Embodiment and phenomenology: meaning lives in engaged practice, not the head

Embodied interaction, the subject of this sixth of ten units, makes one consequential move. It locates meaning in the skilled, unreflective coupling of body, artefact and world, rather than in a representation held inside the user's head. If that is right, physical form and practice become design material, and the user's internal model becomes the wrong unit of analysis. The move is well argued and has changed what designers notice. Its cost is scope: embodiment names a family of claims, so the framework can be claimed for almost any design and predicts little. The diagnosis survives. The programme earns its place only where the sense of the word is specified tightly enough to yield a claim that could fail.

The move: meaning is made in engaged practice, not represented in the head

Dourish (2001) gives the programme its fullest statement. Drawing on Heidegger, Schutz and Wittgenstein, he takes the phenomenological primacy of practice over abstract cognition. He defines embodiment as the way physical and social phenomena unfold in real time and real space, not as physical reality alone. Embodied interaction is organised around the creation and communication of meaning and the maintenance of practice. Tangible and social computing, he argues, share that foundation.

Winograd and Flores (1986) made the earlier move: the rationalist tradition behind artificial intelligence and software design misrepresents understanding. Drawing on Heidegger, Maturana and Austin's speech-act theory, they held that understanding is a pre-reflective engagement with a world that cannot be fully formalised. They concluded that design lies in the readiness-to-hand of tools, and in anticipating their breakdown in use. Dreyfus (1992) supplied the ground. Rule-following over symbols cannot capture the inarticulate background of skilled coping, so such systems succeed only in restricted, special-purpose tasks. Svanaes (2000) wrote the tradition for designers, working out a phenomenology of interactivity in three experiments. He proposes kinesthetic thinking as the designer's mode of reasoning, and the interaction gestalt as its unit. He reports that the tools of the time did not support either. In the first, 7 participants explored all 38 examples of Square World, a small drawing environment. They described aloud what they were doing, while their mouse actions were annotated to the tape, which gave 266 cases for analysis. The third ran over 4 sessions of 45 minutes spaced about two weeks apart, in which the participants designed a tool themselves.

Tangibility becomes a spectrum a task can sit on

Fishkin (2004) refuses to treat tangibility as binary. His taxonomy places systems on two axes, metaphor and embodiment, and makes tangibility a spectrum, so the question becomes where on it a task belongs. The scale is populated rather than empty. Embodiment runs from full, where the input object and the output coincide, through nearby and environmental to distant. The marble answering machine sits at the environmental level: its physical marbles stand for voice messages, and are moved between wells to replay a message or return the call. Metaphor is graded in the same way, as none, noun-based, verb-based or both. Hornecker and Buur (2006) move from the object to the setting. They define tangible interaction as embodied interaction with tangible manipulation, physical representation of data and embeddedness in real space. Their framework interweaves the material and the social, over three case studies. Klemmer, Hartmann and Takayama (2006) give the design consequences as five themes, thinking through doing, performance, visibility, risk and thick practice. That is the promise tested next.

The promised empirical test returned no clean benefit

Two papers tried to make an embodiment claim testable. Hurtienne and Israel (2007) define intuitivity rather than assert it. They propose image schemas and their metaphorical extensions, from cognitive linguistics, as a taxonomy that overcomes limits of the earlier approaches. Tangible interface examples and a worked case illustrate it. Antle, Corness and Droumeva (2009) build the same responsive auditory environment twice: once with an embodied metaphor in the mapping between movement and sound, and once without. They report qualitative findings about the complexities of designing intuitive interaction, not a clean benefit from the metaphor. The experiment varies the mapping, while the artefact's physical and practical character, which the framework recommends, is unmanipulated.

The word has more than one sense, and the sense decides the claim

Wilson (2002) distinguishes six claims made under the heading of embodied cognition. They are situatedness, time pressure, off-loading cognitive work onto the environment, the environment as part of the cognitive system, cognition for action, and body-based off-line cognition. She finds the first three and the fifth at least partially true, and calls the environment claim deeply problematic. She calls the body-based claim least attended yet possibly the most powerful. A claim about embodiment is therefore only as strong as the sense it uses. Fishkin (2004) demonstrates the need, noting that his narrow sense must not be confused with Dourish's, which would incorporate all tangible interfaces as a small subset. Kaipainen and colleagues (2011) add a further sense. Their enactive system couples two elements holistically through bodily and spatial involvement. The coupling is recursive, and the connecting surface is implicit and rests on unconscious psycho-physiological reactions. That is a claim about coupling rather than representation, and it is actionable only if the sense is fixed.

Recent work turns the body from object into design material

The most developed recent extension turns the body from an object of analysis into a design material. Hook (2018) proposes soma design, a shift from a symbolic, language-oriented stance to an experiential, felt and aesthetic one across the whole design and use cycle. That shift requires a slower process than commercial design permits. The claims for better products are the author's expectations, not results. Guspara (2024) attempts an applied test. The study uses somaesthetic design and phenomenology to examine the mobility experiences of people with paraplegia, and treats first-person bodily experience as evidence for assistive devices. Its scope is limited, and it is conceptual rather than comparative. It shows the approach can generate design knowledge, but not that it outperforms the alternatives for older users.

Contested question: does embodiment predict, or only re-describe?

The first position holds that embodiment is a testable design claim. Hurtienne and Israel (2007) treat intuitivity as definable, and image schemas as a taxonomy that classifies interface properties in advance. Antle and colleagues (2009) convert the claim into a comparison, by varying only the metaphor in the mapping. Kaipainen and colleagues (2011) propose enactment as a system-level principle that could be implemented. On this reading the framework predicts, so a design can be wrong about it. What would settle it is a study that names the embodied property, holds task and interface otherwise fixed, and states the mechanism in advance. It would report the outcome the property was supposed to produce.

The second position holds that the framework is analytic rather than predictive. Dourish (2001) organises understanding rather than forecasting it. Winograd and Flores (1986) ask designers to understand readiness-to-hand and to anticipate breakdowns, not to compute an outcome. Svanaes (2000) offers a phenomenology and a design sensibility. On this reading the framework earns its place by making a setting intelligible, and the demand for a prediction misreads it. What would settle it is a demonstration that such a framework changed a design decision a predictive alternative could not have reached.

The third position holds that the argument is at the wrong level until the sense is fixed. Wilson's (2002) six claims are not equally supported. The one that carries this unit's thesis, that the environment is part of the cognitive system, is the one she calls deeply problematic. Clark (1997) presses the point from inside the cognitive science the programme borrows from. His minimal representationalism keeps representations but makes them local, personal and action-oriented, between computationalism and anti-representationalism. The leading statement of embodied cognition keeps a place for representation. The claim that the internal model is simply the wrong unit of analysis is therefore stronger than the evidence supports. A comparison that separated the six claims, fixed one for a task, and reported whether the design differed would separate the three positions.

The evidence on the two sides differs in kind. The predictive side offers one controlled comparison, and it reported qualitative findings rather than a benefit from the metaphor. Hurtienne and Israel (2007) supply a taxonomy and a worked case, and Kaipainen and colleagues (2011) supply concepts. The analytic side is the body of design-centred conceptual work, with Svanaes's (2000) three exploratory experiments. While the dispute is open, a designer should write down which sense of embodiment the project uses. A claimed benefit from a metaphor stays a hypothesis until a version without it has been compared. The strongest objection is that the corrective may not be available. Once the sense is narrow enough to test, as Fishkin (2004) and Antle and colleagues (2009) each make it, it no longer carries the framework's claim about mind.

Where the argument is heading

Hook (2018) makes felt experience the material, and asks designers to work slowly and bodily. Guspara (2024) carries that stance into a domain where a design mistake costs a person's mobility. Kaipainen and colleagues (2011) propose that the connecting surface between person and system is partly unconscious. That would make the coupling, rather than the interface, the object of measurement. Each proposal becomes a testable design claim only when the sense of embodiment is fixed, as Fishkin (2004) does for tangibility and Hurtienne and Israel (2007) do for intuitivity. Wilson (2002) and Clark (1997) remain the standing check. The programme earns its standing when it can say which claim it makes, and name the result that would show it wrong.

The sources this unit rests on

• Antle, A. N., Corness, G. and Droumeva, M. (2009). Human-computer-intuition? International Journal of Arts and Technology 2(3). doi:10.1504/ijart.2009.028927 Ran the same auditory environment with and without an embodied metaphor mapping movement to sound. The findings are qualitative, so the claimed benefit stayed unproven.
• Clark, A. (1997). Being There: Putting Brain, Body, and World Together Again. MIT Press. doi:10.7551/mitpress/1552.001.0001 Argues that cognition is scaffolded and extended by body, tools and environment. Its minimal representationalism keeps representations but makes them local and action-oriented, between computationalism and anti-representationalism.
• Dourish, P. (2001). Where the Action Is: The Foundations of Embodied Interaction. The MIT Press. doi:10.7551/mitpress/7221.001.0001 Grounded in Heidegger, Schutz and Wittgenstein, it defines embodiment as the unfolding of physical and social phenomena in real time and space and organises embodied interaction around meaning and practice. Tangible and social computing are argued to share that foundation.
• Dreyfus, H. L. (1992). What Computers Still Can't Do: A Critique of Artificial Reason. MIT Press. https://archive.org/details/whatcomputerssti0000drey Argues that the representational programme in artificial intelligence mistakes a derivative mode of thought for the whole of it, since skilled coping draws on an inarticulate background no symbolic representation captures. Success is confined to special-purpose tasks.
• Fishkin, K. P. (2004). A taxonomy for and analysis of tangible interfaces. Personal and Ubiquitous Computing 8(5). doi:10.1007/s00779-004-0297-4 Presents a taxonomy using metaphor and embodiment as axes, so tangibility is a spectrum rather than a binary property. Its narrow sense of embodiment must not be confused with Dourish's, which would incorporate all tangible interfaces as a small subset.
• Guspara, W. A. (2024). Exploring Somaesthetic Design and Phenomenology in Understanding Paraplegic Mobility. International Journal of Creative Multimedia 5(2). doi:10.33093/ijcm.2024.5.2.4 Uses somaesthetic design and phenomenology to examine the mobility experiences of people with paraplegia, treating first-person bodily experience as design evidence. It is conceptual and limited in scope, showing the approach can generate design knowledge but not that it performs better.
• Hook, K. (2018). Designing with the Body: Somaesthetic Interaction Design. The MIT Press. doi:10.7551/mitpress/11481.001.0001 Proposes soma design as a shift from a symbolic, language-oriented stance to an experiential and aesthetic one, requiring a slow process. It argues for better products rather than demonstrating them.
• Hornecker, E. and Buur, J. (2006). Getting a grip on tangible interaction: a framework on physical space and social interaction. Proceedings of the SIGCHI Conference on Human Factors in Computing Systems. doi:10.1145/1124772.1124838 Defines tangible interaction as relying on embodied interaction, tangible manipulation, physical representation of data and embeddedness in real space. It interweaves the material and the social, applying the framework to three case studies.
• Hurtienne, J. and Israel, J. H. (2007). Image schemas and their metaphorical extensions: intuitive patterns for tangible interaction. Proceedings of the 1st International Conference on Tangible and Embedded Interaction. doi:10.1145/1226969.1226996 Defines intuitivity, then proposes image schemas and their metaphorical extensions as a taxonomy for tangible interaction, illustrated with a worked case. It is the clearest attempt to make the intuition claim testable.
• Kaipainen, M., Ravaja, N., Tikka, P., et al. (2011). Enactive Systems and Enactive Media: Embodied Human-Machine Coupling beyond Interfaces. Leonardo 44(5). doi:10.1162/leon_a_00244 Proposes the enactive system as an alternative to the standard human-computer system: two elements coupled holistically by bodily and spatial involvement, with the coupling recursive. It develops concepts rather than reporting a study.
• Klemmer, S. R., Hartmann, B. and Takayama, L. (2006). How bodies matter: five themes for interaction design. Proceedings of the 6th Conference on Designing Interactive Systems. doi:10.1145/1142405.1142429 Synthesises five themes from psychology, sociology and philosophy: thinking through doing, performance, visibility, risk, and thick practice. It aims to inspire evaluations integrating the physical and computational worlds.
• Svanaes, D. (2000). Understanding Interactivity: Steps to a Phenomenology of Human-Computer Interaction. PhD thesis, Norwegian University of Science and Technology. https://www.idi.ntnu.no/~dags/interactivity.pdf Works out a phenomenology of interactivity and carries it into three experiments in an interactive system. It proposes kinesthetic thinking and the interaction gestalt, and reports that the design tools of the time did not support kinesthetic thinking well.
• Wilson, M. (2002). Six views of embodied cognition. Psychonomic Bulletin and Review 9(4). doi:10.3758/bf03196322 Distinguishes six claims made under the heading of embodied cognition: it finds the first three and the fifth at least partially true, calls the claim that the environment is part of the cognitive system deeply problematic, and calls body-based off-line cognition possibly the most powerful.
• Winograd, T. and Flores, F. (1986). Understanding Computers and Cognition: A New Foundation for Design. Ablex. https://archive.org/details/understandingcom0000wino Argues that the rationalist tradition behind artificial intelligence and design misrepresents understanding, drawing on Heidegger, Maturana and Austin's speech-act theory to hold that understanding is embodied rather than symbol manipulation. It concludes that design lies in readiness-to-hand and anticipating breakdowns.