The [new] perspective asks us to construct, simultaneously, with the help of available facts, a view from the past, and from within. (Damasio, 2012, 16)

Objective science means straightforwardly taking the world as one’s subject matter; humanistic science means taking the world as one’s subject matter as the world of the subjectivity which functions for it, the world insofar as it is subject-related. (Husserl, 1970, 326)

1 Introduction

That the concept of mind may have some important role to play in understanding the lives, behavior, and cognition of prehistoric hominins is not a particularly controversial idea. Indeed, cognitive archaeologists talk frequently about minds, and some of the best known works in the field refer directly to minds in the title (e.g. Donald’s (1991) Origins of the Modern Mind, Mithen’s (1996) The Prehistory of the Mind, Lewis-Williams’ (2002) The Mind in the Cave, and Malafouris’ (2013) How Things Shape the Mind).

In most of these cases, however, the concept of mind arguably does not denote anything distinctive; it rather appears to function as a fancy synonym for cognition. On closer inspection, it seems instead that a skeptical attitude toward minds prevails in recent cognitive-archaeological literature, with any more substantial notion of it being dismissed offhandedly as Cartesian, dualistic, or internalistic. Surprisingly perhaps, in related fields, such as evolutionary psychology and primatology, talk of minds plays a more notable role. Tomasello, for example, frequently speaks of non-human great apes as “reading other minds”, as understanding “intentional states of others’’, and of their “mental processing” (Tomasello, 2021, 49–64). Talk of inner states also figures prominently in the literature on emotional contagion and empathy (e.g. Keysers et al., 2022), among so many other places.

What is it that cognitive archaeologists and other prehistoryFootnote 1 researchers allude to when they invoke concepts of the mind and the mental? Is there anything distinctive about minds, vis-à-vis the material structure of brains or other, “extended” vehicles, that they should be heedful of, or is it safe to equate talk of minds with talk of brains? Would it be better to abandon talk of minds and the use of other mental concepts, or is there something which an analysis of minds, most notably as elaborated in the philosophical tradition called phenomenology, could contribute to debates on prehistory?

One cannot begin to appreciate the positive role mind-talk -could play without some notion of what minds are. In this paper, I discuss such a notion, which can be of use in debates on (past) primate cognition, and I also defend it against frequently heard criticisms of dualism and internalism. Furthermore, in the final section, I offer some concrete examples that show how taking minds more seriously could help us better address conceptual problems in various disciplines. Here I focus, quite arbitrarily, on symbolic cognition in (past) primates, as this is what I’ve worked on most recently.

Generally speaking, the concept of mind as I defend it is simply that of the first-person viewpoint. I have outlined this concept of mind, based on an interpretation of Husserl’s (1859–1938) works, elsewhere in terms of what I called a “dual aspect theory” (van Mazijk, 2020b). The central (but admittedly hard to grasp) idea is that minds are not distinct things in the world, and consequently talk of minds does not refer to separate entities. Instead, it refers to a certain perspective on any given (real) thing. For example, a table can be considered as a physical, mind-independent being, but also as an appearance for me. The appearance is not a distinct entity; these are two ways of considering the same thing. Not just tables, but any given thing can be considered either as appearance for me, thus from the first-person viewpoint, or as mind-independent thing, as we are prone to do. This way, phenomenology considers “the [same] world, exactly as it was for me earlier [but now] as the correlate of [subjectivity]’’ (Husserl, 1970, 152). There is, as I purport to show later, nothing internalistic or dualistic about this concept of mind, as it refers to ordinary reality, only qua appearance, thus from a first-person viewpoint.

In my view, the central characteristic of the first-person viewpoint is intentionality. This is a technical term that is again borrowed from Husserl, and which primarily denotes world-directedness (see also Jacob, 2023). To analyze intentionality, and thus first-person viewpoints, primarily means to analyze the ways in which things appear(ed) – i.e. the world considered as appearance –, including things like other embodied minds, tools, cultural objects, symbolic meaning, oneself, and so on.

There are two ways in which this account of intentionality can be taken up in the study of (past) primate minds. On a “soft” reading, taking an intentional stance is a merely helpful heuristic device when it comes to theorizing about (past) primate minds and their evolution. This is not entirely unlike Dennett’s (in)famous take on the intentional stance (Dennett, 1988, 1989). For example, I might be interested in knowing whether extant non-human great apes are capable of understanding symbols, or certain aspects of the intentions of others. To this end, it might be smart to take the ape’s own point of view explicitly into consideration. For instance, in making a typical gesture, we can ask: what does the ape try to achieve? What do we need to assume regarding its understanding of others in order to make its behavior comprehensible? Does the ape itself experience the action as a symbolic one, or does it merely seem that way from an outsider’s perspective? And so on.

These are, I believe, very important questions to ask, and in the final sections of this paper I show that they are often not taken seriously enough. The main aim of this paper is to convince the reader of the practical value of thinking through the first-person viewpoint in this sense. At the same time, I argue for something slightly more substantial, as I don’t believe the first-person viewpoint is merely a heuristic device, whereas in reality it would somehow not exist. Instead, the first-person viewpoint is a real part of this world as much as anything else is. Consequently, intentional analysis (naturalized phenomenology) addresses something real, which is not straightforwardly reducible to other, third-person ways of addressing brains or cognition. This position, of which I only offer a fairly general account (in Sect. 4), can be seen as a form of non-reductive physicalism (Searle, 1992; Pereboom, 2005; Baker, 2007, 2009; Murphy, 2013). It maintains the distinctively first-person and intentional character of minds, but without abandoning naturalism, or worse, resorting to substance dualism (Fig. 1).

From the perspective of contemporary philosophy of mind, there may be little new about either of my suggestions (using the intentional stance and non-reductive physicalism). The point of this paper is not to contribute to philosophy of mind debates, but to offer an uncomplicated outline of a concept of mind – based on the concept of intentionality – that may be of use in debates on (past) primate cognition, and which may act as a counterweight to a certain mind-skepticism found in certain recent cognitive-archaeological literature. I therefore arranged the discussion in the following three sections around (what I take to be) three closely related misconceptions about minds. These are the following:

Fig. 1
figure 1

Three misconceptions about minds

Full size image

The point of this critical discussion (Sects. 2–4) is not to argue in favor of internalism or dualism, but to show that the anti-internalism and anti-dualism tendencies in recent cognitive-archaeological literature are overly critical of mind-talk. In short, the baby (mind-talk) is thrown out with the bathwater (internalism and dualism). Finally, in Sects. 5–6, I illustrate somewhat more concretely what a positive approach to minds in terms of intentionality may look like when theorizing about (past) primate cognition. Indeed, as I see it, many ongoing disputes cannot be resolved by means of empirical data only; they require conceptual work, also by thinking through first-person viewpoints.

2 Talk of minds entails internalism

Cognitive archaeologists seem ambivalent when it comes to minds. While they frequently write about them, the term usually functions as a synonym for cognition; it does not pick out anything specific. Moreover, on closer scrutiny, many cognitive archaeologists today write skeptically about mind-focused approaches. Ironically perhaps, this skepticism is partially fueled by representatives of approaches which claim indebtedness to phenomenology, a philosophical movement focused on consciousness and the way things appear from a first-person viewpoint. This includes material engagement theorists and enactivists (see discussions in Knapett & Malafouris, 2008; Malafouris, 2013; Gallagher & Ransom, 2016; Garofoli, 2018; Ihde & Malafouris, 2019).

Material engagement theory can be seen as part of the broader enactivism movement, a relatively new paradigm for thinking about cognition, which was developed in the early 1990s by Francisco Varela and others, with Daniel Hutto and Erik Myin as two of the better known contemporary representatives. Like enactivism and other theories belonging to the 4E-cognition family (extended, embedded, enacted, embodied – see also Newen et al., 2018), material engagement theorists are critical of traditional cognitivist approaches, which they view as Cartesian, mentalistic, or representationalist. Material engagement theory, in this regard, opts for a “co-extension of the mental and the physical” (Malafouris, 2013, 5). It suggests that the mind is not inside the skull; it rather includes objects, things, and materials in the outer world. Thus “broadening our understanding of mind” would mean reconsidering cognitive processes as “relationship[s] among brains, bodies, and things” (Malafouris, 2013, p. 50). In this way, external objects can co-constitute cognition. For example, fingers might be an essential part of early mathematical cognition, with later tally devices reshaping human cognitive processes, etc. (see further discussion in Overmann, 2016, 2023).

Many of the suggestions made by material engagement theorists are inspired by or resemble phenomenological claims, in particular those of Merleau-Ponty, but also Husserl’s (see Varela et al., 1993; Malafouris, 2013, pp. 4–8; Wilson & Foglia, 2015). For example, it is suggested that “the clay on the potter’s wheel should not be construed as the external passive object of the potter’s intentional states, but as a functionally co-substantial component of the intentional character of the potting experience” (Knapett & Malafouris, 2008, xiv). Accordingly, as Clark notes, “[u]nderstanding the human mind […] will require us to attend much more closely to the role of such bodily actions and external media than was once anticipated” Clark (2008, 10). Such ideas about bodily action and environmental interaction are in agreement with phenomenological work by Husserl, Heidegger, and Merleau-Ponty, in which embodiment and environmental embeddedness are central themes.

Clark further maintains a conceptual distinction between brains and minds. He notes that brains are “mind-producing engines”, and “humans, dogs, ferrets – these are, we would like to say, mindful things” (Clark, 2008, 1–2). Even the title of this particular contribution focuses on the question where brains, body, and world collide – but it does not draw minds into that picture, as one colliding thing among others. In Clark’s view, then, pottering humans, their brains, and their pottering wheels are co-productive of mindful activity, but minds themselves are not yet another thing dynamically connected to these things. Put differently, mindfulness (“having” a mind) is produced by these things, but it is not a thing among them.

Malafouris, the leading material engagement theorist in cognitive archaeology, does not usually maintain this important conceptual distinction between brains and minds. Instead, he speaks of “blending the mental and the physical” (Malafouris, 2013, 117), of a “mingling of mind and matter” (Malafouris, 2023, 154), and of a “dynamic coupling between mind and matter”, such that a potsherd is itself a “component of the prehistoric mind” (Malafouris, 2008, 22–24). As I see it, such phrasings tend to blur third-person and first-person concepts. Taken literally, these remarks could seem to suggest that minds are to be thought of as things among other things, and that these things may be mingled or dynamically coupled. Note that this is not how Clark spoke of minds in the discussion summarized earlier. Like Clark, Malafouris (2013) wishes to resist a dualism of mind and matter. By saying minds are on the order of brains and other material things, then, he presumably does not wish to suggest that minds are a special type of immaterial things among other things. Instead, a third-person, “material” viewpoint remains default throughout his assessment, such that minds – in the traditional sense of consciousness or a first-person viewpoint – arguably do not really enter into his picture at all. Put differently, the “mingling” really only concerns “brains, bodies, and things” (Malafouris, 2023, 158) - which are all material.

In my view, this suggests a certain mind-skepticism, which is further developed and justified in light of a string of anti-Cartesian arguments, which are taken to undermine the legitimacy of first-person approaches, that of intentionality included. It is suggested that any more substantial notion of mind must succumb to some of the long familiar criticisms of dualism and internalism. For example, it is noted that “intentionality, known also as ‘object-directedness’ … presents a significant obstacle for any discussion of material agency in the proper sense … Seen from this ‘internalist’ perspective, the issue of intentionality appears to be fairly straightforward – no room for active externalism here” (Malafouris, 2008, 28). In a more recent contribution, Malafouris (2023, 160) suggests that viewing minds in terms of intentionality or aboutness implies that “material stuff … is ontologically separate from the “mental” world”. Not just phenomenology, but the entire “theoretical basis of cognitive archaeology [is] trapped in a Cartesian universe” (Malafouris, 2013, 25).

Note that I admire material engagement theory as a paradigm for thinking about past hominin cognition. There is substantial overlap between it and my own intentionality-based work, as both emphasize the importance of action and practices. At the same time, this criticism of intentionality and first-person methodology seems plainly mistaken to me. There is nothing internalistic or dualistic per se about taking intentionality seriously. In fact, as I discuss in the next section, many – likely most – philosophers of mind today accept that consciousness and intentionality are real, without implying internalism or dualism for that matter. Also, in the Husserlian tradition, or at least within certain interpretative schools (see e.g. Drummond, 1990; Zahavi, 2008, 2010; van Mazijk, 2020a, b, c), consciousness is not regarded a thing “within” the world. Phenomenology rather studies the external world qua appearance. Zahavi (2008), in this regard, has noted that “it would be a mistake to dismiss offhandedly the analyses of the mind-world relation offered by a transcendental idealist like Husserl with the argument that they remain committed to an outdated form of internalism”, and that “the division between inner and outer […] might be inapplicable when it comes to phenomenological conceptions” (Zahavi, 2008, 355, 372).

In the introduction, I briefly summarized my own take on minds, which I defended elsewhere as an interpretation of Husserl (van Mazijk, 2020b). In short, on this reading, third-person and first-person statements often pertain to the same things, with the viewpoint from which they are considered being different. For example, an intentional description might set out to analyze how a tool appears and is used, instead of analyzing its component parts and materials objectively. To invoke intentional description, then, does not require conceptualizing an isolated subject which does not interact with the world. To the contrary, minds are intentional: the world is always already involved in them. This position does not, then, require an isolated subject, or an incomprehensible coupling of two different substances.

Whereas an intentional approach to minds may not sound particularly groundbreaking to any phenomenologist or philosopher of mind, some cognitive archaeology researchers appear to assume that we can simply get rid of all mind-talk once internalism (and dualism, as discussed in the next section) has been sufficiently discredited. To give one more (arbitrary) example: Ingold, in discussing the whereabouts of the mind, suggested that organisms and their environments form “one indivisible totality”, and that an “ecology of life” can address such totalities (Ingold, 2000, 19). This is perfectly agreeable. However, it does not warrant the conclusion that “we [thus] have no need to appeal to a distinct domain of mind” (Ingold, 2000, 19). Here too, a case against internalism (and in favor of externalism) is taken to refute mind-talk. But such a rethinking of the physical structures which are “mind-producing” – as Clark (2008) rightly put it – does nothing to undermine the appearance-reality of first-person viewpoints.

Externalism is increasingly popular among archaeologists today, and I have no particular beef with it. The externalist’s conviction that a study of past cognition or behavior need not be restricted to the study of skulls and brains is attractive to archaeologists for obvious reasons (they study prehistoric artifacts, after all). Moreover, a strong emphasis on externality is compatible with phenomenological methodology. As Zlatev et al. (2008) noted, Husserl “has only recently been properly understood in the Anglo-Saxon world to be concerned not with the nature of private experience”, but also with the body and environment (Zlatev et al., 2008, 3).

At the same time, externalism, at least in cognitive-archaeological debates, often drags along with it a certain material reductivism, and this is a cause for concern. For it does not help bring into view the first-person viewpoint as a legitimate and valuable scientific concern. Instead, anti-internalist sentiments effectively discourage researchers from taking mental concepts seriously. All the while, the anti-internalist arguments we hear do not yield a viable criticism of the first-person viewpoint or of intentionality, and so the baby has been thrown out with the bathwater.

3 Talk of minds entails dualism

Externalist perspectives in cognitive archaeology and related disciplines have a lot going for them. Yet anti-internalist and anti-dualist sentiments are not entirely guilt-free. Malafouris writes as if the choice is between a third-person (material) approach to minds (which arguably is not an approach to minds at all) and substance dualism: “a common thread that runs through the chapters of this book is that minds and things are continuous and inter-definable processes rather than isolated and independent entities” (Malafouris, 2013, 9). In other words, either minds and things are “continuous”, which seems to mean that minds are just like the material processes with which they are “coupled”, or they must be independent, supernatural substances as Descartes suggested. This is, in spite of good intentions, a false dilemma, as I purport to show in what follows.

My baseline in thinking about past minds is as naturalistic as that of the next contemporary philosopher. At the same time, I believe that talk of the first-person viewpoint does not straightforwardly reduce to third-person discourse. I thus resist anti-Cartesian sentiments not because I line up with Descartes, but because anti-Cartesian arguments do not yield a viable criticism of other (non-dualistic) concepts of mind, as I and many others today defend it. There is nothing new or special about my position, and I will in what follows offer only a general account of how I think minds could fit into nature in a way which does not succumb to a superficial criticism of dualism.

Here’s three examples. Put four hydrogen atoms together, and you will not suddenly get a brand new helium atom. Nuclear fusion happens only under specific circumstances of gravitational pressure and mass to energy conversion, namely in stars. After the weak nuclear force has done its job, the properties of the new helium atom are not reducible to properties of the hydrogen atoms started out with. We understand it and its features as representing something new. Likewise, we might say that the molecular structure H2O has properties irreducible to the hydrogen and oxygen atoms considered individually, just as we might say of a watch that it has properties (such as telling the time) which we do not find in its elements when taken apart.

We understand each of these three entities (helium, H2O, and watches) differently than we understand their individual components, although of course for very different reasons in each of these cases. Consequently, we also approach these things differently depending on what we are interested in. For instance, we might approach nuclear fusion in terms familiar to quantum physics, molecular compounds in terms of chemistry, and watches in mechanical terms. In some cases, we can move from one discourse to another without difficulty, and in other cases we cannot.

Things are ultimately not that different when it comes to talk of the first-person viewpoint. Philosophical laymen often assume that physicalism must reduce everything to one basic ontological level, and that any ontology of consciousness enforces a substance dualism incompatible with it. Clearly, this is not true. Among hosts of theoretical options, non-reductive physicalism denies dualism and embraces physicalism, yet it also denies the supposition that physicalism entails the rejection of minds. In other words, on this account, minds are admittedly strange things, but they are real, just like time-telling watches, molecular compounds, and helium atoms are real.

In my view, at some point in natural evolution, organic systems became organized in such a way that a new sort of virtual space gradually opened up, namely one that concerns the perspective internal to these systems. As Colombetti puts it, at some point “living systems realize a perspective or point of view from which the world acquires meaning for them” (Colombetti, 2014, 19). This probably happened with encephalization (the development of central nervous systems and brains), some half a billion years ago, but it may have earlier origins even in unicellular structure (see also Godfrey-Smith, 2021 for a similar account). This new, first-person appearance-reality consists of the things that the system represents or “views”, just as it represents or “views” them. Such organic systems are “built” in this world, but they now also “have” a world, and it is the very same world, but relative to its viewpoint, qua appearance to it.

Naturally, one might continue to address these new first-person structures in the familiar third-person terms, thereby putting them on the order of material things, as material engagement theory does. The first-person viewpoint is, after all, a part of nature. However, third-person terms are generally not suited to assess the matter at hand, as they do not speak immediately to the appearance-reality as considered from the system’s viewpoint. To assess the first-person viewpoint directly, a descriptive assessment of the intentionality which is productive of the (individual, social, and ultimately cultural) world qua first-person appearance is required, which is what intentional analysis (naturalized phenomenology) does.Footnote 2

This line of thought seems quite common sensical to me, and I cannot really see why anyone should have any big concerns about it, save when one departs from an ill-conceived dilemma between reductive materialism and substance dualism. As Searle (1992) once suggested, the fact of the first-person viewpoint cannot be coherently denied, as one cannot doubt its reality, given that its very appearance-to-oneself is its reality. I am thus certain that I have a first-person viewpoint through which the world is given to me. To deny other sentient organisms, past or present, such a viewpoint requires some spectacular philosophical sophistry.

I mentioned before that my take on minds can be taken up in a “soft” (the intentional stance) and a “hard” (original intentionality) way. Although I do not believe that what I have said in this section is particularly controversial today, it is not strictly necessary to accept this somewhat “harder” take on minds for current purposes, as the heuristic value of first-person thinking is arguably what matters most. Nevertheless, it is worth pushing back on anti-dualism sentiments in debates on past minds, insofar as they prevent scholars from taking up a first-person approach in debates where this is in fact much needed. Even Dennett, (in)famous for his rejection of original intentionality (and for characterizing its defenders as “the Bad Guys”), not only accepted but advocated “the role of the intentional stance in interpreting evolutionary history” (1988, 502). I will illustrate this need concretely with some examples in the two sections after the next one.

4 We can do without minds

Granted, then, minds are real. They exist today and they existed in the past. I accept that past minds, to speak with Husserl, had “their ‘world-representation’, […] with all the actualities which are valid for them in it” (Husserl, 1970, 272). I accept this obvious truth, and I understand such minds as real things of this world, just as I take molecular compounds and mechanical watches as parts of this world. I accept that, in virtue of this, minds or first-person viewpoints are a legitimate scientific concern. Be that as it may, I might still be wary of first-person approaches, and be inclined to think that we are better off avoiding fishy mind-talk altogether. So it is worth asking: could the prehistorical sciences also do without any reference to first-person viewpoints?

Let’s say that I find a potsherd or a stone tool. These are basic descriptive terms which any archaeologist must accept. Yet the description of these findings in terms of “pots” and “tools” arguably reveals that I have already interpreted them in relation to past subjects and their intentional states. For I have assumed that something like a “tool” or “pot” appeared to someone in the past, who engaged in the relevant practical and social activities, and thus comported itself to a world within which such cultural objects and related projects and plans were given. Merely in speaking of potsherds or stone tools, then, I have already tacitly reconstructed a part of past hominin intentionality.

This basic example purports to show that archaeologists inevitably rely on interpretations of the intentional activities of past subjects within worlds phenomenologically constituted for them. It seems to me that things are no different in related fields, such as those dealing with extant primate cognition. For example, Michael Tomasello, and many others with him, have assessed the social intentionality of pre-linguistic infants and non-human great apes. Apart from the pervasive talk of “reading other minds”, Tomasello frequently notes that they “understand […] intentional states of others”, that they “imagine the actual psychological content of what others are seeing”, that their “own experience is not part of [their] mental processing” (Tomasello, 2021, 49–64), and so on. Surely we could pretend that such descriptions pertain simply to brains or other “extended” material processes. However, it is hard to see what remains of them if we leave out first-person terminology, such as “mind”, “understand”, “intentional”, “imagine”, “experience”, “mental”, and so on. In other words, such descriptions and the understanding they convey involve an appeal to first-person viewpoints, and they are incoherent without it. Examples like these are numerous, and they effectively reveal that first-person viewpoints cannot be altogether avoided when thinking about (earlier) primate cognition, even if we wanted to.

5 Sign-object and sign-intentionality

So far, I defended a notion of mind characterized by intentionality, and I argued that certain recent anti-internalist and anti-dualist tendencies unjustifiably attempt to reject such a notion. The question remains: what can we do with first-person thinking in the prehistorical sciences? In my view, we need not assume that first-person thinking is going to radically change the way we think. However, it can, I believe, be a valuable aid in thinking about (past) primate cognition and behavior, and it can help straighten out some of the concepts we use. My own recent work uses intentionality and other phenomenological concepts in theorizing about prehistoric symbolic cognition (van Mazijk, 2022, 2024a, b). I shall not repeat these discussions in full, but instead present a fairly general discussion of icons and symbols here, to point out a number of places where I believe first-person thinking may prove useful.

Symbolic signs are traditionally defined in terms of arbitrariness and conventionality. This understanding derives largely from C. S. Peirce, and is still commonplace in discourse in ethology, primatology, and cognitive archaeology. As it is usually understood, icons involve a relation based on resemblance (e.g. a picture), and indexicality a causal relation or a covariation (e.g. smoke indicating fire, by being caused by it). On the other hand, a sign is arbitrary if there is no indexical (causal) or iconic (semblance) relation, and conventional if it requires social learning. Symbol use is universally regarded as one of the hallmarks of our species, but there is currently no consensus about how it evolved.

In my view, a significant shortcoming of Peirce’s semiotics, at least in the way it is adopted by most researchers today, is that it almost entirely ignores the first-person viewpoint. Instead, focus lies on material signs – what I call sign-objects –, rather than on cognition or experience – or what we may call sign-intentionality. Questions about sign-objects are questions about whether something is a sign, and if so, how we ought to classify it (as an icon, index, or symbol). Questions about sign-intentionality are questions about how signs are interpreted, communicated, used, and potentially constructed by embodied and socially engaging subjects. Questions of sign-intentionality deal with the lived, embodied, and intersubjective experiences of animals (primates, in our case), and with their first-person viewpoints.

It is a remarkable fact that many researchers working on symbolic cognition, also in cognitive archaeology, continue to use Peirce’s semiotics as their starting point, even though the classification of sign-objects is quite evidently of no interest to them at all. A cognitive archaeologist should be interested in cognition, and therefore in the mental processes that are implicated in the handling of signs, possibly within various communicative settings. But the basic distinction between sign-objects and sign-intentionality is overlooked here as elsewhere, and this leads to serious conceptual problems. I will in what follows discuss a number of examples from various disciplines, including primatology and cognitive archaeology, to illustrate how such problems appear in the literature, and how a focus on first-person viewpoints and sign-intentionality might help address them.

6 Icons, symbols, and the first-person viewpoint

To start off, in ethology, the relation which the shape of a thorn bug bears to an actual thorn is sometimes interpreted as iconic (e.g. Queiroz and El-Hani (2002). It is, after all, one of resemblance. As such displays are physiologically fixed, it should be clear that this particular example does not involve any communicative intent (see for further debate on communicative intent Moore, 2018a, b; Moore & Brown, 2022). Put differently, the sign is there whether the bug likes it or not. For that reason, it is arguably not a social act (it does not involve a directedness at another’s mental state), and it is in fact not even an intentional act (it does not require directedness from a first-person viewpoint at all).

While there is nothing problematic about the thorn bug’s classification as an icon, things get more complicated when compared to other examples of iconic sign-objects. For example, if I shape my hands to resemble the shape of an egg, then this signal is to be classified as iconic on the very same grounds as the thorn bug’s signal. The same holds for chimpanzee pantomime, which the majority of researchers regard as (at least partially) iconic. Again this is all very well, but by classifying these three forms of signs as icons, we have learned virtually nothing about the cognition they implicate. Instead, we have lost sight of their differences by classifying them together.

As said, the intentionality of signs is what ought to matter from a cognitive viewpoint. For example, it matters that the thorn bug’s signal is not a communicative act, as this means there need be no intentional directedness at others involved. Chimpanzee pantomime, on the other hand, does involve communicative intent, at least in certain cases. It also matters that the egg-signal resembles an object, whereas chimpanzee pantomime generally serves to request immediate actions by resembling those actions. Only humans naturally produce signs which resemble objects rather than actions, and this fact obviously has cognitive and behavioral implications. Whereas a sign-object approaches lumps various signs meaninglessly together, thinking through the first-person viewpoint may help highlight such important differences.

We can add a fourth icon example from the archaeological literature, namely the infamous Makapansgat pebble, which contains a natural pattern resembling a face, and which was carried along a long time ago by an australopithecine (Oakley, 1981). From the perspective of classifying sign-objects, the Makapansgat pebble too must be categorized as an iconic sign. However, it is iconic in virtue of cognitive mechanisms involved in facial recognition; it is essentially a case of seeing “meaningful patterns [a face] in meaningless data” (Bednarik, 2017). Now the ability to recognize faces is neurologically distinct (in the fusiform gyrus) from the ability to recognize a familiar action or an object like an egg (see also Sergent et al., 1992; Kanwisher et al., 1997). Yet the sign-object approach lumps all of these signs together as icons, and fails to capture such important differences.

The examples discussed so far concerned icons, but it is also worth reviewing some cases where researchers have been tempted to posit symbolic cognition. First, in debates on non-human great ape gestural behavior, gestures are often assumed to have a certain iconic origin, in this sense that a gesture resembles some natural behavior which is requested. However, sometimes the gesture develops in such a way that the initial iconicity is lost, as the first part of the act gradually comes to stand for the whole, a process called ontogenetic ritualization. As the iconic relation is now absent, the act is classified by some as symbolic. For example, Tomasello and Call speak of such gestures as representing “meaningful social acts symbolically” (Tomasello & Call, 2019, 462–463), and so do Cissewski and Luncz (2021).

Chances are this classification is incorrect, and we can motivate this by thinking through the first-person viewpoint more carefully. For it is quite plausible that the communicating ape performs the same intentional acts as before, while the interpreting party simply comes to understand the act at the initial stage, possibly through conditioning (see also Byrne et al., 2017), such that the execution of the first part alone henceforth suffices. Remarkably, this interpretation is backed by some of the same researchers who call such acts symbolic. If correct, these acts need not involve symbolic cognition, at the very least not from the standpoint of the communicator. Instead, the interaction merely seems symbolic from the outside, thus qua sign-object, with relatively minor cognitive changes on behalf of the interpreter explaining this change of appearance.Footnote 3

Another interesting example from the primatological literature concerns the so-called attention-getter (Tomasello, 2010, 27). The attention-getter might, according to Tomasello, be exclusive to primates, perhaps even to great apes. In short, attention-getters are likely individually learned behavioral displays that are used to attract the attention of others. An individual may notice that certain extravagant behavior draws attention to themselves, and subsequently learn to exploit this is in diverse settings. The attention-getter is not straightforwardly iconic, as the action does not resemble what it is about. Tomasello and Call (2019, 466) argue instead that the attention-getter “involves something in the direction of reference”, and that it manipulates “the attention of the recipient to specific entities”, namely by drawing attention to oneself. Because of this, they even suggest that it may form a “missing link” in hominin language evolution.

Taking first-person viewpoints better into account may help clarify whether talk of “reference” and “specific entities” is indeed fitting here. A referential act, of which pointing is arguably the most basic example, involves quite specific intentional structures on the communicator’s and interpreter’s side (see van Mazijk, 2024a where I outline this theory in detail). On the communicator’s side, pointing usually involves intentional directedness at the entity pointed out, which they take to be accessible to others as well. This means that pointing at oneself would involve awareness of oneself as an entity which can also be viewed by others. Second, pointing usually requires simultaneous directedness at the intentional state of the other, as the act serves to manipulate the other’s attention to the entity pointed out, in order to establish shared intentionality. This means that the communicator must be able to track the intentional state of the other, which amounts to what in Theory of Mind is called a second order act, as in “I see that you see X”.

The interpreter appears to be in an even more demanding position, however, for they generally need an understanding that the other wants to show them something. This means that they need an understanding of the communicator’s intentional state as directed at their own intentional states. In Theory of Mind, this would be called a third order act, as it involves a double recursion of intentional states (“I see X” is a first order act; “I see that you see X” is a second order act; “I see that you see that I see X” is a third order act).

Thinking through (what is likely) the intentionality of the attention-getter reveals, in my view, substantial differences with basic referential acts like communicative pointing. For one, it does not seem plausible to suppose that the communicator here considers themselves as an entity. It is worth noting in this regard that there is little solid evidence for pointing among non-human great apes in the wild at all, whether at themselves or at other things (see e.g. Leavens et al., 2005; Tomasello, 2021)Footnote 4, although they do sometimes move their own bodies so as to be properly seen before initiating pantomime. Consequently, it is arguably from the start unlikely that the attention-getter would involve reference to an entity, in this case to oneself.

Instead, it is in my view more likely that the communicator has a pre-reflective, embodied understanding of what it is like to be gazed at – an experience chimpanzees are thoroughly familiar with, as it is also needed for dyadic pantomime. By using an attention-getter, the communicator exploits certain attention-drawing behavior in order to achieve this familiar situation of having the other’s attention. Put differently, they know what it is like to be seen by others, and they have individually learned a new way to achieve this, namely by use of an attention-getter, but without necessarily knowing themselves as an object of shared attention for that matter. By comparison, a human newborn might use attention-getting behavior to draw a caretaker’s attention, but without thereby showing signs of being aware of themselves as an entity in the world, as would generally be the case in pointing to oneself (in humans, pointing develops around nine months of age).

This reading becomes more plausible once the interpreter’s viewpoint is taken into account, which does not resemble the pointing scenario either. For the attention-getter does not require a third order (“I see that you want to show me X”) act from the interpreter’s viewpoint, or even a second order one (“I see that you see X”). This is because the interpreter of an attention-getter need not understand the communicator’s intentions at all in order for their attention to be successfully manipulated. After all, they respond simply to a loud sound that was made, possibly without directedness at the other’s intentional states whatsoever. Indeed, attention-getters are probably useful precisely because they effectively circumvent the high cognitive demands of referential acts on the interpreter’s side, by simply enforcing a shift of attention through some extravagant attention-getting behavior.

The concept of the attention-getter is an interesting one, and it bears on ongoing discussions on the origins of referential acts and symbolic cognition. Tomasello and Call (2019, 466), in this regard, suggested that it may constitute a “missing link” in hominin language evolution, because it would involve “something in the direction of reference”. Taking first-person viewpoints into consideration offers one way of assessing the plausibility of such claims, if only for initial purposes. In this case, it reveals that the attribution of complex social understanding to attention-getting behavior is likely unwarranted, which leaves the suggestion concerning a “missing link” unmotivated.

Icons, symbols, and other communicative signals also draw the interest of cognitive archaeologists. Unlike primatologists, archaeologists frequently speak of symbols when addressing the distinctively human production of cultural objects. In particular, the term is often reserved for the sort of behavior long thought to be exclusive to behaviorally modern humans in the Upper Paleolithic/Late Stone Age, as evidenced by figurines, such as the famous lion-man of Hohlenstein-Stadel, dated some 30kya (Dalton, 2003), or the rock art of Chauvet and Lascaux, dated some 28kya-37kya and 17kya respectively (Quiles et al., 2016; Ducasse & Langlais, 2019).

More recent findings have problematized this picture. Use of pigment may go back 400-500kya (Wynn, 2012, 290–291; Rifkin et al., 2015; Dapschauskas et al., 2022), and its decorative use is often taken to indicate modern symbolism as well (Henshilwood and Dubreuil 2009, 50; Vanhaeren & d’Errico, 2006, 1107), although pigment differs in many important ways from hand-made objects like figurines. The function of beadwork and pendants, which are more recent inventions but nonetheless originating well before the Upper Paleolithic and likely not unique to Homo sapiens, are also often said to be symbolic (Zilhão, 2007; Zilhão et al., 2010, 1023; Henshilwood & Dubreuil, 2011, 375; d’Errico and Vanhaeren, 2012; Prévost et al., 2021, 1). Others have been more critical of such assessments, however, and the debate rages on to this day. For example, Wynn and Coolidge have suggested that a certain “intentionality typical of modern human social interaction” is presupposed, but not that “beads stand for anything at all” (Wynn & Coolidge, 2007, 88).

So where does this leave us? It is worth noting that not a single one of the major archaeological findings alluded to – figurines, rock art, ochre use, and beadwork – qualifies as symbolic in the Peircean sense.Footnote 5 Most of them are, qua sign-object, iconic, although it is disturbingly difficult to classify both beadwork and pigment convincingly within Peirce’s tripartite division, which again highlights its shortcomings. It is worth emphasizing again that the classification of sign-objects should not be of particular interest to cognitive archaeologists in the first place, as it is the implicated subjectivity or cognition that matters. Needless to say, symbols might still have been involved in the handling of some or all of these signs. However, the sign-object rarely proves a useful starting point for thinking about sign-intentionality, for they do not seem to correlate strongly. For example, figurative art is an extremely complex, human-unique achievement, yet it is qua sign-object iconic, which is not a human-unique form of signaling. On the other hand, ontogenetically ritualized chimpanzee signals are sometimes classified as symbolic qua sign-object, a description which may not match their sign-intentionality, as I argued earlier.

Consequently, a focus on sign-objects may not be very useful, and it does little to help us understand how signs were actually handled. Instead, the first-person viewpoints of embodied and socially engaging hominins who used these signs must be taken into consideration, also in cognitive archaeology. In the context of ongoing debates on Middle Stone Age symbolic cognition, I have in other recent work repurposed the phenomenological concept of “horizon”, a concept originally developed by Husserl, in order to rethink how symbolic signs may have been used in basic communicative settings. Husserl used the term “horizon” to refer to any kind of background awareness of what is not immediately perceived yet somehow made co-present. Husserl famously distinguished between the “inner” horizon of things (their currently unperceived sides) and their “outer” horizon (their unperceived surroundings), but he used the term in diverging ways, as for instance in speaking of a “horizon of familiarities” and “the world as horizon” (Husserl, 1997, 197; 2001, 40–42; 1970, 31). For example, in perceiving a chair, I am tacitly aware of its unperceived sides (inner horizon), its place in my living room (outer horizon), and it belonging within a larger cultural world in which I stand with others (world horizon).

Heidegger, a former student of Husserl, used the term horizon more specifically in relation to a background understanding of how things are used (are “ready-to-hand”) within an ongoing practice or “equipment-context” which is pre-reflectively understood (Heidegger 2001, 109, 340, 464). Building on these ideas, I have introduced the concept of a “shared practice horizon” in discussions on early symbolic communication (van Mazijk, 2022, 2024a). The basic idea is that early symbolic communication would have relied on a shared background understanding of an ongoing practice. For example, pointing to a tool, or using a general concept like “use object”, would at first have relied on a shared background understanding of a social practice, say Acheulian tool making. This social practice is one which both communicating parties are intimately familiar with; there is a shared understanding of activities, expressions, arm movements, etc., and the intentions and goals they serve in this context. This practice need not be thematically known, but functions as a background “horizon” for communicative activity.

A shared practice horizon would thus serve as a socially enacted background framework which enables the use and interpretation of signaling acts, especially those in which the message is not codified indexically or iconically in the overt behavior. For instance, the communicative success of pointing to an anvil in an Acheulian tool making session would depend on a shared understanding of the ongoing practice of making stone tools. It is only because both parties stand in the same practice horizon (making tools), that the interpreter can grasp the social intention of this pointing gesture (“I want you to use that anvil”). This requires both parties to assume that they are, at least to some extent, cooperatively engaged in the same practice. If, by contrast, the recipient lacks the appropriate practice horizon, this may lead to communicative failure. For example, X’s pointing to an anvil when Y is preparing for sleep would not allow Y to disclose whatever the meaning of that gesture was, as they are not sharing the same practice horizon.

It makes sense, in my view, to posit a reliance on shared practice horizons in early non-iconic, non-indexical signaling, pointing included, particularly because this helps delimit the scope of possible social intentions the communicator has without requiring more complex means to make this explicit, such as more extensive sign systems (van Mazijk, 2024a). Put differently, instead of specifying things in great detail, a single sign like pointing can be effectively used to mean different things, with the shared practice horizon functioning to provide the shared context needed to get the message across.

In summary, first-person thinking can help us approach important questions about signs and other tools in new ways, in particular by emphasizing how they were used and handled, as well as by offering detailed analyses of the embodied and social understanding they may implicate. Considering first-person viewpoints can also help us spot and correct mistakes in our interpretations of (past) primate cognition, or at least further debate about them, as some of the examples in this section illustrated.

7 Conclusion

I used this paper to outline a particular concept of mind which does not succumb to superficial criticisms of dualism and internalism as often heard in today’s cognitive-archaeological debates. This concept of mind is that of the first-person viewpoint, characterized by intentionality. In the final part, I briefly pointed to ongoing discussions on icons, referential acts, and symbolic cognition, which served merely as examples of the many places where first-person thinking could prove practically useful. Here I also introduced a distinction between sign-object and sign-intentionality, and suggested that cognitive archaeologists should focus more on the latter.