Critical computational relations in design, architecture and the built environment: editorial

ABSTRACT

Computation in design, architecture and the built environment, and its practices, methods, and tools frequently offer ‘neutral’ and ‘optimized’ techno-solutions to (social) design problems. Such a portrayal of these computational infrastructures as neutral solutions that open participation in design hides the social, political, and environmental entanglements involved in their creation and expansion. This special issue spotlights power relations between computational practices, technology infrastructures, knowledge, and their reproductions of bias at multiple scales.

KEYWORDS:

• Computation
• design
• architecture
• critical studies
• technology

Introduction

The presence of computation in almost every corner of contemporary life is becoming indisputable. A growing reliance on computational approaches and paradigms to operate and transform sociotechnical systems exerts influence at different scales. This includes the scale of individual interactions with digital devices for spatial navigation, institutional interactions where organizations are being managed according to digital metrics, and the planetary ones where satellite data is used to predict and simulate environmental conditions. Following cultural theorist David Golumbia, computation is simultaneously a ‘metaphor, method, and organizing frame’ of the uppermost state yet under-scrutinized in specific domains (Golumbia 2009, 1). In a Foucauldian governmentality sense, critical scholars have argued that computation and its rhetoric legitimize both novel and established social and political powers (see Amoore 2020; Bridle 2018; Luque-Ayala and Marvin 2020). Computational techniques can not only extend our thinking and imaginations beyond the known and achievable, but also ‘give to each according to their need and to each according to their ability’ (Berry 2011, 10). This, as many empirical studies of computational and algorithmic systems reveal (e.g., Benjamin 2019; Buolamwini and Gebru 2018; Burrell 2016; Jaton 2021), has the potential for inequality and opacity when these new technologies are introduced in different social domains. While computation requires discreteness in the form of abstracted relations between quantifiable entities, it exceeds the plain operation of number crunching by enhancing different media and technologies with new qualities (Fazi 2018).

These qualities suggest that computation and the digital are potentially abstract, but as digital humanities scholar Beatrice Fazi proposes, ‘the becoming of computation corresponds to actual discretising procedures of quantitative systematisation’ (ibid., 56). Prominent paradigms like smart cities and generative design promise precisely such quantitative systematization through computational processes to municipalities, urban planners, governments, and technology developers. Urban computation in this form ‘can also advance a techno-managerial focus’ (Luque-Ayala and Marvin 2020, 222; original emphasis). Critical geographers Andres Luque-Ayala and Simon Marvin’s study of urban computation is one of many examples revealing how computational processes—usually originally derived outside the urban or many other social contexts—increasingly concentrate power and ‘the retreat of that power into ever more narrow domains of experience’ (Bridle 2018, 34). It is precisely these computational relations, i.e., ‘the tensions, contradictions, and limits of the coming together of computational logics and these assemblages’ (Luque-Ayala and Marvin 2020, 3), that this special issue seeks to uncover in the context of design, architecture, and the built environment.

When it comes to computation in this context, practices, methods, and tools frequently offer ‘neutral’ and ‘optimized’ techno-solutions to (social) design problems. Portrayal of these computational infrastructures as neutral solutions that open participation in design hides the social, political, and environmental entanglements involved in their creation and expansion (Benjamin 2019; Bridle 2018; Nakamura 2014). This narrative of neutrality conceals power that computer-aided design (CAD) software monopolies and technology providers hold (Cardoso Llach 2015). Neutrality also obscures these platforms’ embedded values and their relations to inequality, injustice, and racism. Another narrative and mission of CAD industries and their computational technologies is the optimization of processes. As media and communications scholar Alison Powell (2021) notes, optimization is a push for technological development. However, not everything can and should be optimized, especially when it begins to curtail the social (ibid.)

Our call for articles in this special issue on Critical Computational Relations in Design, Architecture and the Built Environment, therefore, sought to attract contributions and studies that uncovered power relations between (commercial) design technologies, computational design practices, technology infrastructures, knowledge, and their reproductions of bias at multiple scales. The issue asks, how might we interrogate the values and relations embedded in computational design tools and processes? What does optimization really mean, what does it look like on the ground, and who benefits? The special issue’s goal is to shed light on and conceptualize power relations between computational design, the built environment, and society.

Critical computational relations: themes, methods, and sites

We bring forth several matters for discourse and critique of computational practices: first, their contribution to concealing and sometimes even erasing human labour, agency, and decision-making; second, their reconfiguration of values, societies, and narratives about social life; third, the embedded values and motivations of dominant groups with power in computational systems; fourth, the problem of technical framings of the social; and, finally, a desired sense-making through embodied knowledge and spatial, material practices.

Different computational processes have relied on different forms of human labour, from the early moments of forming a sociotechnical system to its implementation and later maintenance. As Lilly Irani has demonstrated in her study of the Amazon Mechanical Turk platform, these systems involve multiple forms of human work that are not equally valued and represented: ‘some people are employers, entrepreneurs, and programmers […], and others simulate computation for them’ (Irani 2015, 226). It is precisely this simulation of computation—more recently dubbed as automation-as-a-service or AI-as-a-service—that often gets invisibilized to ensure that a computational enchantment persists (ibid.; Boeva et al. 2023; Tubaro, Casilli, and Coville 2020). Human labour performed for and in computational processes includes data curation and production, data cleaning and sorting, data editing and fixing, making subjective decisions based on expert knowledge, making links across fields, concepts, and technologies, and building physical and digital infrastructure, to name a few (Bates, Lin, and Goodale 2016; Plantin 2019; as well as Dessewffy, Schikowitz, and Davies, in this issue; Hasey, Rhee, and Cardoso Llach, in this issue).

By centering on technological advances and optimization, the human labour and input required to make these systems perform, human welfare, care, and the conditions under which humans work are rarely brought to the fore. The result is little examination and critique of the whos, whats, and wheres of labour and their links with technology. Human labour that makes automation work, however, is very present. Exploited and devalued labour tied to histories of enslavement, indentureship, colonization, and racism has merely been displaced from the factory floor to computer screens at multiple sites including the Global South (Atanasoski and Vora 2019; Birhane 2020; Miceli and Posada 2022; Nakamura 2014). In a similar manner, computation in design and architecture masks diverse labour practices, knowledge practices, and transitional, unstable, and uncertain moments involved in making a building or physical environments (Kitchin, Young, and Dawkins 2021; Mommersteeg 2022). Very often, primacy is ascribed to these (novel) computational design and decision-making systems due to their numerical description of the social, flattening, for instance, the urban scale to a few controllable metrics and parameters (Carlsson 2022; Charitonidou 2022; Kropp, Braun, and Boeva 2022; Mattern 2021).

Additionally, creators of commercial computational systems and logics have decision-making powers that determine who and what is visible, and who and what is invisible in data, interfaces, and simulations (Cardoso Llach 2015; Chun 2013; see also Dessewffy, Schikowitz, and Davies, in this issue; Savić, in this issue). This begins at the level of abstraction that defines principles of computer programming (Malazita and Resetar 2019) and translates into ‘othering practices’ of software developers imagining future users of computational systems (Canizares 2020). Computational practices allow for the splitting of technical knowledge and what it produces from the larger sociocultural and political systems it is embedded in, thereby concealing how they are implicated in potentially harmful data-based and automated decision-making, including the built and physical environment (Gabrys 2020; Önal 2020; Sadowski 2020).

Most hardware and software systems used in our fields are privately owned, driven by profit, and not answerable to the public. Creators of these systems spread their values, agendas, motivations, and ways of seeing, knowing, and organizing knowledge, work, and the social through computational logics and the terms associated with them. The prospect of integrating artificial intelligence, automation, or robotics presented in a techno-optimistic manner, while gradually gaining infrastructural sovereignty and reconfiguring organizational structures via ‘disciplining’ application programming interfaces (APIs) and cloud services that intermediate and connect, continues to be ignored in the field of design and architecture. These integrated ecosystems made of corporate-driven computation should open up to the scrutiny of their political techno-economy in the same way digital and social media have been explored (see e.g. Bucher 2013; 2020; Helmond 2015; van Dijck, Nieborg, and Poell 2019).

Computational tools, systems, and work often redefine and reshape social life, identity, and participation. By framing social problems as technological ones with technological solutions, we reconfigure civic action and engagement by our publics, and reduce public debate and demand for public services to polite digital participation through interactive public displays (Powell 2021). These systems can mislead publics to think that civic participation is the goal when, in fact, it is not (see e.g., Savić, in this issue). Computational systems can reconfigure the identities of communities in a way that their identities are constructed around relationships to technology, shifting values from social/communal ones to more economic/individualistic ones.

Critical feminist and intersectional perspectives have, however, sought to deconstruct computational practices and systems by playfully engaging with computation’s ‘hidden’ and equally necessary materiality, thereby giving voice and presence to mostly marginalized actors, practices, and issues, such as the weaving of computer’s memory core (Rosner et al. 2018), feminist hacking for open source technology development (Dunbar-Hester 2020; Toupin 2016), human-robot knitting collaborations (Treusch 2021), the use of grease pencils in digital fabrication (coons & Ratto 2015), or the cultural histories of wire-bending in the Trinidad Carnival (Noel 2020; Noel, Boeva, and Dortdivanlioglu 2021). These research projects reveal ways to account for differences—bodies, gender, race, culture, geography—in computational relations (see also Lee, in this issue; Voigt, in this issue).

Technical descriptions and their framings of the social, thus, require robust understandings through the inclusion of social science and humanities-based methods. Building on historical knowledge and cultural context to make sense of and interpret computationally-derived results can help overcome technical deficiencies. Simultaneously, externalizing history in physical artifacts such as styles of architectural ornamentation (Özkar 2020) or Victorian-style dresses and female dressmakers’ stories by exploring their patents as a technolology (Jungnickel 2023) can become a method for computational design and exploration.

The issues and topics discussed here are not conclusive; they only capture the scholarly worlds and research closer to the editors. While we aim to address what seems pertinent yet underexplored in current debates on computational relations in design, architecture, and the built environment, more topics outside this special issue’s scope require further attention. For instance, computational processes in design and architecture are largely envisioned in a techno-optimist way to resolve planetary problems related to climate change and sustainability (Wiessner 2022). At the same time, these imaginaries and solutions often lack discussions on how computation connects to existing practices of extraction and depletion—extractivism—of natural resources by technology providers, e.g., the data centers needed for large-scale computation (Bridle 2018; Velkova and Plantin 2023) or the rare minerals needed for high-tech developments (Arboleda 2020).

Computation in design and architecture tied to its commercial software technologies and their providers also involves outsourced, invisible, and underpaid labour similar to other IT domains. The more these systems converge and become embedded into Internet-based infrastructure like cloud platforms, the more inscrutable labour operations behind them—e.g. data annotation for machine learning—may become for users. Additionally, questions related to form, geometry, or aesthetics were not directly addressed. However, the persistent polarization between ‘parametricism’ (the individual) and efficiency paradigms (the standard) interrogates the computational relations behind that and who benefits, in what way, and from what. Finally, questions addressing the centers and peripheries in design computation have also not been discussed in-depth in this issue.¹ We notice that defining sites of (design) computation remain in the Global North. This may be a result of advanced higher education degrees that prioritize curriculum in new technologies, resources made available, and the location of architecture and engineering firms as sites of application and employment. Given the interwoven relations between computation, labour, knowledge, and their environmental entanglements, we believe that more research on these topics is needed.

Contributions

The special issue includes contributions that position computational design and technologies in wider sites of world-building and their power relations. The five articles give critical perspectives and insights into computation’s relations and entanglements with architecture, design, the built environment, and society. Each article combines a specific empirical case with a critical reading of its computational relations. These include accounts of data practices involved in design simulations, the application of machine learning (ML) and digital humanities methods in studies of form and heritage as well as unearthing techno-optimist ideas of digital architecture, the conjuncture of the virtual and material through the expansion of computational processes with embodied practice, and on feminist hacking practices carried out to counteract heteronormative values of the digital being further scripted into urban design. Moreover, the articles propose innovative mixed-methods approaches on the cusp of the technical and social.

The first article, ‘Techno-optimism and optimization in media architecture practice and theory,’ by researcher and trained architect Selena Savić, examines how narratives of optimization and efficiency in the media architecture community propagate the field’s techno-optimistic values and aspirations. By combining topic modeling techniques on social media posts with a review of media architecture publications, Savić argues that the profession has reframed governance and civic participation into technical problems that can be addressed with technology-driven solutions while making invisible socio-political problems that ought not to be framed in technical terms. Just as this article used machine learning (ML) methods to detect patterns in text sources and enable their exploration, the following one also applied ML methods to detect patterns and enable exploration, but this time patterns in architectural features based on a dataset of 3D digital models.

In ‘Form data as a resource in architectural analysis: an architectural distant reading of wooden churches from the Carpathian Mountain regions of Eastern Europe,’ computational design researchers Michael Hasey, Jinmo Rhee, and Daniel Cardoso Llach present a novel method for critically analyzing large historical datasets of 3D architectural forms combining data curation, digital modeling, deep learning, subjective decision-making, and direct engagement with archival materials and oral histories with local experts. Through a case study of wooden churches from the Carpathian Mountains, the authors gain insights into their styles, geographic distributions, and evolutions by combining technological and humanities-based methods for robust results, interpretations, and evaluations while critically reflecting on the time and labour involved in constructing datasets.

Data labour and how it informs computational results is at the heart of the third contribution. In their article ‘Tracing (in)visibilising practices: engaging with simulations for architecture and spatial planning,’ science and technology studies researchers Esther Dessewffy, Andrea Schikowitz, and Sarah R. Davies examine the development of simulations for architectural design and spatial planning in academic contexts. By conducting ethnographic work and producing ethnographic vignettes, the authors reveal the politics of (in)visibility in making simulations and demonstrate how these examinations render power relations underpinning simulation practices visible. Their study makes a case for further examination of other computational practices like digital design and fabrication, robotic fabrication, and building information modeling, in which the amount of human labour is often made invisible. While this paper foregrounds the human labour behind computational simulations and their erasure from view, the fourth paper centers on the social aspects of computational material practices.

Computational design researcher Yi-Chin Lee, in ‘Digital tufting bee: expanding computational design boundaries through collective material practice and social play,’ challenges dominant design and technology paradigms of innovation, efficiency, individualism, universalism, and solutionism by centering social interactions and play through a collective material practice. The article draws on a series of design workshops focused on the textile craft of machine tufting to explore the material embodiment of computation. Lee demonstrates how centering collective effort in computational making practices—rather than individual achievement—can enlighten research and researchers on social meaning, connections, values, and interactions. Similar to Lee’s article bringing emphasis on community in computational practices, the fifth and final contribution highlights the importance of feminist hackerspaces and their communities in digitized cities.

In ‘We build this city on rocks and (feminist) code: hacking corporate computational designs of cities to come,’ urban designer and researcher Maja-Lee Voigt argues for feminist hackerspaces as urban co-creators in digitized cities, broadly termed as smart cities. Computational logics in algorithmic systems, cloaked as neutral, often represent the needs, values, and political motivations of dominant groups with social, spatial, political, and economic power. While underrepresenting the needs of marginalized groups, these systems make marginalized groups—those not white, cis-normative, not disabled, or young—extremely visible. Conducting ethnographic research in feminist hackerspaces in Germany and Austria, Voigt describes how (future) smart cities can be technologically, culturally, and spatially hacked to create diverse realities and futures by building infrastructures and spaces that challenge biased, gendered, and racialized representations in technical systems. This article not only emphasizes the need to challenge the normative in computation; it also relates to the opening contribution by Savić by highlighting how profit and power are embedded in digital infrastructures cloaked even in the civic participation in our cities.

In closing, this special issue opens up questions and issues addressed in research within computational design and digital architecture to a smaller extent. Much attention has been given and continues to go to proposals on emerging technological developments and methods seeking to transform design, architecture, and the built environment. At the same time, their subtle convergence with and, at times, subjugation to practices, principles, and values of the computational domain remains unseen. The articles and their contributors illustrate the importance of opening up to more critical inquiry by bringing up issues on the margin, demonstrating computational methods’ limits, and asking this community how much and when computation is needed in relation to the social. In the end, we should not forget that computation is also about rendering subjects and things programmable, and thus they can become more vulnerable and fragile. Sometimes not all that can be programmed should be.

Acknowledgments

We like to thank all anonymous reviewers for their immensely constructive and thoughtful comments on the papers in this special issue and the participants in our panel at the 2021 4S Society for the Social Studies of Science Annual Meeting for our fruitful discussions.

Correction Statement

This article has been corrected with minor changes. These changes do not impact the academic content of the article.

Additional information

Funding

The work of the first author and co-editor is partially supported by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) under Germany's Excellence Strategy – EXC 2120/1–390831618.

Notes on contributors

Yana Boeva

Yana Boeva is a Postdoctoral Researcher at the Institute for Social Sciences and the Cluster of Excellence Integrative Computational Design and Construction for Architecture (IntCDC) at the University of Stuttgart. She is part of the IntCDC social sciences and humanities focus area, where she examines processes of algorithmization, datafication, and platformization in architecture and construction, the transformation of skills and practice through computational design and automation, and their intersection with urban development and sustainability.

Vernelle A. A. Noel

Vernelle A. A. Noel is a design scholar, architect, artist, and Director of the Situated Computation + Design Lab at the Georgia Institute of Technology (Georgia Tech). She investigates traditional and digital practices, interdisciplinary creativity, and their intersections with society by building new frameworks, methodologies, and tools to explore social, cultural, and political aspects of computation and emerging technologies for new reconfigurations of practice, pedagogy, and publics.

Notes

1 Overall, this topic has remained on the margins in design computation research. A notable exception is the special issue on Other Computations of Dearq journal (see Cardoso Llach and Burbano 2020).