How Function Shapes Form
– by Murillo Pagnotta
Symbolic artefacts as tools of the mind
In the study of human evolution, evidence of symbolic behaviour is often taken to indicate when the mind of our deep past relatives became equivalent to ours. While language is the ultimate yardstick of symbolic behaviour, in a broad sense this also includes other meaning-making practices such as drawing, sculpting, making ornaments, singing, and dancing. And while language does not fossilise, the production of materials laden with symbolic meaning can leave traces in the archaeological record.
Some of the oldest evidence of symbolic behaviour come from two well-known archaeological sites in South Africa. These are the engraved markings found in the ochre pieces from the Blombos Cave and the ostrich eggshell pieces from the Diepkloof Rock Shelter. These materials date back as far as ~100 thousand years ago and hint at a practice spanning tens of thousands of years. The engravings are widely interpreted as intentional markings. They are certainly not accidental—but what function did they actually serve? This has been a matter of debate among archaeologists and three main hypotheses have been proposed: the markings served i) a decorative purpose, ii) as markers of group identity, or iii) as denotational symbols (similar to a proto-writing system).
The basic motivation for our recent article published in Evolution and Human Behaviour is to generate experimental evidence for or against these competing hypotheses. Looking closely at the engraved patterns, they appear to have changed over time: earlier (older) patterns are simpler configurations of roughly parallel lines, while later (more recent) ones show complex patterns such as cross-hatching. We believe the structural properties of these engravings—and especially how they changed over time—might provide clues about their function. This work is actually the third in a series, so let’s view it in its proper context.
The cognitive profile of the Blombos and Diepkloof engravings.
In the first study in the series, the late Kristian Tylén and colleagues (2020) argued that symbolic artefacts such as the Blombos and Diepkloof engravings might be seen as ‘tools of the mind’—they can shape our thinking. The analogy is with instrumental tools, such as the stone cleavers, handaxes, and flakes our deep past relatives produced. Instrumental tools act on the physical environment and are often refined over time to better serve their function—stone flakes, for instance, became increasingly efficient cutting tools, and their structure reveal their basic function. Symbolic artefacts, by contrast, primarily act on perceptual and cognitive processes, facilitating reasoning, communication, and aesthetics in senders and receivers—and like instrumental tools, they too may be incrementally refined as they are repeatedly produced for the same purpose across generations. Our underlying hypothesis is that structural changes in symbolic artefacts can likewise reveal their function, all with cognitive implications that we can measure in the lab.
How can structural changes inform us about function? Tylén and colleagues hypothesised that each candidate function should map onto a distinct cognitive profile: decoration (hypothesis 1) predicts later patterns becoming more salient and purposeful-looking; group identity (hypothesis 2) predicts them becoming easier to remember and more stylistically distinct; denotational communication (hypothesis 3) predicts them becoming easier to discriminate within a set, so as to carry specific meanings efficiently.
To test these predictions, they designed five experiments measuring these cognitive implications: saliency, discriminability, style, intentionality, and memorability using drawings derived from early, intermediate, and late engravings as stimuli. They found that, over time, the engravings became more salient, more stylistically distinct, more intentional-looking, and easier to remember and reproduce—but no easier to discriminate within their own site and period. This cognitive profile points to decoration and/or group identity marking (hypotheses 1 and 2), but not denotational communication (hypothesis 3).
The cognitive profile of experimental transmission chain drawings—and back to the engravings.
This foundational analysis assumed that functional context constrains how artefacts change over time, with different functions producing distinct trajectories of structural change—and thus different cognitive profiles. Our two subsequent studies tested this assumption directly. Both studies followed a two-part approach. First, a transmission chain methodology generated lineages of related drawings: participants were assigned a position or ‘generation’ in a chain, shown drawings from the previous generation, and asked to reproduce them digitally after a short delay (the first participant in a chain saw drawings derived from the early engravings). In the second part of the study, a subset of these transmission chain drawings was used as stimuli in the same five cognitive experiments.
Our two transmission chain experiments different in the following way: One study implemented a simple delayed reproduction task to provide a baseline (Pagnotta, Tylén et al. 2026). We found that the drawings became increasingly regular, and this had cognitive consequences: they became easier to discriminate, looked more intentional, and became easier to remember—but their saliency did not change, and they became less stylistically distinct between chains. These changes appear to be shaped by working memory constraints, perceptual grouping, and other basic aspects of visual perception. In our latest study, we embedded the reproduction task within three functional contexts operationalising the three hypotheses: decoration, identity-marking, and denotative communication (Pagnotta, Wisher et al. 2026).
We found that different contexts indeed led to different cognitive profiles. For example, discriminability increased only under the denotative communication condition. Comparing the cognitive profiles of the transmission chain drawings with those of the original engravings, we found matching patterns suggesting that the original engravings were probably used for decorative or identity-marking functions—or, not inconceivably, a mix of these—but not denotative communication. These findings corroborate the conclusions from Tylén et al (2020).
In sum, these studies contribute to a growing corpus of experimental and quantitative evolutionary semiotics by showcasing how cognitive scientists and archaeologists can collaborate and use experiments to test competing hypotheses about ancient symbolic artefacts and the cognitive capacities they index.
This study was funded by the EU ERC (eSYMb, 101044626) and the Interacting Minds Centre (IMC) at Aarhus University.
References
Tylén, K., Fusaroli, R., Rojo, S., Heimann, K., Fay, N., Johannsen, N. N., Riede, F., & Lombard, M. (2020). The evolution of early symbolic behavior in Homo sapiens. Proceedings of the National Academy of Sciences, 117(9), 4578–4584. https://doi.org /10.1073/pnas.1910880117
Pagnotta, M., Tylén, K., Qvist, A.S., Kjeldsen, R.F., Rojo, S., Heimann, K., Fay, N., Johannsen, N.N., Riede, F., Lombard, M. and Fusaroli, R. (2026), Simulating Symbolic Evolution in the Lab: Potentials and Implications of Using Transmission Chains to Study Early Symbolic Behavior at the Emergence of Homo sapiens. Topics of Cognitive Science, 18: e70036. https://doi.org /10.1111/tops.70036
Pagnotta, M., Wisher, I., Petersen, M. L., Riede, F., Tylén, K., & Fusaroli, R. (2026). The evolution of symbolic artefacts: How function shapes form. Evolution and Human Behavior, 47(5), 106902. https://doi.org/10.1016/j.evolhumbehav.2026.106902
