Evolution of abstract thought

Abstract thought, broadly defined as the ability to mentally represent and manipulate concepts detached from immediate sensory experience, is among the most distinctive cognitive capacities of Homo sapiens. It encompasses theory of mind (the ability to attribute mental states to others), symbolic reasoning (the capacity to use arbitrary signs to represent objects, events, and relationships), mental time travel (the projection of oneself into past and future scenarios), and analogical thinking (the recognition of structural similarities between different domains).^{5, 13} Understanding how and when these capacities evolved is one of the central challenges of human origins research, complicated by the fact that cognition leaves no direct fossil record and must be inferred from archaeological proxies, comparative studies of living primates, and the neuroscience of the human brain.^{4, 9}

Comparative cognition: humans and other primates

Great apes, particularly chimpanzees and bonobos, share many cognitive abilities with humans, including basic tool use, social learning, numerical competence, and rudimentary planning.⁷ The question of whether non-human primates possess theory of mind has been debated since Premack and Woodruff posed it in their landmark 1978 paper. Subsequent decades of experimental work have established that chimpanzees can track what other individuals know (for example, whether a competitor has seen where food is hidden) but appear unable to represent others' false beliefs, the classic hallmark of full theory of mind in human developmental psychology.^{5, 6} Kaminski, Call, and Tomasello's 2008 experiments demonstrated that chimpanzees understand what conspecifics have and have not seen, but do not attribute beliefs that differ from reality, suggesting that their social cognition operates on a "knowledge-ignorance" system rather than a full belief-desire framework.⁶

The gap between human and non-human primate cognition widens dramatically in the domain of symbolic representation. While apes trained in captivity can learn to associate arbitrary tokens with objects or actions, no wild great ape population has been observed to spontaneously create or use symbols.⁷ Deacon argued in The Symbolic Species that the critical difference is not the ability to learn individual symbol-referent associations but the capacity for symbolic reference: the understanding that symbols form a system of relationships where meaning derives from the relations among symbols themselves rather than from direct connections to the physical world.¹⁰ This capacity for relational, hierarchical, and recursive representation, which Deacon linked to the expansion of the prefrontal cortex, underlies uniquely human cognitive feats including grammar, mathematics, and moral reasoning.¹⁰

Neural substrates

Comparative neuroanatomy provides clues about the brain changes underlying abstract thought. The human prefrontal cortex is not simply larger than that of other primates in absolute terms; it is disproportionately larger relative to the rest of the brain, and its internal connectivity is qualitatively different.¹¹ Schoenemann, Sheehan, and Glotzer demonstrated in 2005 that prefrontal white matter volume, reflecting the density of long-range axonal connections between the prefrontal cortex and other brain regions, is disproportionately expanded in humans compared to other primates, even after controlling for total brain size.¹² This enhanced connectivity is thought to underlie the human capacity for working memory, the ability to hold and manipulate multiple pieces of information simultaneously, which Coolidge and Wynn identified as the cognitive faculty most directly responsible for the gap between human and non-human primate abstract reasoning.⁴

Working memory, as described in Baddeley's influential model, comprises a central executive system that coordinates information from multiple subsidiary systems including a phonological loop (for verbal information) and a visuospatial sketchpad (for spatial and imagistic information).³ Coolidge and Wynn argued that the enhancement of this central executive capacity, perhaps through genetic changes affecting prefrontal cortex development, was the key innovation that enabled modern human cognition, including the ability to plan multi-step tool manufacture, to hold social obligations in mind during complex interactions, and to reason about hypothetical and counterfactual scenarios.⁴ Molecular genetic studies have identified several genes showing evidence of positive selection in the human lineage that affect brain development, including ASPM, Microcephalin, and ARHGAP11B, the latter of which has been shown experimentally to promote basal progenitor amplification and neocortex expansion when introduced into mouse and ferret embryos.^{14, 15}

Archaeological proxies for abstract thought

Because thoughts do not fossilize, researchers rely on archaeological evidence that implies symbolic or abstract cognition. The most widely cited proxies include the production of personal ornaments (which require shared understanding of symbolic meaning within a social group), representational art (which requires the mental capacity to depict absent objects or beings), abstract geometric engravings (which demonstrate pattern recognition and intentional design), and composite tool manufacture (which requires planning multiple steps in advance and mentally integrating components that have no function individually).^{8, 9}

The African Middle Stone Age provides the earliest evidence for several of these behaviors. Engraved ochre pieces from Blombos Cave, South Africa, dated to approximately 77,000 years ago, display cross-hatched geometric patterns that are clearly intentional and non-utilitarian, providing the earliest unambiguous evidence for abstract mark-making by Homo sapiens.⁸ Perforated marine shell beads from Blombos, Skhul (Israel), and Grotte des Pigeons (Morocco), dating between 80,000 and 100,000 years ago, represent the earliest known personal ornaments, artifacts whose function is entirely symbolic since they serve no subsistence or technological purpose.⁹ Middle Stone Age composite tools, including hafted stone points requiring the manufacture of separate components and their assembly with adhesives, appear in Africa by at least 71,000 years ago and demonstrate multi-step planning and hierarchical reasoning.⁹

The temporal distribution of these proxies is uneven, with symbolic behaviors appearing, disappearing, and reappearing in the African record over tens of thousands of years, a pattern that has been interpreted as evidence either for demographic instability (small populations losing cultural innovations) or for the gradual, geographically patchy accumulation of cognitive capacities rather than a single moment of cognitive breakthrough.^{1, 9}

Social cognition and the social brain

The social brain hypothesis, articulated by Robin Dunbar, proposes that the primary selective pressure driving the evolution of large brains and complex cognition in primates was not ecological problem-solving but the demands of navigating complex social relationships.¹ In humans, social cognition requires not only tracking the behavior and knowledge states of multiple individuals but also reasoning about their beliefs, intentions, and emotional states, forming mental models of social relationships, and predicting how others will respond to one's own actions.^{1, 2} These demands place enormous loads on working memory and executive function and may have been the primary driver of prefrontal cortex expansion in the hominin lineage.^{2, 4}

Dunbar and Shultz's comparative analyses have demonstrated that across primates, neocortex ratio (the size of the neocortex relative to the rest of the brain) correlates strongly with social group size, and that this relationship is specific to anthropoid primates, suggesting that social complexity has been a uniquely important driver of brain evolution in the human lineage.² The extension of this argument to abstract thought proposes that the cognitive machinery originally evolved for social reasoning, including the ability to represent mental states, to reason recursively ("I know that you know that I know"), and to maintain large numbers of social relationships, was subsequently co-opted for other domains of abstract thought including mathematics, narrative, and philosophical reasoning.^{1, 10}

Unresolved questions

Several fundamental questions about the evolution of abstract thought remain open. The timing of the critical cognitive transition is uncertain: if the neural architecture for modern abstract thought was present by 100,000 years ago or earlier, as the African archaeological record suggests, then why do some markers of symbolic behavior appear only sporadically in the Middle Stone Age before becoming ubiquitous in the Upper Paleolithic and Later Stone Age?^{9, 16} Possible explanations include demographic thresholds (symbolic behavior requires a minimum population density to be maintained across generations), ecological triggers (specific environmental pressures that favored symbolic communication), and cultural ratcheting (the gradual accumulation of cultural practices that, once established, create the conditions for further innovation).^{4, 9}

The relationship between language and abstract thought presents another unresolved problem. Some researchers argue that language is a prerequisite for symbolic thought, since the manipulation of abstract concepts depends on having an arbitrary sign system in which to encode them.¹⁰ Others contend that abstract thought can operate independently of language, pointing to evidence from spatial reasoning, musical cognition, and mathematical intuition that appears to engage non-linguistic neural systems.³ The question of whether Neanderthals and other archaic hominins possessed some form of abstract cognition adds another dimension to the problem: if Neanderthals produced symbolic artifacts such as cave paintings and personal ornaments, as some evidence suggests, then the neural substrates for abstract thought may have been present in the last common ancestor of modern humans and Neanderthals, predating the divergence of the two lineages by at least 500,000 years.^{14, 16}