The job of archaeologists is often signal detection, that is, to correctly judge the presence or absence of a target phenomenon based on noisy archaeological data. As such, two types of errors are possible: False positives and false negatives. In his article, Stibbard-Hawkes vividly illustrates the threat of one type of false-negative reasoning for symbolic archaeology, providing strong empirical support for the view that, in Moffett (Reference Moffett2013)'s words, “many of the symbols employed by recent hunter-gatherers would leave little or no archaeological signature.” In this commentary, we extend Stibbard-Hawkes' analysis by elaborating on the corresponding problem of false-positive reasoning in the symbolic context. Specifically, we focus on the problem of false symbols – intuitively special entities that are judged to be symbolic in some broad sense, but which are not. The false-symbol problem is distinct from the more familiar one of interpreting possible indirect evidence of symbolic behaviour, for example, ochre processing at certain Middle Pleistocene sites. There, the issue is that the evidence might have been produced via non-symbolic, utilitarian behaviours (e.g., skin protection). In contrast, as we explain below, false symbols generally have no clear-cut connection to utilitarian behaviour, yet they are still non-symbolic. False symbols are prototypically both non-utilitarian and non-symbolic.
The false-symbol problem is thrown into sharp relief by cultural evolution work in comparative cognition. Non-human great apes (“apes,” hereafter) can individually develop certain utilitarian behaviours that at first seem to require cultural access, plus human-like types of cultural transmission. However, as Stibbard-Hawkes acknowledges, the fact that these behaviours can develop individually (in at least some individuals) makes them lie instead within their zone of latent solutions (Tennie, Braun, Premo, & McPherron, Reference Tennie, Braun, Premo, McPherron, Haidle, Conard and Bolus2016). Latent utilitarian ape behaviours are strong candidates for latent utilitarian behaviours in all hominins (ibid.; see also Reindl, Beck, Apperly, & Tennie, Reference Reindl, Beck, Apperly and Tennie2016; Tennie, Reference Tennie, Tehrani, Kendal and Kendal2023; Tennie, Premo, Braun, & McPherron, Reference Tennie, Premo, Braun and McPherron2017). But our main point here is that such latent behavioural capabilities exist outside the utilitarian domain, too, and it is especially these abilities that can give rise to outcomes that become instances of false-symbol interpretations.
Given primates' close phylogenetic proximity to humans, “primate art” is a prime example of how this issue may play out. Here, we are uninterested in the monetary or aesthetic value which humans attach to such items (which may be considerable). Nor are we interested in attempting to define “art.” Instead, we simply note that there are things primates make which humans routinely label “art,” and we note that art is generally linked to symbolic cognition by archaeologists (Stibbard-Hawkes included). What significance do these items have for the symbolic interpretation of similar items found in the archaeological record (regardless of their specific medium)?
A good start on this question can be made through a careful re-analysis of the extant literature on primate art (a task we are currently undertaking). Such a re-analysis must filter evidence in three ways. First, some of the primates involved were human enculturated. As is widely acknowledged, such enculturation can push primate minds and/or behaviour generally in the direction of the human range (e.g., Kanzi the bonobo). Second, some of the primates were trained to perform various “artmaking” tasks. Third, some had artmaking demonstrated for them by humans (e.g., how, where, when, etc., to make art). Strictly speaking, only cases involving unenculturated, untrained, and demonstration-naïve primates (or other animals, e.g., elephants) are relevant. The simple reason is that humans did not (and we suspect, will not) use time-machines to “go back” and enculturate, train, or demonstrate art for their Pleistocene relatives.
Sadly, the vast majority of cases of primate art prove irrelevant by these three standards. We plan to conduct work soon that avoids these confounds. But for now, a single proof-of-principle case is enough to show our approach holds water. Fortunately, there would appear to be at least one such case: In the 1990s, Westergaard and Suomi (Reference Westergaard and Suomi1997) provided capuchin monkeys with clay, paint, and various other materials (sticks, stones, etc.). These monkeys were unenculturated, untrained (personal communication), and the paper mentions no human modelling. These relevant monkeys nevertheless spontaneously made non-utilitarian marks – in a variety of ways – including with paint, and including parallel scratch and even chisel marks. By “non-utilitarian,” we here mean that the marks were made for no apparent practical purpose, nor as a by-product of some activity done for a practical purpose, for example, cutting food. An onlooker may well have concluded the capuchins were making art, and that they were creating “symbolic material culture.” However, most would deny that (untrained, unenculturated, naïve) capuchins produce and use symbols in any interesting sense. Yet, such capuchins clearly produce outcomes (markings) on par with outcomes that many archaeologists treat as symptomatic or at least suggestive of symbolic cognition and culture. Were we to discover principally similar markings (e.g., non-iconic etchings on a cave wall) known to be made by hominins, they would likely be interpreted as symbolic (see, e.g., Fuentes et al., Reference Fuentes, Kissel, Spikins, Molopyane, Hawks and Berger2023 on potential Homo naledi engravings).
The main point is this: The presence of non-utilitarian markings made by hominins at an archaeological site is therefore not in itself a safe means of inferring the presence of human-like symbolic cognition (or human-like culture). To show a symbolic origin, more data are needed. The simple reason is that at least some primates are apt to spontaneously produce such outcomes, despite (likely) lacking symbolic capacities in anything like the sense archaeologists seem interested in when they speak of “symbolism.” Thus, in addition to false negatives, archaeology faces a false-positive problem in regard to symbol identification. Although the symbolic status of (e.g.) representational cave art cannot be seriously doubted, that is not true in every case of, say, engraved zigzag designs, parallel lines, or finger flutings.
You have
Access
The job of archaeologists is often signal detection, that is, to correctly judge the presence or absence of a target phenomenon based on noisy archaeological data. As such, two types of errors are possible: False positives and false negatives. In his article, Stibbard-Hawkes vividly illustrates the threat of one type of false-negative reasoning for symbolic archaeology, providing strong empirical support for the view that, in Moffett (Reference Moffett2013)'s words, “many of the symbols employed by recent hunter-gatherers would leave little or no archaeological signature.” In this commentary, we extend Stibbard-Hawkes' analysis by elaborating on the corresponding problem of false-positive reasoning in the symbolic context. Specifically, we focus on the problem of false symbols – intuitively special entities that are judged to be symbolic in some broad sense, but which are not. The false-symbol problem is distinct from the more familiar one of interpreting possible indirect evidence of symbolic behaviour, for example, ochre processing at certain Middle Pleistocene sites. There, the issue is that the evidence might have been produced via non-symbolic, utilitarian behaviours (e.g., skin protection). In contrast, as we explain below, false symbols generally have no clear-cut connection to utilitarian behaviour, yet they are still non-symbolic. False symbols are prototypically both non-utilitarian and non-symbolic.
The false-symbol problem is thrown into sharp relief by cultural evolution work in comparative cognition. Non-human great apes (“apes,” hereafter) can individually develop certain utilitarian behaviours that at first seem to require cultural access, plus human-like types of cultural transmission. However, as Stibbard-Hawkes acknowledges, the fact that these behaviours can develop individually (in at least some individuals) makes them lie instead within their zone of latent solutions (Tennie, Braun, Premo, & McPherron, Reference Tennie, Braun, Premo, McPherron, Haidle, Conard and Bolus2016). Latent utilitarian ape behaviours are strong candidates for latent utilitarian behaviours in all hominins (ibid.; see also Reindl, Beck, Apperly, & Tennie, Reference Reindl, Beck, Apperly and Tennie2016; Tennie, Reference Tennie, Tehrani, Kendal and Kendal2023; Tennie, Premo, Braun, & McPherron, Reference Tennie, Premo, Braun and McPherron2017). But our main point here is that such latent behavioural capabilities exist outside the utilitarian domain, too, and it is especially these abilities that can give rise to outcomes that become instances of false-symbol interpretations.
Given primates' close phylogenetic proximity to humans, “primate art” is a prime example of how this issue may play out. Here, we are uninterested in the monetary or aesthetic value which humans attach to such items (which may be considerable). Nor are we interested in attempting to define “art.” Instead, we simply note that there are things primates make which humans routinely label “art,” and we note that art is generally linked to symbolic cognition by archaeologists (Stibbard-Hawkes included). What significance do these items have for the symbolic interpretation of similar items found in the archaeological record (regardless of their specific medium)?
A good start on this question can be made through a careful re-analysis of the extant literature on primate art (a task we are currently undertaking). Such a re-analysis must filter evidence in three ways. First, some of the primates involved were human enculturated. As is widely acknowledged, such enculturation can push primate minds and/or behaviour generally in the direction of the human range (e.g., Kanzi the bonobo). Second, some of the primates were trained to perform various “artmaking” tasks. Third, some had artmaking demonstrated for them by humans (e.g., how, where, when, etc., to make art). Strictly speaking, only cases involving unenculturated, untrained, and demonstration-naïve primates (or other animals, e.g., elephants) are relevant. The simple reason is that humans did not (and we suspect, will not) use time-machines to “go back” and enculturate, train, or demonstrate art for their Pleistocene relatives.
Sadly, the vast majority of cases of primate art prove irrelevant by these three standards. We plan to conduct work soon that avoids these confounds. But for now, a single proof-of-principle case is enough to show our approach holds water. Fortunately, there would appear to be at least one such case: In the 1990s, Westergaard and Suomi (Reference Westergaard and Suomi1997) provided capuchin monkeys with clay, paint, and various other materials (sticks, stones, etc.). These monkeys were unenculturated, untrained (personal communication), and the paper mentions no human modelling. These relevant monkeys nevertheless spontaneously made non-utilitarian marks – in a variety of ways – including with paint, and including parallel scratch and even chisel marks. By “non-utilitarian,” we here mean that the marks were made for no apparent practical purpose, nor as a by-product of some activity done for a practical purpose, for example, cutting food. An onlooker may well have concluded the capuchins were making art, and that they were creating “symbolic material culture.” However, most would deny that (untrained, unenculturated, naïve) capuchins produce and use symbols in any interesting sense. Yet, such capuchins clearly produce outcomes (markings) on par with outcomes that many archaeologists treat as symptomatic or at least suggestive of symbolic cognition and culture. Were we to discover principally similar markings (e.g., non-iconic etchings on a cave wall) known to be made by hominins, they would likely be interpreted as symbolic (see, e.g., Fuentes et al., Reference Fuentes, Kissel, Spikins, Molopyane, Hawks and Berger2023 on potential Homo naledi engravings).
The main point is this: The presence of non-utilitarian markings made by hominins at an archaeological site is therefore not in itself a safe means of inferring the presence of human-like symbolic cognition (or human-like culture). To show a symbolic origin, more data are needed. The simple reason is that at least some primates are apt to spontaneously produce such outcomes, despite (likely) lacking symbolic capacities in anything like the sense archaeologists seem interested in when they speak of “symbolism.” Thus, in addition to false negatives, archaeology faces a false-positive problem in regard to symbol identification. Although the symbolic status of (e.g.) representational cave art cannot be seriously doubted, that is not true in every case of, say, engraved zigzag designs, parallel lines, or finger flutings.
Financial support
Ronald J. Planer and Claudio Tennie gratefully acknowledge research support received from the DFG-Centre for Advanced Studies in the Humanities Words, Bones, Genes, Tools (DFG-KFG 2237).
Competing interest
None.