Museums and cultural memory in an age of networks

Assemblages

Using aspects of DeLanda’s (2006, 2010) assemblage theory as a materialist framework for cultural analysis, ontological assumptions subtending discussions on persons, networks, media, cultures, institutions, states and various other social wholes, are highlighted and challenged, especially where these are conceived in general terms. In DeLanda’s world, assemblages are the only legitimate entities that may be said to exist, not persons, or societies, or other reified generalities such as ‘the media’ or ‘the market’. Since DeLanda (2006) uses the terms ‘network’ and ‘assemblage’ more or less interchangeably, we will do the same. After all, networks or assemblages are both sets of social wholes that are simultaneously individuals. As DeLanda (2006: 40) puts it:

The ontological status of any assemblage, inorganic, organic or social, is that of a unique, singular, historically contingent, individual. Although the term ‘individual’ has come to refer to individual persons, in its ontological sense it cannot be limited to that scale of reality.

While human persons may be considered assemblages in DeLanda’s world – they are composed of heterogeneous components, some of which possess a degree of autonomy, for example, a heart may be transplanted from one body to another and continue functioning – the point is that we find individuals nested in sets at multiple levels of scale. Assemblages exist in a multi-scalar reality, constituting individual arrows of time. At the interpersonal scale, assemblages are networks of individual persons, material and expressive components, and the interactions between them (nodes and links). As emergent social wholes, these possess properties such as identity and density, ‘a measure of the intensity of connectivity among indirect links’, as well as memory – ‘the capacity to store local reputations’ – and a degree of stability (DeLanda, 2006: 56).

For DeLanda, a crucial question confronting any attempt to think about history is the nature of the historical actors that are considered legitimate in a given analysis. Assemblage theory asks us to define the historical agents that we would be prepared to defend the reality of in our theories. For example, is it reasonable to speak of ‘the rational individual’ or ‘cultures in general’ as real entities? At the same time, it critiques three common reductionisms, micro, macro and meso. ⁴ In the first instance, also known as ‘methodological individualism’, the individual person constitutes the core ontological commitment. Larger interpersonal networks may well be acknowledged, such as ‘society’ or ‘culture’, but always as entities that are nothing more than the sums of their parts, lacking emergent properties such as agency, memory and collective thought. In the second case, macro-reductionism, it is society at large that is the fundamental ontological commitment, with individuals given the role of ‘products’, constructed and constrained by prevailing large-scale structures. This stance is also known as ‘methodological holism’ or ‘collectivism’. It may be useful to refer to Branko Mitrović’s (2010) discussion of individualism and holism-collectivism for clarity on these terms.

Individualist historiography assumes that the name of a collective refers exclusively to a set of individuals, that any activity of a collective is ultimately describable by listing the activities of its individual members and their interaction.… According to the opposing, collectivist viewpoint, collectives are the primary entities of any historical account while individuals, their thoughts and actions, are derivative. (2010: 1–2)

At stake between individualists and collectivists is the question of which scale to situate causal explanation at, the micro or the macro? Individualism reduces agency or causality to the level of individual persons, whereas collectivism-holism reduces the same to the level of groups or social structures. The former emphasizes upward causation while the latter emphasizes downward causation. In the last case, meso-reductionism takes something from the middle of the micro and macro levels. DeLanda refers to Anthony Giddens’ ‘praxis’ as an example, where praxis is the true stuff of reality, making the individual and social unit epiphenomenal. Meso-reductionism entails a commitment to ‘middle-out’ causal explanation, in other words, praxis or some other intermediary event, such as relations or interactions, produces everything observable at individual and social levels. In contrast to these tendencies, assemblage theory conceives of causality in complex and non-linear terms; it does not restrict causal explanation to a single level of scale, whether micro, macro or meso. It takes the assemblage, simultaneously individual and social, as the basic ontological commitment, the only entity it is prepared to defend as real, and gives it the characteristic of self-causation (causa sui).

DeLanda’s more recent work, Deleuze: History and Science (2010) develops assemblage theory further, differentiating it clearly in many respects from the sense in which it was used by the originator/s of the concept, Félix Guattari and, subsequently, Gilles Deleuze. Perhaps the chief innovation lies in the parametrization of the concept of assemblages, where the operations of territorialization and coding are given in ‘degrees’, ultimately determining the degree of hierarchization or stratification of each assemblage. ⁵ Some explanation may be helpful at this point. Parametrization simply means assigning a parameter or ‘control knob’ to a concept. Territorialization and deterritorialization consist of processes by means of which pre-individual components and their interactions become individuated and de-individuated as a social whole, while coding and decoding are processes whereby, for example, linguistic codes consolidate and reinforce (and undo) an assemblage’s identity and consistency. Together de/territorialization and de/coding are two core parameters exerting a recurring influence over the configuration, identities and possibilities available to an assemblage. These parameters function as control knobs in the sense that if the degrees are turned up, we should expect a higher degree of hierarchization or ‘institutionalization’, and corresponding homogenization. On the other hand, if the degrees are turned down low, there is a greater possibility for the coexistence of heterogeneous parts, but also a less clearly defined overall identity as an assemblage.

Given popular misconceptions of networks as non-hierarchical, it is important to emphasize that networks may take hierarchical and non-hierarchical forms, as well as combinations of the two. According to DeLanda (1997), organizational forms exist along a continuum ranging from hierarchies to meshworks. Deleuze and Guattari (1987) call these poles ‘trees’ and ‘rhizomes’, respectively. Organizations at the hierarchical end of the continuum are characterized by a concentration of power in chains of command and the unidirectional flow of orders from central to subordinate nodes. State armies and corporations, which imported military strategies through operations research and management science, are without doubt the best exemplars of hierarchies. Organizations at the meshworks end of the continuum are by contrast, characterized by a diffusion of power across component nodes and many-to-many information flows. Drawing on the economic history of Braudel, DeLanda uses the town market as an example of a meshwork. Meshworks are self-organizing, meaning that organizational coherence and functionality arise as emergent properties of a swarm of reciprocal interactions and many-to-many information flows between nodes.

The two organizational logics are literally poles apart, constituting entirely different solutions to the social problems of resource interdependence and existential uncertainty. Hierarchies characteristically attempt to eradicate uncertainty through rational planning and centralized decision-making by expert managers and consultants, and, wherever possible, to eliminate resource interdependencies through horizontal and vertical integration. Such a collectivizing tendency is actualized in rigid organizational processes in which subordinate nodes possess very little autonomy and decision-making power. On the other hand, meshworks characteristically cope with resource interdependencies by making the best of them, benefiting from synergies and cooperation as well as changing local circumstances wherever possible. In contrast to hierarchies, meshworks tend to be actualized in flexible and flat structures, in which individual nodes are relatively autonomous and enjoy greater decision-making power.

All sorts of networks exist along the continuum between the two extreme tendencies, hybrid networks combining hierarchies and meshworks in different configurations. In fact, strictly speaking, DeLanda’s hierarchies and meshworks, and Deleuze and Guattari’s trees and rhizomes, may be considered as pure differences in nature or kind, real tendencies that are actualized only as composites. Hybrids are thus the rule, with all assemblages existing as composites of hierarchies and meshworks. The usefulness and relevance of DeLanda’s continuum is that it allows us to identify the degree of hierarchization in any given network, to discern its degree of concentration–diffusion of power, without falling into the error of conflating all networks with meshworks, thereby missing important power differentials. This is an error made in an article in Frieze, to cite but one example, where Cornell and Varnelis (2011) observe that ‘broadly speaking, the art world is vertical’ that is, characterized by ‘escalating levels of privilege and exclusivity’, whereas ‘the web is horizontal (based on free access, open sharing, unchecked distribution, an economy of attention)’. Searches for well-established artists in regions such as South Asia in highly popular English-language social networks, such as Tumblr and Twitter will rapidly reveal the hyper-visibility of Euro-American artists and art histories on one hand, and the invisibility or near-invisibility of the artistic histories and practices of territories including India, Pakistan and Bangladesh on the other.

If we assume a longue durée perspective ⁶ on the relations between hierarchies and meshworks, what stands out in the internetworked globalization of world economies is the fragile opportunities for small-scale actors towards the meshworks (peer-to-peer) end of the continuum to gain a surer footing in political-economic claims-making vis-a-vis large-scale and highly stratified institutions such as states and corporations, at least in some areas. Think for example of the decentralized ‘Anonymous’ movement. Braudel’s (1977) economic history shows how hierarchical institutions globalized themselves at an early age, with capitalist actors more or less holding land-locked markets to ransom. Small-scale actors simply did not (in most cases) possess the access to credit and command of capital to compete with the claims and contention of these capitalists. The internetwork changes the situation somewhat by unmooring populations of small-scale actors from geopolitical boundaries, accounting for a contemporary explosion of translocal networked publics and markets, accommodating the expressive activities of meshworks alongside those of hierarchies, making it increasingly difficult for institutional actors to eliminate or repress the claims and contention of small-scale actors, including traders, activist and dissident groups and so-called media ‘pirates’. ⁷ State assemblages are predictably responding in many parts of the world with repressive measures in the digital sphere, with censorship, surveillance and the persecution of dissidents.

Cultural memory

If memory tends to be conceived in individual or personal terms, this may reflect the resilience of strains of methodological individualism and the preponderance of psychology and the neurosciences in the development of theories of memory and its functioning, and in its shaping of common-sense understandings of time, presence and the past. This tendency has been criticized, in the literature on social and cultural memory, as reductive of the social complexities involved in memory practices, notably in the pioneering work of sociologist Maurice Halbwachs (1952), a contemporary of the art historian Aby Warburg (coincidentally, both independently developed theories of large-scale memory). Rowlinson et al. (2010: 72) observe that Halbwachs problematized the very notion of individual memory, and is largely ‘credited with introducing the concept of collective memory into contemporary usage’. For sociologically oriented thinkers, psychological accounts of memory too often suffer from a lack of attention to the social contexts, relations and dynamics within which recollection takes place. Sociological accounts of memory, on the other hand, remind us how ‘we actually remember much of what we do only as members of particular communities’ (Zerubavel, 2003: 3). Even within social and cultural memory studies, however, there are individualist and collectivist tendencies, or what Olick (1999: 336) describes as ‘two radically different concepts of culture … one that sees culture as a subjective category of meanings contained in people’s minds versus one that sees culture as patterns of publicly available symbols objectified in society’.

This is not a problem for assemblage theory, because, as intimated earlier, the assemblage is at once subjective and objective, individual and social. Its objectivity consists of its actual presence, its material and expressive components, and its subjectivity, of its memories and recollections. Its individuality is evident from its identity, and its sociality, in its relations with populations of other social wholes. Cultural memory in assemblage theory designates the flux of space and time proper to the objective and subjective lines of historically situated assemblages; it involves perceptions, actions and recollections distributed across complex circuits traversing a museum’s past and present. In contrast to common-sense understandings of time as a successive layering of pasts and presents, and approaches that conceive of memory as a storage bin of static representations, memory at any scale for Deleuze is best conceived in terms of movements, specifically, two entwined temporal movements, ‘the passing of the present, and the preservation of the past’ (Deleuze and Parnet, 2002: 151). For Deleuze, following Bergson, the past is not something that ceases to exist upon the succession of the present (actual) moment. Rather, the past coexists with the present in the form of memory, possessing a virtual reality of its own; the past is virtual, and as such, ‘real without being actual, ideal without being abstract’ (Proust, in Deleuze, 1988: 96).

Strictly speaking, the present is nothing other than a ‘dimension’ of the past. As Deleuze (2004: 103) explains, this is ‘the Bergsonian idea that each present present is only the entire past in its most contracted state’. Memory is not the result of a succession of present moments, but rather it is the present, the actual, that emerges from the contraction of memory. This contraction is synonymous with the passing of the present, or actualization, given as an experience (perceptions–actions) in the objective (spatial) world. Simultaneously, because for Deleuze–Bergson, the present is an interval and scission of time into two ‘jets’, the passing of the present as contraction towards the future is always accompanied by the self-preservation of the past as dilation or expansion towards the ‘pure’ (ontological) past. For simplicity’s sake, the contraction or actualization may be considered to relate ‘virtual images’ to actual objects, and the dilation, which we will call virtualization, to relate actual objects to virtual images. Image in this context has many senses, designating ‘slices’ of the world, the after-image (double) of the present and recollection. The important point is that virtual images exist along a continuum, constituting circuits interpenetrating (linking) the virtual and actual, memory and the present, actualization and virtualization (Deleuze and Parnet, 2002: 148–52).

An illustration of this theory of memory and time by way of a brief example of networked life may be helpful. Consider users logging into a social network such as Facebook to chat with each other using its native instant messaging client. Logging in, composing, sending and receiving messages involves the passing of the present, the actualization of a virtual image, for example, a wish for company based on recollections of past experiences. It is a contraction of memory involving actual perceptions–actions as communication partners interact. Simultaneously, the past also preserves itself alongside the passing present, as a dilation whereby the actual object becomes a virtual image. The display picture characteristic of digital social networks provides us with a literal instance of this passage of virtualization. The digital photograph when ‘taken’ in the first place extracts from the passing present elements corresponding to the photographer’s interest, typically a face, which it transforms and encodes as a visual idea. Uploaded into a social network profile, the visual idea takes the place of an actual face, and from there passes into a cloud of virtual images, into memory, thus contributing aspects of actual experience to the virtuality that is part of an assemblage’s duration. As another example, where users talk to each other over a social network like Skype, the sounds of different voices are transformed and encoded as sonic and linguistic ideas, and the contents of actual experience are donated as virtual images to the stream that is the past’s self-preservation.

To speak of the cultural memory of a museum assemblage is to account for the memory (and emergent properties deriving from the interactions) of its component parts, which function like so many neuronal circuits making up a brain, each virtual–actual circuit contracting and dilating at variable rates. The brain for Deleuze, however, is by no means confined to human beings, a fact made evident when he writes with Guattari (1994: 212), that even ‘the plant contemplates by contracting the elements from which it originates – light, carbon, and the salts’. In fact, the brain might as well be synonymous with actual presence and the interval between excitation and reaction – ‘the brain is wholly on the line of objectivity’ (Deleuze, 1988: 54) – with memory as its virtual ground, the ontological past from which it continually actualizes. If the brain is presence, interval, then mind is effectively to be found everywhere – ‘everywhere there are forces that constitute microbrains’ (Deleuze and Guattari, 1994: 213) – and the same must be said of memory and the flux–scission of time. The theory of memory presented here belongs to panpsychism, the philosophy of mind-in-life, for which reason it would be a mistake to understand it simply in (human) psychological terms. ⁸ The cultural memory of a museum assemblage must therefore also include the memories of nonhuman ⁹ components alongside the mnemonic activities of its human parts.

As traces of the actualization of memory, the actions–expressions in the examples of networked life above are preserved for as long as the material components last, upon which the internetwork depends. Similarly, the expressions making up the material and digital collections of museums are preserved in themselves, that is, for as long as their composite materials last. Concerning networked expressions, these distribute themselves more or less autonomously across the internetwork, preserving in themselves the capacity to affect others in ways independent of the original desires and intentions of those involved. For example, where members of a network pass away without deleting their accounts, profiles frequently become memorials, remaining affectively charged and indefinitely embedded in the interpersonal networks of the deceased. An intriguing website that emerged in this regard is MyDeathSpace, where information on dead MySpacers and links to their profiles are provided to the public. Despite their enduring autonomy and affectivity, networked expressions should by no means be confused with the entirety of memory. Memory, as this section has attempted to portray, consists of twin temporal movements, a contraction and a dilation, in which the actual is only the most contracted point of the virtual past. Conceptualizing memory in terms of two movements, actualization and virtualization, gives us dynamic cultural memory in place of static, unchanging and unreal representations of reality.

The previous sections have portrayed networked museums as assemblages constituted by the movements of memory and organized under different collectivizing logics. The discussion until now has largely been theoretical, incorporating brief examples for illustrative purposes where necessary, but leaving detailed examples to the last section on museums. The intention was to lay out a set of concepts first, with which to evaluate innovations in cultural memory practices. If it seems strange to readers to conceive of museums and memory largely in terms of the virtual past, the actualization of virtuals, and the virtualization of actuals, it is worth remembering Paula Findlen’s (1989: 63) observation: ‘The use of the term museum was not confined only to the tangible; museum was foremost a mental category and collecting a cognitive activity that could be appropriated for social and cultural ends.’ Cultural memory has several senses, one being that of the memory of a culture-in-general. In assemblage theory, we are dealing instead with historically individuated social wholes, individuals at different levels of scale – villages, tribes, suburban and urban communities, artistic practices and movements, individual nation-states and so on. Each of these assemblages corresponds to a set of cultural memories as its virtual dimension. All assemblages are composites of virtuals and actuals.

Varieties of museums

The wealth of literature on the development of collections from antiquity to the modern era makes it unnecessary for us to dwell on the extended history of museums. ¹⁰ The focus here is on the networked museum, in which repositories of physical objects are increasingly supplemented (and in some cases, entirely displaced) by collections of digital objects in the form of networked images and sounds, linguistic and non-linguistic information particles, which, by virtue of their digitization, are easily copied, manipulated and repurposed, and in this manner widely distributed, and occasionally ‘remixed’ or ‘mashed up’, often without regard for the physical locales and cultures from which the objects are derived. Networked museums may be characterized as assemblages linking peoples, locales, social forces and experiences together across time and space through cultural memory practices. If classical and modern museums have on the whole tended to be ‘object-oriented’ in their emphasis on collecting and displaying material artworks and artifacts, networked museums evince a tendency to foreground the incorporeal, informational and conversational dimensions of cultural memory.

The internetwork as living museum presents us with an image of a museum without walls, making it tempting to conceptualize collections of digital images, sounds and conversations, as well as the relations they make possible across institutional and political-economic borders, in terms of virtuality, disembodiment and separation from the objective (actual) world of bodies, actions and physical co-presence. Such a temptation should be resisted, for the incorporeal media traversing the internetwork would not exist without material components such as cables, computers, servers, screens, interfaces and electricity, not to mention actual individuals and organizations populating networked locales. At the same time, the temptation to conceive of museums as simple containers or stores of memory, whether in the form of material or digital collections, must also be resisted, as museums are networks of objects and events that facilitate acts of cultural recollection. What is distinctive about the internetwork as museum is the manner in which the field of cultural memory practices is vastly expanded to include non-institutional assemblages, and the degree of autonomy from geopolitical constraints that products of cultural memory practices may acquire upon digital distribution.

Also remarkable is that the internetwork, by providing conditions for individuals and groups to express themselves more or less ‘freely’ in relation to the circumstances of everyday life, and by permanently storing such expressions (unless deleted), affords networked publics that constitute dynamic topologies of cultural memory manipulable by users in perpetuity (coexisting uneasily with private and often corporate owners of many of the platforms in question). Photographs and videos of social events uploaded to sites like Flickr and YouTube, status updates on micro-blogging platforms such as Twitter and Tumblr, as well as display pictures and comment–conversations across social network profiles such as Facebook, record and memorialize the cultural encounters and circumstances of millions of individual lifetimes in digital codes and physical servers, with myriad translocal connections enabling populations of individual durations to coexist and thus to coalesce into composites of cultural memory proper to the lifetimes of larger assemblages. It is in this sense that the internetwork itself may be considered a living museum, or rather, museum of museums, a set of globalized and self-organizing temporal fluxes built on nodes that fire and rest like the neurons of so many collective brains.

Towards the hierarchical end of DeLanda’s organizational continuum, well-established and endowed institutions have of course been at the forefront of numerous developments in emerging networked cultures, largely enabled by their command of capital. Thus, museums such as the Whitney and the Tate have, for the better part of the last decade, supported innovation by developing networked archives and exhibitions and investing in commissions of networked art. The Tate also facilitates accessible critical discourse through its online research journal TatePapers. Then there are the examples cited earlier, Google’s Art Project as well as the engagement of major art institutions with social media platforms. A notable project that comes to mind is the Museum of London’s Streetmuseum, launched in 2010 as an iPhone ‘app’ (application) with the intention of transforming the city of London itself into an open-air museum for iPhone users. Developed by a professional creative agency, Streetmuseum draws on the museum’s collections of art and photography to illustrate over 200 actual sites in London through augmented reality. Users discover sites by using the in-built map, and once at a site, the iPhone camera overlays images from the Streetmuseum collection onto the camera view, displaying fragments of the site’s history such as buildings or streets that no longer exist, and providing historical information to accompany the images.

‘YouTube Play’ by the Guggenheim Museum is another useful example of a hierarchical institution incorporating greater participation in its activities. Launched in 2010, the project involved an open competition for a ‘Biennial of Creative Video’ using the YouTube platform. It attracted over 23,000 submissions, of which Guggenheim curators shortlisted 125 and the YouTube Play expert jury subsequently narrowed down to the top 25. The videos selected by the jury were presented at the Guggenheim Museums in New York, Bilbao, Berlin and Venice during October 2010, and are of course, available for viewing online on YouTube itself. Considered on its own, YouTube Play may be positioned a little closer to the meshworks end of DeLanda’s continuum, insofar as it is embedded within a pre-existing meshwork of video-makers, and to the extent that the competition maintained an open access policy, stipulating only internet access as a (very minimal) condition for entry. It is at the point of decision-making, however, that significant power differentials become impossible to ignore. The fact that the final decision-making remained the privilege of a handful of experts means that concerning the YouTube Play assemblage overall, a small number of nodes still ended up being much more powerful than all the others. YouTube Play is nevertheless a promising initiative, recognizing the capacity of non-professional communities to produce ‘exceptional talent’, and marking a modest movement of de-hierarchization. It could have increased participation had it incorporated crowd-sourced jurying as well.

Towards the meshworks (non-institutional) end of DeLanda’s continuum, projects such as Historypin and Tales of Things stand out as user-driven and generated networked museums. As the name suggests, Historypin is an online platform that enables users to pin photos on a shared map (powered by Google). ¹¹ Besides tagging photos to specific places, monuments and buildings, users are also able to specify the time depicted in their photos, and write stories to accompany the photos. In this way, Historypin socializes previously isolated memory practices, constructing user-generated histories of places without depending too heavily on a centralized curatorial or editorial node. Tales of Things is a network based on users creating an online museum of objects linked to videos and texts describing their histories, as well as to a world map. Then there is Panoramio, a social network which overlays landscape and cityscape photography generated by users onto Google Maps and Google Earth, and also displays the content in situ via ‘augmented reality’ through its integration with the Layar augmented reality browser.

Besides the examples discussed so far, the expressive activities of users of popular networks such as Facebook and Twitter, not to mention the millions of blogs on Blogger, Wordpress and Livejournal, all contribute to the networking of cultural memory practices, simultaneously transforming the actual locales and social relations within which they are embedded, and donating actual experiences to the self-preservation of the past. Like Historypin and Tales of Things, these assemblages have a relatively low degree of hierarchization, and are an important expansion of the field of cultural memory practices, counterbalancing tendencies in official histories and institutional memories towards censorship, selective amnesia and even historical revisionism.

While this article has emphasized the gains of meshworks relative to hierarchies over the long duration, the existence and influence of what we can call ‘power nodes’ should not be underestimated. Power nodes are immediately visible in network visualizations – they are the ones with an enormous number of connections. Search engines such as Google and major news websites such as the New York Times or the BBC are good examples of power nodes. By virtue of their centrality in many social media networks, such nodes may often function as opinion-homogenizing machines. As DeLanda (1997: 244) explains: ‘The overall effect of mass newspapers and news agencies was homogenizing. Newspapers aimed their presentation to the lowest common denominator, while news agencies attempted to create a product that would be acceptable to all their subscribers.’ Combined with the global reach of the internetwork, such homogenizing machines may exert formidable subjectivating pressures over wide swathes of networked populations.

To be fair to museums of the institutional variety, user or visitor participation may not always be a high priority, nor are high levels of participation always desirable. Museums may have very specific mandates in relation to their collections that are not served by high levels of participation. They may be charged with the task of protecting and preserving collections of fragile objects and (sacred) sites of cultural significance for instance. Also, while some arts and culture projects clearly depend on high levels of user participation to succeed, others typically require the sustained vision and direction of a single person or small team over extended periods of time. Having said all that, there are still many public arts and cultural institutions with mandates to encourage and develop greater visitor and community participation, which could learn a great deal from the examples and ideas discussed here. The sheer volume of expressive media flows traversing the internetwork suggests that there are entire communities with at least the desire to participate more fully in cultural memory practices, a decisive factor for networked participation being the emergence of (more or less) ‘free’ and easy-to-use expressive repertoires.

DeLanda’s organizational continuum, structured around two extreme tendencies of hierarchies and meshworks, allows for a nuanced social analysis in which organizational logics may be identified with regard to their consequences for user participation in cultural memory-making. Hierarchies, with their concentrations of capital and power, make large-scale innovation possible. Think for example of the Google Art Project and Google Books. Hierarchical organizations also have proven capabilities in organizing large groups to achieve planned objectives with a reasonable degree of precision. Unfortunately, hierarchies also come with a tendency towards a homogenization of differences, and the concentration of power is frequently if not characteristically abused. Meshworks, on the other hand, exemplify a diffusion of power and greater flexibility in accommodating individual differences and experimentation. While meshworks and hierarchies are distinct collectivizing logics, it is important to reiterate that they are not mutually exclusive; we are always dealing with composites. The use of ‘hierarchies’ and ‘meshworks’ in this article should be understood as shorthand for composites tending towards either end of the continuum.

YouTube Play is a useful example of how institutions may momentarily incorporate a meshwork as a semi-autonomous node. Its success depended at least partly on relatively open parameters for participation, and a necessarily expansive concept of art and artist, to engage a pre-existing meshwork of video-makers in their ‘own’ territory. Networks like Historypin and Tales of Things demonstrate how meshworks may successfully self-organize as living museums without major centralized curatorial or editorial direction. These examples of museums across the hierarchy–meshwork continuum portray a space of possibilities for cultural memory-making in an age of networks, while the expansion of concepts in the previous sections contributes a rigorous theory of cultural memory-making in terms of assemblages. In response to the popularity of the concept of virtuality in explaining the internetwork, this article also tries to bring some clarity to the understanding of virtuals and their relations to actuals. That which is virtual in networked encounters is no different from that which is virtual in face-to-face encounters. It is in every case memory that is virtual, memory and recollection as a continuum of virtual images, and it is always the brain that is actual, the interval and scission of time in which perceptions–actions (presence) and recollections coalesce.

Conclusion

I have not stressed the history of the concept of assemblages in any great detail, given the limitations on the article length. In conclusion, however, it should be said that the theory of assemblages articulated here diverges from DeLanda’s assemblage theory in some ways, just as DeLanda diverges from Deleuze (and Deleuze and Guattari) in some ways. DeLanda does not emphasize the subjective line of the assemblage to any great extent – he seems more interested in the objective line of actualization, although he does mention ‘absolute’ lines of flight. Where DeLanda designates the virtual half of the assemblage as a diagram, and the actual half as a territorialized and coded social whole actualizing the diagram, I have incorporated Deleuze’s Bergsonism to account for the virtual half of the assemblage as cultural memory. In the Bergsonist (1988: 22–3) account, everything is portrayed in terms of an ontological memory, in terms of movements of contraction towards the future (actualization) and dilation towards the past (virtualization). This is not to discount DeLanda’s rigorous reconstruction of Deleuze’s world. What the return to Deleuze–Bergson emphasizes, though, is that assemblage theory need not only be understood in terms of non-equilibrium thermodynamics (a major touchstone for DeLanda), but may also be conceived in Bergsonist terms of matter and memory, presence and an ontological past as actual–virtual composite.

As for the lessons to be learnt from an assemblage analysis of hierarchical institutional structures and experimental peer-to-peer expressions and artifacts of culture, we can take from Simon (2010) the book drop project implemented at Haarlem Oost as an example, which involved visitors expressing opinions on library books by returning them to shelves with specific tags (described in chapter 1 of Simon’s book). The project demonstrates how participation may ultimately prove incompatible with the hierarchical organization of an institution. While it was extremely popular with visitors, the library eventually removed it. This raises two questions. First, are organizational and architectural changes necessary if publics are to be more actively engaged in cultural institutions in the long term? Second, should we, as cultural professionals, not be looking more closely at non-institutional museums as legitimate and valuable assemblages of cultural memory-making?

Where to from here? For a start, in individual assemblages, we as components could make an effort to understand the workings of collective thought and agency; ‘collective intelligence’ as Pierre Lévy (1997) puts it. After all, since there is no ontological difference between a person and an entire community, we must seriously consider processes of distributed perception, action and memory proper to the scale of familial, ethnic, religious, political, and professional communities. Such an effort may reveal that the notion that order arises from top-down chains of command is not necessarily accurate; that groups of individuals using the affordances of their locales may well self-organize and achieve periods of ritual equilibrium. Further research might investigate different forms of management and organizational structure to evaluate the feasibility of self-organized initiatives as well as a shift away from planned economies of scale.

A degree of pragmatism would not go astray where hierarchical organizations consider developing participatory projects, recalling the library experiment results reported by Simon (2010). It may be better for arts institutions wishing to experiment with participation to establish a separate organization for the purposes of such work, and to experiment with ‘self-management’ strategies in place of corporate chains of command in these settings (without exposing the main institution to risks of ‘brand’ damage). It would also be worthwhile to consider the peer-to-peer theory and political economy proposed by Michel Bauwens (2005). Finally, efforts could be made not just to transpose insights from social media (e.g. strategies of ‘viral marketing’) into cultural and heritage settings, but also to study non-institutional assemblages of cultural memory-making on their own terms, as contemporary forms of networked museums. It is Findlen’s (1989: 63) notion of museum that is invoked in this article’s conclusion, the museum as ‘a mental category’, an organ of memory with its contractions and dilations.