The cognitive constitution of symbols
The use of symbols is considered among the decisive accomplishments separating humans from other species (Deacon, 1998). Symbols are tools of the mind. While we apply instrumental tools to make changes to our material surroundings, the use of symbols (material or virtual) changes the way we think (Bjørndahl, Fusaroli, Østergaard, & Tylén, 2014; Clark, 2006b; Roepstorff, 2008). This happens in contexts of communication where they facilitate interpersonal sharing of information, in contexts of reasoning where symbols aid our cognitive processes, or in contexts of aesthetic activity where symbols induce meaningful and pleasurable experiences. And just like instrumental tools and technologies have evolved dramatically over the last 150.000 years, the development of ever more sophisticated symbols (and symbol systems) has enabled unprecedented human cognitive abilities: unlimited external memory stores (cf. books, libraries, and the internet), cumulative scientific insight, math, religion, narrative, art and literature, as well as complex social organization and institutions (Clark, 2006a, 2006b; Latour, 1996; K. Tylén & McGraw, 2014).
While the functional efficacy of instrumental tools such as flake knives or spear heads depends directly on aspects of their physical properties and affordances (Gibson, 1977; Greeno, 1994), symbolic artefacts do their work by engaging mental processes (henceforth cognitive affordances) of perception, attention, learning, memory, and representation (Clark, 2006b; Kirsh & Maglio, 1994; Roepstorff, 2008; K. Tylén & McGraw, 2014). For instance, an abstract geometrical symbol may be associated with significance in non-trivial ways, depending on its engagement in contexts of decoration, communication or reasoning. To uncover the symbolic function of a past symbolic artefact, we thus need to approach the cognitive processes they elicit in contextualized practices of use: by systematically testing the potential of symbolic forms from different periods to engage particular cognitive affordances, we can make systematic inferences about their plausible symbolic functions.
An integrative approach to the study of symbolic evolution
The objective of eSYMb is to develop a data-driven, systematic and scientific approach to inform the interpretation of early human symbolic behavior and how it evolved during the late Middle Paleolithic as a precursor of later conventionalized practices of representation, communication and language. This is achieved by directly engaging symbolic artefacts as stimuli in experimental studies to elucidate their potential to engage processes relevant for aspects of symbolic cognition. Conceptually, myapproach is grounded on the assumption that symbolic behaviour unfolds as an interaction between three interdependent levels of organization: the artefacts’ 1) symbolic form, 2) their cognitive affordances, and 3) their pragmatic use. While such stratifications resonate with previous theoretical work in semiotics, psychology, and cognitive science (Clark, 2006b; Gibson, 1977; Hutchins, 2010; Kissel & Fuentes, 2017), the primary strength of the model is the way it allows me to operationalize novel empirical approaches to the investigation of symbolic behaviour.
Symbolic forms are directly accessible from the artefacts as geometrical or figurative motifs, and can be quantitatively approached with structural analysis (entropy, complexity, symmetry, etc). An artefact’s cognitive affordances, however, are thecognitive implications of these forms on processes of attention, discrimination, memory, and representation, and are only accessible through experimental and psychophysical methods. Last, the pragmatic use concerns the kind of symbolic work that the artefacts were engaged and purposefully crafted to do, that is, their symbolic function. We cannot directly observe or measure the past symbolic function of an artefact. However, the temporal development of artefacts can, to the extent that they are accessible from the archaeological record, hold important clues. Based on the assumption that symbolic artefacts evolve adaptively over time to better fulfil their functions, we can investigate these structural changes and their cognitive implications to inform inferences about their past use. For instance, if we observe a collection of symbolic artefacts to evolve to become easier to reproduce from memory with high fidelity, we might infer these affordances to have been adaptive for their past use. That is, the artefact could be part of a practice of reproducing the same forms. And these observations can, in turn, inform interpretations of their particular symbolic function.
Hypotheses
Three core hypotheses guide eSYMb:
From these general hypotheses, the project will generate specific predictions for particular archaeological cases and materials. These are repositories of recently published archaeological findings for which we have an extended chronological sequence: a series of artefacts from different periods in time that is an expression of a continuous symbolic practice at the site. The investigations will consist of three complementary and integrated lines of experimental and computational modeling work outlined in the following sections.
Experimental suite 1: cognitive affordances
In this line of research, archaeological findings (i.e. photographic depictions/outlines) are used as stimuli in behavioral and psychophysical experiments in order to investigate their cognitive affordances and how they change over time, to test hypotheses concerning their past symbolic functions (N. Fay et al., 2008). In previous work, we have laid important methodological groundwork, which has established methods for measuring a number of cognitive parameters relevant to the symbolic function of artefacts. These include experimental paradigms to test the relative visual saliency, perceived intentionality, discriminability, style, memorability, and reproducibility of symbolic artifacts (Kristian Tylén et al., 2020). This repertoire will be further developed, for instance, with the addition of improved measures for uncovering the potential denotational nature of some symbolic artefacts (Mellet et al., 2019; Yu, Pöppel, & Bao, 2020). By observing particular profiles of changes in the cognitive affordances of artefacts over time, we will test hypotheses about their likely past function, for instance, as aesthetic decorations, signifiers of group identity, tools for reasoning, or vehicles for communicative meanings.
Experimental suite 2: mechanisms of change
In this line of research, we will address the underlying mechanisms driving the evolution of particular symbolic developments. A number of such mechanisms pertaining to aspects of cognitive biases, individual and social learning, interaction and feedback has been suggested to have shaped past symbolic practices in an incremental and cumulative way (Caldwell & Millen, 2008b; Christiansen & Chater, 2008; Dean, Kendal, Schapiro, Thierry, & Laland, 2012; Nicolas Fay et al., 2018; Kirby et al., 2008). By seeding early symbolic artefacts from the archaeological records into transmission chain experiments subject to different contrastive mechanisms of learning, interaction and symbol use, we will record structural changes to the artefacts and thus test which set of mechanisms provide changes more similar to the historical-archaeological record (Morgan et al., 2015).
Experimental suite 3: symbolic forms
Recent developments in scientific approaches to cultural evolution has utilized a number of structural measurements to descriptively map changes to artifacts in terms of their structure, including, for instance, measures of entropy, compression, complexity, and symmetry (Morin, 2018; Morin & Miton, 2018; Sigaki, Perc, & Ribeiro, 2018; Tamariz & Kirby, 2015; Tran, Waring, Atmaca, & Beheim, 2021). It has, however, not previously been investigated how these structural properties relate to cognitive affordances (e.g. saliency, memorability, or discriminability) of symbolic artefacts. This third experimental series explore systematic and generalizable relationships both between cognitive measures, and between the structural traits of artifacts and their cognitive implications. Using computational modelling (neural networks), the goal is to investigate how structural and cognitive properties functionally interact to support the symbolic function of artefacts and their generalizability between studies (Fan, Hawkins, Wu, & Goodman, 2020).
