Sociopetal and Sociofugal Space: How Spatial Arrangement Shapes Social Interaction
I first encountered the distinction between sociopetal and sociofugal space during an accelerated semester at Cornell, where I was studying environmental psychology and thinking particularly about how research might be translated into evidence-based guidance for design. What surprised me was discovering that the idea was closely connected to the work of Robert Sommer, a researcher I had admired for a long time.
Sommer had, in many ways, introduced me to architectural psychology before I properly knew what the field was. His work on personal space, seating behaviour and the subtle ways in which environments structure social relationships was one of the things that first made me realise there was an established body of research asking questions I had already begun asking as an architect: why people choose one seat over another, why some spaces make conversation effortless while others make it awkward, and how seemingly minor spatial decisions shape what people actually do.
Sociopetal and sociofugal space became particularly memorable because the idea is both immediately intuitive and considerably more complicated than it first appears.
Mini overview Origin: Humphry Osmond Publication Year: 1957 Developed through: Robert Sommer’s research on spatial behaviour, personal space and seating arrangements Core idea: Spatial arrangements can make particular forms of social interaction easier, more comfortable or more likely, while others support separation, privacy or individual activity Key spatial considerations: Orientation, interpersonal distance, visibility, furniture configuration, enclosure, movement and user control Design relevance: Healthcare, workplaces, education, hospitality, residential environments, cultural buildings and public space. Key caution: Sociopetal and sociofugal describe tendencies within a person–environment relationship. They should be understood probabilistically rather than as fixed properties of a room or deterministic predictions of behaviour. |  |
Background.
Architecture inevitably positions people in relation to one another. A chair faces another chair, turns towards a window or sits alongside a wall. A table gathers people around a shared surface. A booth creates a small inward-facing territory. Rows of seats direct a group towards the same focal point. Even before anybody enters a room, its arrangement establishes possibilities for where people might sit, who they might see and how easily they might interact.
The concepts of sociopetal and sociofugal space provide one of environmental psychology’s earliest ways of describing this relationship between spatial arrangement and social behaviour. Psychiatrist Humphry Osmond introduced the terms in the 1950s while considering the design of psychiatric environments.
Sociopetal arrangements bring people into spatial relationships that facilitate interaction. (Osmond, 1957).
 | Chairs clustered around a small table can support conversation because people are within speaking distance and can see one another. |
Sociofugal arrangements tend to separate people or orient them away from direct interpersonal contact (Osmond, 1957).
 | Rows of fixed seats facing the same direction make conversation between neighbours more difficult. Yet the value of the theory extends beyond this familiar comparison. |
The distinction is easy to visualise, but it asks a more useful architectural question:
What forms of social relationship does a spatial arrangement make easier to establish and maintain?
From the psychiatric ward to architectural psychology.
Osmond developed the concept while working at Saskatchewan Hospital in Weyburn, Canada, during a period in which psychiatric care was beginning to move beyond purely custodial models. His interest extended to the physical environment and to the ways apparently functional design decisions could affect everyday social life within an institution (Osmond, 1957).
One environment that attracted particular attention was a refurbished geriatric ward. In conventional terms, the refurbishment appeared successful: it was clean, orderly and newly furnished. Yet much of the furniture had been positioned around the perimeter or arranged so that patients sat shoulder-to-shoulder and back-to-back. The room provided somewhere to sit, but relatively few comfortable opportunities to face another person and talk.
Robert Sommer subsequently worked with Osmond and began examining these relationships more systematically. In an early study with Hugo Ross, furniture in a geriatric ward was rearranged into smaller conversational groupings around tables. The researchers observed a substantial increase in social interaction following the change (Sommer & Ross, 1958). It is a striking example because the intervention was so modest. The architecture itself had barely changed. The social geography of the room had.
Sommer later extended this work through studies of personal space and seating behaviour. Observations in a cafeteria examined which positions around rectangular tables were associated with conversation. Interaction was not distributed equally across every available seating relationship. Positions around corners, adjacent seats and directly opposing positions produced different behavioural patterns (Sommer, 1959).
This already makes the sociopetal diagram more complicated. Interaction is not created simply by pointing two chairs towards each other. Distance, angle, eye contact, activity and the relationship between the occupants all matter. Sitting directly opposite another person may facilitate conversation in one context and feel confrontational in another. Sitting at an angle allows visual contact while also making it easy to look elsewhere. Two people sitting side-by-side may communicate particularly well when their attention is directed towards a shared object.
The social qualities of a space therefore emerge from relationships between people, objects and spatial configuration rather than furniture geometry alone.
What makes a space sociopetal?
At its most basic level, sociopetality concerns spatial conditions that make interpersonal engagement easier.
Orientation is one of the clearest mechanisms. People who can comfortably see one another can establish eye contact, interpret facial expressions and recognise when somebody is preparing to speak. Sommer’s later classroom research similarly identified relationships between seating position, visual relationships and participation (Sommer, 1967a).
Distance matters because conversation operates within interpersonal spatial conventions. Somebody who is visible at the opposite side of a large room is socially available in a very different way from somebody sitting across the corner of a table.
At a larger scale, proximity and visibility can influence opportunities for encounter. Workplace research, for example, suggests relationships between proximity and face-to-face interaction, although the evidence is more complicated than a simple opposition between open and enclosed offices (Sugiyama et al., 2021).
Furniture, lighting and partial boundaries can also establish smaller territories within larger rooms. A table, rug, pendant light or change in enclosure can make a group of chairs read as a coherent social setting rather than unrelated objects on an open floorplate.
Then there is control. Movable chairs allow occupants to adjust distance and orientation according to the people present and the activity taking place. Fixed furniture determines much more of that relationship in advance. Moving a chair slightly closer, turning it towards a group or angling it away may appear trivial in plan, yet such adjustments allow people to regulate social contact for themselves.
Sociopetality is therefore better understood as a relational quality than a visual style.
A circular table is not inherently sociopetal. An enormous one may separate people by distance; a tightly packed circle may make participation uncomfortably exposed. A restaurant booth may be highly sociopetal for the four people sitting within it and strongly sociofugal towards neighbouring groups. An open office might appear sociopetal at the scale of the floorplate while its individual desks provide few comfortable opportunities for conversation. Scale changes the interpretation.
Sociofugal space has a purpose.
The term sociofugal can sound undesirable, partly because the original discussion emerged from institutions where isolation was already a concern. In contemporary architecture, however, reduced interpersonal engagement is often useful. A library carrel supports concentration. A window seat allows somebody to occupy a shared room while temporarily withdrawing from it. Gallery seating may orient visitors towards an artwork rather than towards each other. Focus rooms in workplaces intentionally reduce visual and social exposure. Even within highly social environments, people rarely want uninterrupted interaction.
Meagher and Marsh (2017) demonstrated this particularly clearly. Across a series of experiments, people who had recently experienced social ostracism showed a greater preference for sociofugal room arrangements. The desirability of the spatial configuration therefore changed according to the psychological and social state of the person encountering it. This matters because it challenges the assumption that increasing opportunities for social interaction is necessarily the preferable design outcome. A room does not possess a universally desirable level of sociability. The design challenge is often to provide different degrees of social exposure and some ability to choose between them.
Interaction is more varied than the binary suggests.
Sociopetal and sociofugal remain useful because they direct attention towards the social consequences of spatial configuration. They are nevertheless broad categories. People can converse, observe, cooperate, share attention, participate peripherally, simply occupy the same place or deliberately withdraw from one another. Recent work examining social interaction in public space similarly describes a wider spectrum of behaviours, from withdrawal and co-presence to exchange, coordinated activity and collective assembly (Mohammadi et al., 2026).
Consider a museum bench. Two strangers sitting alongside one another looking at the same painting may never speak. Calling the arrangement simply sociofugal misses something. It supports co-presence and shared attention. The interpersonal relationship remains relatively distant while both visitors participate in a common relationship with the artwork. A lecture theatre provides another example. Rows of seats discourage extensive face-to-face conversation between students, yet they efficiently organise hundreds of people around a shared focal point. The same configuration might be highly restrictive for a seminar built around discussion.
The more useful question is therefore not simply whether a space is social.
What kind of social behaviour is it trying to support?
Healthcare: interaction without compulsory sociability.
Healthcare remains one of the clearest applications of the theory because social interaction, privacy and vulnerability often have to coexist. Shared lounges, day rooms and family areas can benefit from arrangements that allow small groups to form comfortably. Yet a waiting room organised entirely around face-to-face interaction may expose patients to social contact they do not want. Somebody may prefer to sit beside a family member while avoiding eye contact with strangers; another person may seek a more protected position altogether. A richer environment can therefore offer different possibilities: conversational clusters, paired seats, perimeter positions and more sheltered settings. The principle is not simply to make healthcare more sociopetal. It is to allow social engagement without making it compulsory.
Workplaces: interaction, interruption and choice.
The workplace provides another obvious application because collaboration, knowledge exchange and informal conversation are frequently treated as organisational objectives. Spatial configuration can contribute to those behaviours. Proximity, visibility, circulation routes and the placement of shared amenities influence opportunities for encounter, and systematic evidence suggests relationships between several spatial characteristics and face-to-face interaction (Sugiyama et al., 2021). Sociopetal thinking can therefore inform team tables, project spaces, shared kitchens and informal meeting areas. Two armchairs positioned at an angle create a different conversational relationship from chairs placed formally across a desk. A shared table on a circulation route can allow brief interaction that an enclosed meeting room cannot. But workplaces also demonstrate the limits of treating sociopetality as an objective in itself. Concentration requires periods of reduced interaction. Confidential work requires enclosure. Some people prefer to regulate visibility and social exposure more actively than others. Environments saturated with opportunities for encounter can turn interaction into interruption. A useful workplace therefore contains a landscape of social conditions: places for active collaboration, peripheral participation, quiet co-presence, concentration and withdrawal.
Learning environments: start with the behaviour.
Classrooms are frequently used to illustrate the distinction. Circles and seminar tables are associated with discussion; rows with didactic teaching and attention towards an instructor. Sommer’s classroom studies found relationships between seating location and participation (Sommer, 1967a). More recent research, however, helps complicate any direct translation from furniture arrangement to behaviour. Nehyba et al. (2023), comparing circles and rows during reflective group activities, found more intensive interaction in rows under some conditions, with effects varying according to the subject and facilitator. Furniture therefore operates within a larger behavioural setting that includes teaching method, group norms, activity, familiarity and expectations about participation. For architects, this distinction matters. The task is not to identify which seating configuration is generically more social. It is to understand what people need to do, then consider which spatial relationships make those behaviours easier.
Hospitality, domestic and cultural environments.
Some of the clearest examples of sociopetal and sociofugal relationships occur in everyday settings.
A dining table draws a household together. Chairs positioned beside a fireplace create a conversational setting. An armchair facing a window allows somebody to withdraw while remaining part of a shared room. A restaurant booth encourages interaction within one group while reducing engagement with adjacent diners.
These nested relationships show how different degrees of social exposure can be created without walls.
Cultural environments add another dimension: shared attention. Museum visitors may want to observe each other, experience collective presence, discuss what they see or become absorbed individually in an artwork. Seating directed towards an exhibit may be sociofugal in relation to conversation while entirely appropriate for the experience the architecture is supporting. The distinction becomes most useful when it prompts analysis rather than classification.
What can architects take from sociopetal and sociofugal space?
For design practice, I find the theory most useful as a series of questions rather than a set of forms to reproduce:
Who might interact here? Friends, colleagues, strangers, families, students and patients bring different spatial expectations.
What type of interaction matters? Conversation, collaboration, observation, shared attention, informal encounter and quiet co-presence require different conditions.
What spatial relationships support it? Consider orientation, distance, visibility, enclosure, acoustics, circulation and shared focal points.
How much control do people have? Movable furniture, alternative positions and different levels of enclosure allow occupants to adjust their social exposure.
What happens when somebody wants something different? Privacy, withdrawal and concentration need to remain legitimate possibilities.
Read this way, Osmond’s distinction sits comfortably within a probabilistic understanding of architecture and behaviour. Spatial configuration can change the likelihood, effort and comfort associated with particular actions. It can make eye contact effortless, conversation awkward, observation comfortable or withdrawal difficult. Yet those relationships are interpreted through social expectations, activity, individual preference, culture and the immediate circumstances of the person occupying the space. This is why environmental psychology does not need to tell architects exactly what to design in order to be useful. The evidence gives us another way of seeing the problem. Architectural judgement determines what we do with it.
Designing relationships as well as rooms.
More than sixty years after Osmond first described sociopetal and sociofugal space, the distinction remains useful because it draws attention to something fundamental: architecture arranges relationships between people. Plans show walls, furniture, circulation and thresholds. They also contain possible patterns of encounter, avoidance, observation, gathering and withdrawal. The value of sociopetal and sociofugal thinking therefore lies less in deciding whether a room is social or antisocial and more in learning to look carefully at the spatial conditions surrounding interaction: who faces whom, at what distance, with what visibility, around which shared objects, and with what possibility of joining, leaving or turning away.
For architects, perhaps the most productive question is therefore:
Sociopetal for whom, towards whom, and for what?
Archontia Manolakelli is an ARB Registered Architect and Behavioural Design Researcher based in Manchester, UK. Her work embeds behavioural science into the design of everyday spaces as a quiet force for change, helping workplaces, schools, and civic environments better reflect the people who use them. Grounded in evidence-based practice, her approach bridges the gap between theory and application, academia and practice, science and intuition.
Hello. Thank you for stopping by, I hope you have enjoyed your reading! If you have any questions or feedback on this article, please don't hesitate to drop me a line on LinkedIn or via email.
References.
Meagher, B. R., & Marsh, K. L. (2017). Seeking the safety of sociofugal space: Environmental design preferences following social ostracism. Journal of Experimental Social Psychology, 68, 192–199. https://doi.org/10.1016/j.jesp.2016.07.004
Mohammadi, M., Heurkens, E., & Mohammadi, M. (2026). Social interaction in public space: A meta-narrative review. Journal of Environmental Psychology, 112, Article 103080. https://doi.org/10.1016/j.jenvp.2026.103080
Nehyba, J., Juhaňák, L., & Cigán, J. (2023). Effects of seating arrangement on students’ interaction in group reflective practice. The Journal of Experimental Education, 91(2), 249–277. https://doi.org/10.1080/00220973.2021.1954865
Osmond, H. (1957). Function as the basis of psychiatric ward design. Mental Hospitals, 8(4), 23–27. https://doi.org/10.1176/ps.8.4.23
Sommer, R. (1959). Studies in personal space. Sociometry, 22(3), 247–260. https://doi.org/10.2307/2785668
Sommer, R. (1967a). Classroom ecology. The Journal of Applied Behavioral Science, 3(4), 489–503. https://doi.org/10.1177/002188636700300404
Sommer, R. (1967b). Sociofugal space. American Journal of Sociology, 72(6), 654–660. https://doi.org/10.1086/224402
Sommer, R. (1969). Personal space: The behavioral basis of design. Prentice-Hall.
Sommer, R., & Ross, H. (1958). Social interaction on a geriatrics ward. International Journal of Social Psychiatry, 4(2), 128–133. https://doi.org/10.1177/002076405800400207
Sugiyama, T., Hadgraft, N., Clark, B. K., Dunstan, D. W., Chevez, A., Healy, G. N., Cerin, E., LaMontagne, A. D., Shibata, A., Oka, K., & Owen, N. (2021). Office spatial design attributes, sitting, and face-to-face interactions: Systematic review and research agenda. Building and Environment, 187, Article 107426. https://doi.org/10.1016/j.buildenv.2020.107426
