Some introductory remarks

I greatly appreciate the effort of Lindstrøm and Sørensen in expanding the discussion on object agency (Lindstrøm Reference Lindstrøm2017; Sørensen Reference Sørensen2018). I believe this discussion is of great importance as it has allowed us to address details and inconsistencies that often go unnoticed or unreported when addressing agency in archaeology, and I truly believe we have made headway with this discussion. Since it has been established that Lindstrøm and I are pretty much on the same page (Lindstrøm Reference Lindstrøm2017; Ribeiro Reference Ribeiro2016a), I will be focusing mostly on the latest response put forward by Sørensen (Reference Sørensen2018). I hope Sørensen views my persistence not as direct criticism of his ideas, but rather as a demand for continued clarity on a topic fraught with confusion, both inside and outside archaeology. As Sørensen claims, it is unlikely that we will reach an agreement, but I believe it is not a bad idea to outline our disagreements in a clear fashion and let the readers decide for themselves which side sounds more reasonable. This response will be divided into three parts: the changing definition of agency, the many contradictions of agency in the New Materialist movement, and agency studies within the discipline of archaeology.

Changing definitions of agency

Despite Sørensen’s claim that we are in disagreement (Reference Sørensen2018, 95), we seem to have agreed from the beginning that agency is not something ‘in itself’; that is to say, having agency is not the same as having opposable thumbs or a key chain. Additionally, agency is, as Sørensen and I seem to have established, something that changes from context to context. This reasoning, while coherent, can nevertheless lead to some confusion. Just because something changes from context to context, does not mean that it has to hold a different definition in every context. ‘Religion’ is something that can manifest in a plurality of ways, yet can also have a somewhat consistent definition. This is not to say that definitions cannot change over time, because this will happen whether we like it or not, but it is one thing to adapt concepts to changing realities, and something else entirely to intentionally restructure a concept to suit ideological needs and, in the process, make concepts barely understandable. This is what has happened to agency.

A question can be asked: but what if this restructuring of agency is a good thing? My contention is that it is not, for reasons I have elaborated upon previously (Reference Ribeiro2016a) and will also provide in this paper. By completely restructuring agency, Latour (e.g. Reference Latour1993; Reference Latour1999; Reference Latour2013), Gell (Reference Gell1998) and associates are trying to radically alter the way we practice the social sciences, from science studies, to anthropology, to archaeology. At no point have the supporters of Latour or Gell argued that the research conducted under previous understandings of agency is empirically wrong – the arguments come solely from a new metaphysical understanding of agency. My question is: why change how we researched agency? Was it wrong? Was it boring? And if the answers to these questions are ‘no’, then why change something that is not broken in the first place? To pursue this argument further, imagine if Latour’s or Gell’s understanding of agency were to change so drastically, to a point where all their research would have to be considered irrelevant, would their supporters not express objections to the change? The changes to agency have been so radical in recent years that what was considered high-quality and fruitful research on agency just some decades ago, is now considered outdated or invalid. This, in turn, leads down a path where, instead of expanding the methods, theories and research interests of the social sciences and the humanities, we are forced to narrow them down. As Simon Choat has pointed out, accepting Latour’s ideas on how to conduct research on agency implies accepting all his ideas in full (including the inconsistencies in ideas between books) – topics of research such as the market, the state or capitalism are precluded from study – they are maxi-transcendences (Latour Reference Latour2013, 394 ff.) which cannot be understood properly in the framework of actor-network theory (Choat Reference Choat and Latour2016, 139).

Thus, a contradiction arises out of some of Sørensen’s claims: agency is complex and nuanced, manifesting in a plurality of ways across different types of agents and contexts, yet not complex or nuanced enough to recognize a distinct type of human agency; that is, one that recognizes actions as a dialectic established between subjects and objects. Any study of subjectivity is precluded – the subject is nothing more than an affectation of modernity (Latour Reference Latour1993), with some New Materialists in philosophy going so far as to claim that engaging with subjectivity, death or language is anti-realist (Bryant, Srnicek and Harman Reference Bryant, Srnicek and Harman2011, 4)! Changing definitions to suit whatever trendy ideas are in vogue is not only misleading, it is downright intellectually dishonest. Anyone can come up with a new definition of reality, and then argue that everyone who does not follow this ‘new’ and trendy notion of reality is an anti-realist.

Therefore I cannot agree with Sørensen’s comments on how definitions change because he is wilfully disregarding the fact that some definitions change because reality changes, whereas some definitions are changed in order to shape reality according to new trends, in the current case the New Materialisms.

The contradictions of agency in the New Materialisms

We now move to the topic of agency and its context. On this issue, agency has been described either as something that is specific to certain contexts, or, if we prefer Gell’s understanding, as something more akin to a dynamic ‘force’ immanent to all matter (Ribeiro Reference Ribeiro2016a, 231–32). However, as Sørensen correctly argues, Gell also provides a much more nuanced description of agency, one which recognizes a necessary distinction between primary and secondary agents (Sørensen Reference Sørensen2018, 96). This can come off as somewhat contradictory – how is it possible that agency manifests in two different ways according to Gell – first as ‘a global characteristic of the world of people and things in which we live’ (Gell Reference Gell1998, 20), and second as something that manifests differently depending on the context and the agent itself? Is this an actual contradiction in Gell’s ideas or a mischaracterization of his work? I do not believe it is either and I will explain why.

It is entirely possible and logically sound to argue that agency is a universal element of the world and at the same time describe it as something ‘relational’, ‘labile’, ‘heterogenic’, ‘scalable’ (to use Sørensen’s own descriptors). I believe that Gell has not contradicted himself – it seems to my reading that what Gell is saying is that agency is a universal element that can manifest in a heterogeneous and nuanced way. I do not see why this understanding of agency would be a problem for a supporter of the New Materialisms, as it is what allows the natural alliance between Latour’s and Gell’s ideas with ‘vibrant matter’ (Bennett Reference Bennett2010). Addressing agency in these terms is, however, nothing more than metaphysical speculation – it bears no consideration whatsoever of the state of affairs of the world, and most importantly, it has no bearing on the relation between agency and responsibility.

When it comes to responsibility, New Materialists usually bring up Latour’s most famous example of human–object interaction: that of a gunman (e.g. Harris and Cipolla Reference Harris and Cipolla2017, 132), which can be found in his book Pandora’s hope (Reference Latour1999, 176 ff.). As has been consistently repeated, the issue of gun violence is not so much an issue concerning either guns or people but rather the hybrid actor composed of a gun and a person. Given this new understanding, Latour puts forth a new understanding of responsibility, one that must be recognized as shared between the gun and the gunman (ibid., 180). This sounds awfully brilliant on paper, but a closer look reveals some discrepancies: when it comes to responsibility, should we treat a school shooter in the US the same way as a child soldier in Africa? From the perspective of object agency, both a school shooter and a child soldier are metaphysically identical – they are both gun + human. Therefore should we treat both these cases the same when it comes to responsibility?

With this comparison, Sørensen could argue that I am oversimplifying and caricaturing agency according to the New Materialisms and symmetrical archaeology, as he accused me of in his response (Reference Sørensen2018), but if that is indeed the case then simple and caricatured examples, such as Latour’s gunman or Gell’s car breaking down (Gell Reference Gell1998, 18–19, 22), need to be ignored given that they obviously misrepresent reality. From my perspective, a network such as that composed of human + gun can only exist because there is already an ethical framework in place, a framework that creates an asymmetry not only between human and objects, but also between humans. Therefore, I find it strange that we should recognize symmetrical responsibility in the case of the gunman (Latour Reference Latour1999, 80) when it is precisely asymmetrical responsibility that conditions and allows humans + guns to come together in the first place.

Thus, against Sørensen, I believe it is the New Materialisms and symmetrical archaeology that have oversimplified reality. It makes it seem complex, by medium of neologisms, but at its essence the New Materialisms and symmetrical archaeology want to reduce the world to nothing more than relations between objects. It is precisely here where the caricature and oversimplification lie: New Materialisms and symmetrical archaeology want to lay claim to a universal understanding of agency (relational and involving humans and objects), yet at the same time hide this universality behind adjectives such as those used by Sørensen – ‘relational’, ‘labile’, ‘heterogenic’, ‘scalable’. Thus, despite the fancy words used to describe agency, it remains too one-dimensional in the New Materialist movement and symmetrical archaeology. And this, of course, inevitably leads to rather boring and uninspiring research, as Latour has come to realize: if everything is a network, then everything starts looking the same (Latour Reference Latour2013, 35).

A much more productive way of thinking about agency is to recognize it as eminently historical, in the sense that not only are there cases where agency does not exist, but also cases where agency simply cannot be thought of coherently. What this means is that the metaphysical principles supported by the New Materialisms cannot be universal. Additionally, this forces us to turn the New Materialist strategy on its head: whereas the New Materialists start by conceiving agency in universal metaphysical terms, and from there argue that responsibility should be recognized as such and such, the alternative I propose here is to look first at the history of the societies under analysis. It just seems more coherent to me to research responsibility empirically, through sociological methods, when this is possible. This would allow us then to understand agency based on how different human groups assign action (whether it is to collectives, individuals, families, animals, deities, nature, the past, the cosmos, history, etc.). The New Materialists, however, prefer starting from the ontological presupposition that all agency is relational – even in cases where it is not, like in the case of modern Western society, where responsibility is seen primarily as owned by individuals, where humans have acquired the means to self-legislate, i.e. create rules and responsibilities for themselves (Pippin Reference Pippin2008).

This brings us to the modern Western belief that human individuals are agents (Habermas Reference Habermas and Habermas1979, 110; Siedentop Reference Siedentop2014). This belief does not stem from modernist philosophy, as many assume; on the contrary, this belief, and modernist philosophy in general, are reflections of actual historical, social and economic conditions, as Hegel shrewdly argued (Reference Hegel2001). Accordingly, to research the belief that human individuals are agents, we should be looking at actual historical events such as the Enlightenment, the Industrial Revolution, and the French Revolution, research that Jonathan Israel has already started (Reference Israel2001; Reference Israel2006; Reference Israel2011; Reference Israel2014). Thus humans have agency not as some sort of inherent property, but because humans created the historical conditions for agency to be recognized. Modernity was crucial in this respect since it legitimized agency – it made humans into political agents during the French Revolution through the principle that all humans are created equal, and it also made humans into economic agents by giving them power as consumers.

This is perhaps what confuses me most about the New Materialists: despite the persistent reference to modernity they have yet to provide an actual materialist study of modernity itself (Hornborg Reference Hornborg2017, 14). Many New Materialists blindly trust Latour’s comments on modernity yet at no point has he ever put forward an actor-network study of modernity. There are two books in which he engages with modernity and these are We have never been modern (Reference Latour1993) and On the modern cult of the factish gods (Reference Latour2010). In both these books, his method of analysis is Hegelian (which is incredibly popular among French philosophers). Following Hegelian dialectics, what Latour does is to identify conceptual dualisms and resolve them by resorting to a higher third term, such as ‘purification’, for instance. In fact, it is precisely from Hegel that we obtain the title We have never been modern: it follows the principle of negation in dialectics, meaning that modernity can never truly exist because it negates itself. Thus with Latour we find ourselves with a classic case of ‘do what I say, not what I do’: Latour can employ Hegelian dialectics and create new non-material concepts out of thin air, but the research he asks others to commit to is materialist network theory and Hegel has to be ignored tout court because he is an idealist.

Agency and archaeology

The problem of object agency in archaeology is one that goes beyond agency itself. It is a problem highlighted in another paper (Ribeiro Reference Ribeiro2016b), which is that of blindly following trends. In spite of repeated claims that archaeology is, in fact, making unique contributions to the philosophical discourse of the New Materialisms (Edgeworth Reference Edgeworth2016; Nativ Reference Nativ2018; Olsen et al. Reference Olsen, Shanks, Webmoor and Witmore2012), it seems that these unique contributions exclude any critical scrutiny of New Materialism itself. Because of this situation, we now find archaeological theorists citing Heidegger, Latour and Meillassoux approvingly (e.g. Olsen et al. Reference Olsen, Shanks, Webmoor and Witmore2012) even though the philosophies of these three thinkers are mutually incompatible.Footnote ¹ For instance, Meillassoux’s notion of ‘correlation’ was adopted specifically to critique Continental philosophical work since Kant, and this includes the work of scholars such as Latour and Heidegger. According to Meillassoux (Reference Meillassoux2008), ‘agency’ would fall squarely into correlationist philosophy. An accurate representation of the world is one that recognizes the absolute contingency of reality, and this recognition can only be attained by resorting to set-theoretic logic, and not through vague and esoteric notions such as ‘vibrancy’, ‘withdrawn objects’ or ‘embodiment’ (ibid., 82 ff.). I am disconcerted that this section of Meillassoux’s book (the best section in my opinion) has gone unread.

It is now easy to understand my reluctance to accept object agency: it is built on a house of cards, i.e. premises and ideas which are theoretically questionable. Instead of logical clarity, much of the writing surrounding object agency is hidden behind nonsensical allusions and poetic metaphors, or, as Alf Hornborg (Reference Hornborg2017, 2) has described it, it is something more akin to ‘dinner conversations after some glasses of wine’. It seems obvious that archaeology requires much more theoretical engagement than it currently manifests, but I also believe that this engagement needs to be less cryptic and considerably more sober.

Furthermore, my critique of object agency does not imply a rejection of network theory. Network theory has proven itself one of the best methods of analysis in archaeology, as evinced in the excellent work of Chris Fowler (Reference Fowler2013) and Peter Whitridge (Reference Whitridge and Crothers2004). In fact, I fail to see why this type of research cannot be conducted side by side with research based on more conventional methods of analysis, such as those involving normativism, power and subjectivity. As far as I can see, the only reason why more conventional sociological methods should be avoided is because Latour says so.

Finally, there is the fact of what agency represents in archaeology and why its more traditional understanding remains important. When agency first became popular in archaeology, it was used (quite effectively) as an argument against the somewhat determinist models and explanations of human behaviour provided by archaeologists of the processual tradition (Stanton Reference Stanton2004). At the time, agency had some association with intentionality and the idea that humans could act purposefully – and not just as a result of, for example, ‘extrasomatic means of adaptation’ (Binford Reference Binford1962, 218). This idea of purpose is more important now than it ever was. In light of a third scientific revolution (Kristiansen Reference Kristiansen2014), and a desire for a return of processual archaeology (e.g. Kintigh et al. Reference Kintigh, Altschul, Beaudry, Drennan, Kinzig, Kohler, Limp, Maschner, Michener and Pauketat2014), the traditional conception of agency, which highlighted the role played by human choice, is in dire need of a return.