vs. non-group-member. This is evolutionarily ancient: pre-primate, even mammalian. At some point this capacity must have been developed to include perceptual and conceptual features that are rather arbitrary from a non-human standpoint: physical features that demarcate an individual from a group and link it by resemblance to members of another group (skin/hair pigmentation etc.); or behavioral features that mark aspects such as social standing (via membership in a particular social group). All these human features rest on a feature of H. sapiens sapiens that must be older than human culture: an awful lot of our species' cortex is devoted to visual perception of facial features and expressions and other visual evidence to similarity and difference to our own group(s).
Self-awareness: some idea of the self as a distinct individual, and self-recognition (recognizing self in mirror as distinct from conspecifics) (these are apparently pre-hominid capacities). Was there a more recent development in self-awareness? Increasing degree thereof?
Recognition of specific individuals within a group of proximates; development of other recognition channels besides smell: via visual physical characteristics and especially, facial recognition. Development of cortical structure adapted to facial recognition. Link to memory capacities, leading to recognition and memory of large numbers of distinct faces.
Increasingly powerful memory, making possible the ability to store large numbers of cognitive categories, and later, symbols, i.e. cognitive categories associatively linked with some consistent form of expression. Modern human active (production) vocabularies have numbers of units in the high tens of thousands, and more for passive (recognition) vocabularies. Great apes show some basic primary categorization (distinguishing of kinds of things) in the wild and experimental settings, and in experimental settings with training, the ability to learn a relatively small number of lexical symbols (usual claims: anywhere from 60-300). Most researchers are not convinced that grammatical symbols are demonstrably acquired by non-human primates in primate symbolic acquisition experiments. But see Savage Rumbaugh's work on Kanzi. Lieberman claims that chimps DO acquire some simple syntax in the form of meaningful Bword order differentiations.
Increasing complexity of intentional planned goal-directed behavior, involving more complex sequences to reach final goal.
Greater capacity to develop more complex kinds of cognitive categories: various kinds of categories of increasing complexity, cf. the various types of linguistic categories postulated in Langacker's Cognitive Grammar. For example, were "thing"-type categories deployed first, before categorizations for kinds of properties and kinds of events? Did categories for entities with relatively obvious, concrete perceptual attributes precede categories .for conceptualizations more divorced from perceptual properties? E.g. an internal category/conception for 'banana' vs. a conception of 'not'?
Development of not only mental representations, but some external associated manifestation for those internal categories: symbols (whether gestural; vocal; or other); see above point on memory.
Development of complex kinds of lexical symbols: Were "thing"-type symbols first, or at least more prevalent first? Were symbols for more concrete entities first, and later abstract symbols? Cf. Linguistic development in modern human ontogeny (not an altogether simple story: see child language development research on concepts and concept acquisition: Mandler, Tomasello and associates, Clark, Bowerman).
Development of properties of modern lexical symbolic categorization: vagueness or "fuzzy boundaries" of symbols; prototype organization; pervasive polysemy; other non-one-to-one relations of labels to concepts such as homonymy.
Greater ability to manipulate symbols, and relate them into more complex constellations of thought. Development of basic event structures composed of a few participants and an event linking them. Development of more complex, grammatical symbols linking lexical symbols in syntagmatic hierarchies and arrays. Was there a a progressive increase in the complexity of grammatical symbols? What about the development of syntactic devices like subordination, relativization, etc.?
Greater capacity to detach cognitive processing from immediate sensory input and output (production). For example, at some point the ability to communicate something that is not spatially and temporally tied to an immediate strong emotion (desire, fear) must have developed. (Hockett's "displacement" feature, which he says only humans have, and bees to a very limited extent.) Non-human primates show extremely limited development in this respect compared to humans.
Ability to do "cognitive projection" and cognitive blending. Basis of parable (relating a simple, relatively concrete narrative sequence, to perceptually quite different situations in which a similar structure and correspondences among entities can be abstracted); metaphor; problem solving of various kinds. An important basis of creative symbol-use.
Understanding of conditionality: "if I do this/if X happens (which it might)..."
Increased understanding of causal relations. Simple spatial causal relations with two events in relatively immediate temporal contiguity are apparently understood by chimpanzees. Understanding of causation is a difficult topic. It's difficult to know what causal relations are grasped by non-human primates, as it is with precommunicative children. Children show some devlopmental processes in causal understanding. Complex causal relations, beyond spatio-temporally contiguous event relations, are difficult for even developmentally mature humans. Failure to agree on causal relations is a typical source of human strife in society.
Understanding of counterfactuality: "if I had done X (but I didn't)".../"if X had occurred (but it didn't), Y would have happened or not happened". Linguistic means for coding conditionality and counterfactuality are underdetermined by the cognitive understandings of these notions. In other words, humans understand and deploy these concepts without necessarily using unambiguous linguistic means for encoding them. Similarly with causality: the understanding (and communications of it) can be there without any explicit coding of causality.
The cognitive processing involved in discrimination of the finer grained sounds produced via above development. Perceptual discrimination adapted to language developed, presumably as an overlay to the existing auditory perceptual capacities (the last two points relate to the development of modern phonological organizations
The cognitive processing involved in handling complex relations among symbols: formal and functional groupings involving orderings and linkages that constitute some kind of syntax. Was there a progressive development in complexity here?
The cognitive processing allowing discourse tracking and complex sequences of related events that form an overarching narrative
An individual's recognition of others as selves like him or her; understanding of the possibility of a distinct viewpoint different from one's own. This is what philosophers call a theory of mind. I think of it as a mental model in one person's head of what might be going on in another person's head. Most simply, it can be an understanding that a person might not SEE all the things that you can see, or might see things that you can't. Very small children sometimes think that when they cover their eyes others can't see them! A more complicated version is realizing that a person might not know everything you do, or conversely, might know more.
Empathy with social proximates. Development of degrees of empathy with degrees of social proximity (the empathy hierarchy)
Altruistic behavior, including providing feeding/protection for weaker/sick/injured conspecifics beyond a mother's own young
The attentional triad linking two individuals and an external entity, and the pointing behavior that facilitates (as well as reflects) it (cf. work of Tomasello)
Increasing complexity of dyadic interactions: from non-human primates' simple sensitivity to when the other is looking at them (thus a potential recipient of a gestural communication), and ability to invite reciprocality with shared attention, to modern humans give and take of getting the floor, holding the floor, giving up the floor to the other, creating an easy interactional flow with many layers of discourse structure. Human infants seem to understand the back-and-forth interactive nature of communication well before they can speak; the pre-linguistic infant and mother sometimes participate in interactions with appropriate intonation sounding quite like an actual linguistic exchange, even when the child is only cooing or babbling.
Development of another kind of increasingly complex interactional skills: behavior sensitive to social hierarchies that depends on increased memory of past behaviors of other individuals. (Possibly) some development of a predictive capacity for anticipating reaction of the other individual to the behavior of an acting individual (relates to understanding of causality/force dynamics, but on the interpersonal plane)
Learned patterns of communicative behavior: development of conventions, including shortening/ellipsis that moves a symbol away from obvious iconic motivation (auditory iconicity; gestural iconicity; other vocal iconicity?) Non- human primates show shortening/ellipsis in deontic (command mode) communications
Communications that are representative (descriptive), not just manipulative (imperative). Non-human primate communications are primarily manipulative, although in experimental settings the primates can learn to do descriptive communications. (They also have expressive vocalizations, but these appear to not be intentional, just automatic reactions). Humans seem to start out with primarily expressive and manipulative communications in infancy, and later start descriptive communication (beginning with simple naming)
Social transmission of learned communicative behaviors (e.g. use of symbols) to young of the species (not by active teaching but by the young observing, imitating, generalizing)
Attunement to intentionality and goals of others; picking up on what causal consequence is intended by another individual. Increased understanding of intentionality might have occurred during the evolutionary process, much as there is development during ontogeny in modern humans (?)
Development of cooperative behavior: 'helping' (first ad hoc in isolated situations, then learned as a practice?). Helping an individual by protecting, vs. helping by recognizing goal of other and participating in sequence of activity toward that goal (e.g. an individual helps another reach a piece of food it can't get; helps gather; helps make). At what point did hominids acquire an understanding of an implied reciprocity of helping?
Further cooperative behavior: extension of planning, goals from individual to shared, social realm: dyads or (perhaps later?) group plans and goals
Ability to deceive. Great apes supposedly have some level of deceptive capacity, but when it comes to vocalization, they can't control it very well. For example they cannot keep from producing vocalizations that reveal to conspecifics that they see something good to eat, even when it would be in their interest to do so.
Development of motor musculature for new bipedal gait. (Goes back to point of common ancestor of human ancestor line with australopithecines. See first point under neuro-cortical development. Some believe this, along with the neurological structures designed for controlling and sequencing these movements, is a prerequisite for or at least facilitates development of, the following skills.)
Increasing size of brain pan and brain inside it; increase in convolution of cortex leading to huge increase in cortical capacity, ultimately overcoming physical limitation on brain size/skull size due to dimensions of female pelvis. Hard to know when convolution took place, as it is all soft tissue
Development of motor musculature for throwing, aiming
Development of motor musculature for finely sequenced manual control, which can underlie (1) tool-making and (2) gesture
Development of motor musculature in the vocal tract that goes with the neural circuitry involved in controlling that musculature. Enlarged thoracic region of spinal chord is locus of most of these muscles
Modification of the vocal tract and pharynx so that an appropriately resonant chamber is present for speech sounds of sufficient discriminability. Accomplished at least in part by an extreme lowering of the larynx (anti-adaptive for feeding, but adaptive for communication). Development to final configuration was late, per Philip Lieberman and Michael Corballis
Development of genes controlling the prenatal and postnatal development of some of the skills (precisely which?), e.g. genes for getting the musculature and associated cortex in place for when child begins to speak. Presumably at some point behaviorally driven changes leading to enhanced skills must have come to be encoded in some degree genetically. Even if there is no "language box" or "language organ" that genes instruct the brain to form during childhood, still the cortical and other development distinguishing us from other primates has some genetic substrate that makes a human brain develop in a baby, rather than a chimp brain. Human brains are different from chimp brains structurally and functionally, but many of the differences have to do with the cognitive and social capacities described above that go way beyond language. In other words, while some capacities underlying language may have an innate basis, "language" itself as a structural or functional whole does not.