On urban trajectology: algorithmic mobilities and atmocultural navigation

ABSTRACT

In this piece, we introduce the notion of ‘atmoculture’ as a conceptual tool to analyse the new forms of mobility supported and enacted by digital algorithms. In historical perspective, we analyse how modernity has created a movement-space where the problem of finding one's way through an increasingly ‘displaced’ urban space first emerged, with noticeable psycho-social consequences. Reconstructing the new digital media as a continuation of this spatial imagination, we seek to zoom in on the forms of mobility facilitated by digital algorithms. Urban digital navigation, we suggest, proceeds in parallel with a reorientation of the urban experience towards atmospheric considerations, maximizing safety and pleasure in the user's encounters with the environment. In this context, atmoculture appears a spatial-aesthetic, psycho-cultural, and bio-technological milieu that prepares space for convenient navigation. We discuss a number of consequences: first the disburdening effect, whereby subjects delegate to a number of perceptions and decisions to algorithms, expropriating the natural problem-solving aspect of subjectivity; second, the invisible transformations of urban space due to the biases that are built in algorithms themselves; third, the tensional, even contradictory outcomes of atmocultural expectations, whereby the goal of a secure and pleasant environmental interaction is undone by the very quantity of information provided and the level of alertness required from the user.

KEYWORDS:

• Spatial perception
• urban mobility
• new media
• hodology
• urban navigation
• urban atmospheres
• atmoculture

Acknowledgment

Andrea Pavoni's research is funded by FCT/MCTES under CEEC Individual contract [CEECINST/00066/2018/CP1496/CT0001].

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 We wish to acknowledge two anonymous reviewers appointed by the journal for very stimulating comments and remarks.

2 A classic illustration is Google Maps’ ‘areas of interest’, which, it has been remarked, are ‘less about how to get around than about where to go’ (Grabar 2016).

3 There is a rich and growing literature on urban atmospheres inspired by the three main directions of ‘atmospheric’ thinking: the French tradition (in particular the work of Jean-François Augoyard, Jean-Paul Thibaud at the CRESSON lab in Grenoble), the German tradition (in particular the work of Hermann Schmitz, Gernot Böhme, Peter Sloterdijk), and the more recent Anglo-Saxon tradition (in particular the work of Ben Anderson, Derek McCormack, Kathleen Stewart). We have reviewed these developments elsewhere (Pavoni and Brighenti 2018). For a useful introduction, see also Bille, Bjerregaard, and Flohr Sørensen (2015). Similarly, mobility has been constituted into a veritable sociological paradigm by the Lancaster school associated with John Urry, Mimi Scheller, and others (e.g. Sheller and Urry 2006, 2016). Human geographers have also developed the mobility paradigm extensively, see e.g. Cresswell and Merriman (2011); Merriman et al. (2013), and Cresswell (2011, 2012, 2014).

4 ‘Between object and subject, between objective and subjective there is an enormous gap: the trajective [le trajectif]. Object, subject and trajectory are one single being’ (Virilio in Offner, Sander, and Virilio 1991, 48).

5 Incidentally, a similar idea is foreshadowed in Freudian psychology, specifically in the notion of Besetzung (energy investment, or ‘charge’). While we cannot venture more extensively into this topic here, arguably Freud's Besetzung laid the foundations of a fully ‘cathectic’ approach to the urban psyche.

6 Specifically, Carter (2002, 9) develops this argument reflecting on the etymological co-presence, in the word agoraphobia, of the notion of agorà (the public square) and its dialectical opposite, the agròs (the countryside, the ‘faraway place’), arguing that since its inception the agora is ‘twinned with wilderness’, whereas the repression of this constitutive wilderness is typical of the modern imaginary. Esposito (2002) develops a similar argument, focusing on the ‘impossible’ relation between community and immunity.

7 We refer here to the difference between map and diagram proposed by the anthropologist Peter J. Wilson (1988). According to him, whereas hunter-gatherer societies are intrinsically ‘map-making’ societies, insofar as their problem is fundamentally one of reaching out for resources scattered over large territories, domesticated societies should rather be considered as ‘diagrammatic’ or ‘recipe-based’: they function through instructions and templates allowing to perform certain tasks – such as for instance, ‘How, what, and when to plant, and where; how, where and when to herd; who should (has to, has the right) to be where and to have what … ’ (Wilson 1988, 153), and so on. A domestic life then requires a diagrammatic way to organize, connect and integrate movement across its ‘wild’ in-betweens: something like an infrastructure of mobility is called forth. To avoid confusion, the word diagram is used here in a different sense from Gilles Deleuze's homonymous notion. In Deleuze, the diagram has to do with the immanent surfacing of a ‘consistence’ among the different parts of a given formation, for instance a social group, a territory, or more generally a ‘machine’ or ‘assemblage’. The diagram is what keeps a formation together, it is its way of working, or actual mode of functioning. While Wilson's connotation of the term is different, its algorithmic premises, as we are to see, are remarkably similar. Recently, in this vein, Matteo Pasquinelli (2019) has characterized algorithms as ‘an abstract diagram that emerges from the repetition of a process, an organization of time, space, labor, and operations: it is not a rule that is invented from above but [one that] emerges from below’.

8 Cerdà's intention to domesticate the urban in its entirety is later echoed in Le Corbusier's [1887–1965] vision of fast urban mobility unhampered by the unpredictable chaos of the urban crowds and the intermingling of pedestrians and cars (Adams 2014). As Le Corbusier (1987[1924], 168) put it in Urbanisme, ‘traffic can be classified more easily than other things. Today traffic is not classified – it is like dynamite flung at hazard into the street, killing pedestrians. Even so, traffic does not fulfil its function. This sacrifice of the pedestrian leads nowhere’.

9 Specifically, imageability is defined as ‘that quality in a physical object which gives it a high probability of evoking a strong image in any given observer’, that is, an object able to polarise a significant valence – not necessarily seen, but in any case ‘presented sharply and intensely to the senses’ (Lynch 1960, 9).

10 As known, the word ‘cybernetics’ derives from the Greek kybernetes, i.e. steersman, which in Latin translated as gubernator: the verb to govern initially developed from the field of navigation, and cybernétique was introduced by André-Marie Ampère in nineteenth-century French to designate ‘the science of governing humans’.

11 See for instance the following passage: ‘Let the mishap of disorientation once occur, and the sense of anxiety and even terror that accompanies it reveals to us how closely it is linked to our sense of balance and well-being. The very word “lost” in our language means much more than simple geographical uncertainty; it carries overtones of utter disaster’ (Lynch 1960, 4). One can contrast this view with the Situationists’ practice of dérive, or drifting, which proceeded through the deliberate production of spatial disorientation as a prelude to urban play and discovery.

12 In the original: ‘L’être percevant est le même que l’être agissant : l’action commence par une résolution des problèmes de perception ; l’action est solution des problèmes de cohérence mutuelle des univers perceptifs’.

13 In the original: ‘le chemin est à la fois monde et sujet’ (Simondon 1995[1964], 209).

14 In this sense, we may also understand better Simondon's critique of Lamarckian evolutionism, where the process of adaptation is understood as carried out by pre-formed individuals in pre-given environments: a ‘biology without ontogenesis’, glosses Simondon (1995[1964], 208).

15 ‘Just as entropy is a measure of disorganization, the information carried by a set of messages is a measure of organization’ (Wiener 1950).

16 Literally, the individual is that which cannot be cut down into pieces – or better, that which cannot be cut down without changing nature. By contrast, the dividual is divisible without undergoing a change in nature. This makes it, to some extent, scale-less, and always mixable, or even always currently mixed: a new principle of social composition is envisaged.

17 As Kasy (2019) explains in different terms, machine learning for the most part is grounded on two basic concepts: regularisation, that is, the selection of relevant patterns and the discounting of irrelevant ones; and tuning, that is, the comparison of ‘the predictions for the validation data to the actually observed outcomes’.

18 As Matteo Pasquinelli and Vladan Joler write, ‘a statistical model is said to be trained successfully when it can elegantly fit only the important patterns of the training data and apply those patterns also to new data “in the wild”’ (2020, 10).

19 The artificial intelligence paradigm devised by I.J. Good, Marvin Minski and others in the 1960s was mostly moulded upon a formalistic vision of human rationality. The neural network model changed the approach, reformulating of the relation between the machine and the world, ‘basing the performance of prediction on the world itself’ and in this way ‘renewing the adaptive promises of the reflection machines of cybernetics’ (Cardon et al. 2018, 30). For an interesting exploration, see also Errol Morris’ 1997 feature film, Fast, Cheap & Out of Control covering Rodney Brooks’ work in robotics.

20 As Cardon et al. (2018, 9) summarise, ‘The characteristic feature of the architecture of these machines is that their coupling with the environment (the world) is so organic that it is not necessary to grant the calculator its own agentivity’. In his useful introduction to the concept, Pedro Domingos (2015, 7) offers an illuminating comparison – unwittingly, perhaps, a fully vitalistic one: ‘In farming, we plant the seeds, make sure they have enough water and nutrients, and reap the grown crops. Why can't technology be more like this? It can, and that's the promise of machine learning. Learning algorithms are the seeds, data is the soil, and the learned programs are the grown plants. Th­e machine-learning expert is like a farmer, sowing the seeds, irrigating and fertilizing the soil, and keeping an eye on the health of the crop but otherwise staying out of the way’.

21 This, he remarked, is the opposite of an ‘hypertelic’ adaptation to the task that deprives the machine of flexibility. Increasingly pushing forward the constraints of the technical evolutionary phyla they belong to, open machines are able to assume within themselves a number of ‘critical points’, dealing with margins of indetermination to be temporarily and strategically ‘localised’. However, Simondon warned that, contrary to cybernetics, the analogy between living things and technical objects cannot be stretched too far: while the living thing is an actually concrete reality, the technical object contains an ‘abstract’ part, the partial concretisation of which corresponds precisely to the given degree of technical evolution. Reasoning with Simondon, until machines will not be in measure of posing the coordinates of their own vital problem, they will always contain a degree of abstractness.

22 ‘By “the experience of infrastructure,” we point to the ways in which infrastructure, rather than being hidden from view, becomes visible through our increasing dependence upon it for the practice of everyday life. By “the infrastructure of experience,” we want to draw attention to the ways in which, in turn, the embedding of a range of infrastructures into everyday space shapes our experience of that space and provides a framework through which our encounters with space take on meaning’ (Dourish and Bell 2007, 417).

23 Sloterdijk has traced this spatial genealogy in his remarkable three-volume spherology (2011, 2014, 2016). We engaged more directly with Sloterdijk's work elsewhere (Brighenti and Pavoni 2018, 2019).

24 For a similar take on control as ‘environmental modulation’, see the growing literature on so-called environmentalities, inspired by Foucault's lectures on biopolitics and Deleuze's postscript (e.g. Anderson 2012).

25 In this sense, Holert and Mende (2019) write: ‘Navigation, instead of framing or representing the world, continuously updates and adjusts multiple frames from viewpoints within the world. Navigation in the digital realm is the modelling and mapping of an elusive environment – in the service of orientation, play, immersion, control, and survival’. See also Galloway (2006).

26 For remarkable examples of urban governance tailored to citizens as users and explicitly using gamification strategies, see for instance the cases of Gainesville, Florida, and San Francisco (Budds 2016a and 2016b).

27 The world is rendered as a vector space supporting a topological system of pattern recognition based on statistical proximities (Topological Data Analysis). ‘According to Yann LeCun, the goal of the designers of connectionist machines is to put the world in a vector’ (Cardon et al. 2018, 24).

28 While few would doubt that behind the interface, multiple informatic protocols intersect, it is rather harder to convey the idea that such protocols are not innocent, and that behind their technocratic functionality there are always asymmetries. Authors such as Noble (2018), Benjamin (2019) and Espeland and Yung (2019), for instance, have drawn attention to the racism that may be encapsulated in algorithmic language, and made accordingly ‘infrastructural’ and invisible.

29 Curiously, the proponents of the word ‘nomophobia’ for this psychic condition seem to have been unaware about the earlier meaning of the word, referring to a fear of the law (nómos).

30 We have dealt more extensively with some of the legal and political consequences of atmoculture elsewhere (Brighenti and Pavoni, 2020).

31 ‘The discernment that Kant called understanding (Verstand) has been automatized as the analytical power delegated to algorithms executed through sensors and actuators operating according to formalized instructions that lie outside any intuition in the Kantian sense – that is, outside any experience’ (Stiegler 2018, 28).

32 ‘This figuring-out of connections was one of the key skills and preoccupations of residents inhabiting popular districts’. Simone is referring to African and Southeast Asian cities, such as Khartoum, Kinshasa, or Jakarta. Admittedly, our discourse is grounded in an eminently Western genealogical trajectory; nonetheless, the planetary dimension of both contemporary urbanization and global computation means that it can be applied to a variety of non-Western contexts. Although it takes place in very different ways, atmoculture is a general reality of the contemporary urban.

33 An example in this vein is for instance fairbnb, https://fairbnb.coop/.

Additional information

Notes on contributors

Andrea Mubi Brighenti

Andrea Mubi Brighenti is Professor of Social Theory and Space & Culture at the Department of Sociology, University of Trento, Italy. Research topics broadly cover space-power-and-society. His most recent book is co-authored with Mattias Kärrholm, Animated Lands. Studies in Territoriology (University of Nebraska Press, 2020). His research website is www.capacitedaffect.net

Andrea Pavoni

Andrea Pavoni is Research Fellow at DINÂMIA’CET ISCTE-IUL, Instituto Universitário de Lisboa, Portugal. His research explores the relation between materiality, normativity and aesthetics in the urban context. He is a fellow at the University of Westminster Law and Theory Lab and associate editor at the journal Lo Squaderno, Explorations in Space and Society. His book, Controlling Urban Events. Law, Ethics and the Material, is out with Routledge.