Systems Theory & HCI in the Design of New Cycling Technologies
Integrating Mind, Body & Communication
1. Introduction
It is widely acknowledged that the historical development of human-computer interaction (hereafter, HCI) can be modeled according to a notion of three major ‘waves’ of evolving research considerations. These waves represent a gradual widening of HCI’s domain, expanding from human-technology coupling and the modeling of human factors, to cognitive psychology approaches toward human information processing, to meaning making viewed through socio-cultural lenses.
This essay outlines a conceptual framework for ‘integrated HCI’ that can rigorously structure HCI making and inquiry across this varied discursive and methodological terrain. The framework demonstrates that, at least with regards to some kinds of human-technology interactions, a perspective that embraces a ‘multi-wave’ approach can be productively brought to bear when the interactions require complex physical and embodied coordination with informatic devices. The framework is based on Niklas Luhmann’s systems theory, in which he distinguished Mind, Body and Communication as distinct autopoietic systems, taking each other as their environment.
In a recent survey of technologies and methodologies associated with third wave HCI (Filimowicz and Tzankova, 2018a/b), I proposed the possibility of using Niklas Luhmann’s systems theory as a basis for integrating HCI’s divergent methodological positions. Luhmann’s theory is triadic in its design, seeking to overcome the negative epistemic effects and impasses that resulted from Enlightenment-era Mind-Body dualisms which continue to affect inquiry today.
Incorporating Maturana and Varela’s concepts from biological phenomenology, he introduced a third term into the mix, Communication, on the empirical basis that the human realm could be modeled as Biological, Psychological and Social autopoietic systems that take each other as their environment. I argued that this triadic formulation was a good fit for understanding the expanding forms of research and production that HCI has historically undertaken.
I gave the example of a prototype interactive technology for horse riders to show how Luhmann’s Mind/ Body/ Communication concepts can be productive for new integrative approach to new technology designs:
The successful operation of such systems involves considerations at several levels that closely correspond to the three waves associated with HCI. First, the system requires to be physically constructed which engages technical and ergonomic concerns — such as physical design of the equipment, posture of rider, and kinesiological characteristics of the horse. This level corresponds to problem conceptualization characteristic of first wave HCI.
Second, the design of the system should take into account cognitivist considerations — e.g. not distracting the rider through misallocation of limited attentional resources — problematics essential to second wave HCI.
Last, the system should effectively communicate to the rider by providing feedback that makes sense — facilitating interspecies communication between technology, horse, and rider through embodied interactions. This level of ‘meaning making’ is a distinct theme of third wave HCI.
A system such as this — especially coming from a sports context where all three levels are vital to the safety and security of the sportsperson engaged — exemplifies the growing necessity of a research agenda that integrates all three HCI waves through discursive and practical variations based upon Luhmann’s three autopoietic systems categories of Body, Mind, and Communication. (2018a/b, p.6)
In this essay, I will expand further on these ideas via a formalized conceptual framework that will show more clearly how Luhmann’s ideas can be expanded into a practical design framework. I will argue for its robustness by applying the framework to preliminary design considerations for new computational media designs in the area of interactive sports technologies for cycling.
2. Interactions Between Autopoietic Systems & Their Environment
In autopoietic theory, systems exhibit operational closure which allow them to be distinguished from their environments, and indeed distinguish the environment in the first place. In contrast to certain intellectual movements currently in vogue based on forms of radical symmetry, such as object-oriented ontology (OOO) and actor network theory (ANT), Luhmann’s systems theory is based on the asymmetry in which all distinctions occur within the observing system.
The environment itself makes no distinctions, and all observations are one-sidedly within the system, starting most importantly with the distinctions between self-reference and other-reference. The autopoietic system’s first distinction is between itself and its environment, which is a process that occurs within its operational closure from the environment. All subsequent distinctions are likewise made only internal to the observing system.
Luhmann takes up the key figure in George Spencer Brown’s calculus of form and distinction as the conceptual cipher for indicting this asymmetry at the origins of all production of distinctions, where the mark — which is an indication of distinction or observation itself– exhibits the form of autopoietic operational closure:
This asymmetrical perspective within Luhmann’s thought also aligns well with the widespread interested within HCI of phenomenological approaches. Don Ihde, in his essay “You Can’t Have It Both Ways: Situated or Symmetrical,” (Selinger 2006), pointed out the conceptual difficulties of trying to claim both a situated perspective or orientation in one’s research, while also trying to take up the theoretical resources of ANT and OOO which are based on radical symmetries.
ANT and OOO take up somewhat ‘omniscient’ (in the narratorial sense of a detached third person observer) perspectives outside the systems/environment distinction by producing forms of analysis in which both sides of the sociotechnical equation– human and non-human– are ‘equal’ in their discursive and productive effects. An autopoietic perspective on all forms of cognition, observation, and distinction, however, makes radical symmetry impossible, since all the conceptual labor of a network analysis, for example, can only be accomplished within the operational closure of the observing system (i.e., an ANT theorist). Nonliving objects are part of the environment, and thus make no distinctions themselves.
In systems theory, since autopoietic forms produce themselves, and produce their own constructions within their internal operations, they cannot steer or control the operations in other autopoietic systems. Thus, the interactions between systems, or between systems and their environment, are limited to certain kinds of effects and processes, since they take each other as their local ecology. While there are many specific interactions that can be identified between particular systems, staying at a high conceptual level within Luhmann’s general systems theory, there are three main categories of interactions between systems and the environment as follows:
- Perturbance: systems can impinged upon, disturb or perturb, other systems or the environment, irritating them to higher or lesser degrees.
- Observation: systems make distinctions, producing its forms of cognition within operations of self/other reference.
- Structural Coupling: the operational closure of systems does not foreclose functional connections between the system and other systems or the environment.
Below are short illustrations to ground these concepts.
Perturbence
In Luhmann’s social theory, the Political and the Economic are each different autopoietic function systems within the ecology of society. No matter how much politicians may try to take credit (or blame others) for the performance of the economy, politicians are not in a position to steer the economy, since such steering in fact occurs by actors within the system of the economy.
Politicians can try to produce certain economic effects, such as by increasing/decreasing taxes on the wealthy, or expanding/contracting the monetary supply, but these actions merely impinge upon or perturb economic operations, rather than steer them. If this were not the case, then it would be a rather simple manner for all politicians to simply grab ahold of the policy steering wheel and direct all economies to produce prosperity for all citizens. That this never happens is a result of the functional differentiation of society into distinct autopoietic systems of economy and politics.
Observations
All distinctions, observations, and meaning production occurs within the operational closure of autopoietic systems. The ‘primordial’ distinction made is between self and other reference, since the system needs to close itself off operationally from its environment in order to be a system in the first place. The system produces its forms of distinction based on its difference from the environment, and this constitutes its basic form of cognition.
But systems are limited as to what they can distinguish. The law can distinguish between what is legal or not legal, but not what is red or what is blue, or what is conservative and what is progressive, or what is beautiful and what is ugly. Visual processes, political discourses, art worlds and so on produce forms of distinction particular to those kinds of systems. Distinctions are made on the terms of the system’s internal operations.
Structural Coupling
No system exists in a vacuum. That the respiratory and circulatory are different systems does not mean that healthily circulating blood can do without air. While minds have their own grounding in unique biological brains, they structurally couple to communication systems to produce meaning between them.
Communication itself is a distinct autopoietic system from minds and bodies– otherwise we would need telepathy to communicate. Clothing may be a form of structural coupling to the body, but it is not a biological process within our cells or even between our bones and muscles. Many objects with which we can even become physically intimate with do not enter into the operational closure of our bodies but rather form a layer of intimacy between us and the environment.
3. The Framework
To form the first iteration of the conceptual framework for what I call Integrated HCI– approaches to HCI research that are inclusive in its perspective across the disciplinary variations that have been identified under the heading of the three ‘waves’ above– I begin by developing axes based on each wave and by each form of system interaction.
For purposes of concision, an abstraction layer is introduced which associates the three HCI discourses with Luhmann’s terms Body (first wave, e.g. ergonomics, human factors, use cases), Mind (second wave, e.g. perceptual organization, information processing, memory, attention) and Communication (third wave, e.g. meaning making, situated context, socio-cultural dimension).
This gives us a matrix of 9 initial conceptual interactions for Integrated HCI:
The use of this framework is straightforward: at the intersecting cells of the rows and columns, design specifications, methodologies and prototype inspirations appropriate for each intersection are explored for whatever technology is under development.
This systems theoretic approach offers the HCI researcher and prototyper a highly flexible and robust framework for structuring rigorous interdisciplinary R&D that takes into account the broadest possible considerations proposed to date for integrating all the methods and approaches in designing human-computational interactions and artifacts. Such an approach can be highly productive when the designs are of sufficient complexity to call for research that transgress beyond the more narrowly defined methods that are associated with particular cultures of HCI.
I consider the development of this framework as a speculative project and not intended polemically to challenge or replace other frameworks, nor is it aiming to be a ‘universal’ framework that would somehow be located in a grand narrative space outside of contingent developments in the history of ideas. Autopoietic theory is a powerful lens for understanding sociotechnical systems and relations, but it is also just one lens amongst many.
I have produced this framework in an exploratory manner to test its conceptual resources against the realities of prototyping robust interactive systems in challenging use contexts. Like many if not most conceptual frameworks in HCI, its validity is largely of a pragmatic character, and should eventually show its efficacy in the development of new artifacts.
4. Applying the Framework
In this section, the framework is applied to the cycling context. The matrix’s nine intersecting conceptual positions are identified in the headings below. Each heading is explored through a looser and more speculative style (hence the italics) as a way of introducing a wide creative latitude to the design possibility space and to foster lateral thinking.
4.1 Coupling/Body
The coupling of bike to rider already has well-established practices, with the main methods being either the custom fitting, or the bespoke bicycle built to the body type and dimensions of the rider. Bike fittings typically entail a set of adjustments to the main physical touchpoints between bike and rider–namely handlebars, saddle and pedals–and often require replacement of parts when adjustments prove insufficient.
In the field of mechatronics there is the notion of ‘smart steel’ and ‘dumb steel.’ Dumb steel has no informatic component, an example of which would be the tracks of a roller coaster. Smart steel comprises the sophisticated environment-reactive networked sensor and actuator arrays of the often media rich ‘smart’ cars that ride the dumb steel, and which pivot constantly to point riders in the direction of a theme park’s moment-by-moment attractions.
Mechatronically smart steel on a bike could play a role in relieving body stresses that build up over time, for instance by providing novel ways of stretching a rider’s arms, shoulder and torso that go beyond simply moving the hands down to the drop bars, which is a traditional way of changing upper body stresses. A smart steel bike could elongate and contract while in motion to alter the rider’s position over time, and relieve stresses that may build up in the body.
4.2 Observation/Body
Consumer electronic devices for observation of the cyclist’s body tend toward cardiovascular monitoring, sometimes featuring wireless connectivity to a bike computer which is capturing ride data via GPS and a map-based navigation database. Devices have yet to be designed to better monitor other aspects of a rider’s physiological changes over time.
Rider’ rely on much internally ‘organic’ information processing, such as the feeling of hunger or lightheadedness which correlates to caloric expenditure, or the sensation of ‘hot foot’ (Metatarsalgia) from prolonged squeezing of nerves and joint tissue in the ball of the foot. Felt stresses in the muscles after hours of riding are also not part of a formal data set beyond what might be stored in the records of a massage therapist. A goal of better body-observing equipment would be to identify areas where improvement is either needed or has occurred in the effects of cycling on the body.
4.3 Perturbance/Body
There are many body-perturbance factors in cycling, such as climbing hills, rain and of course, accidents and injury. Bikes with electric motors already relieve the perturbances from hill climbing, and waterproof clothing mitigates the effects of bad weather. Designs to protect from the impacts of accidents are mainly focused on the rider’s head, though in some cycling use cases, such as with mountain biking, injury protection is also paid to the knees, elbows and spine.
Road cyclists don’t have good options for expanded body protection since the body armor that might work in a downhill off-road context is unsuitable for longer distance, racing or touring styles of cycling. There are more recent advances in wearable technologies that can convert a shawl into a helmet in a very short order of time based on sensor readings that indicate a likely head impact in imminent.
Key advances would be to equip road cyclists with better and adaptive body protection without sacrificing overall rideability. Another area yet to be fully solved is in keeping the hands warm and dry in cold wet weather, since even the most waterproof cycling gloves eventually fail given the high PSIs (pounds-per-square-inch) that are present in the glove seams.
There may be a role of embedded technologies in making riding gloves more active, e.g. by warming them and pumping mositure out, since strictly analog (materials-based) approaches have not proven satisfactory (much marketing and many an online product description notwithstanding!).
4.4 Coupling/Mind
Brain computer interfaces present perhaps the most ideal realization of this matrix position. The design possibilities of mapping thought directly to aspects of the rider’s experience are rich in sci-fi-esque possibilities! On a more prosaic level, a good example of ergonomic coupling to cognition would be the use of bone conduction headphones for obtaining sonic information while cycling, which could range from aesthetic information (i.e. music for co-organizing one’s internal rhythms or cadence) to sonified information directed to the rider’s awareness, or even communication systems between riders.
Since bone conducting headphones don’t cover the ears, coupling to cognition in this manner doesn’t obstruct the sounds of the natural environment which provide important cues, such as the approach of oncoming vehicles! The best placement of a bike computer for ease of view while riding, or the use of a heads-up display to obtain visual information of the space behind the rider in real-time, would be other sites for mind/coupling design.
4.5 Observation/Mind
Information presented to a rider in situ needs to be formatted to provide rich, though not constant, information flow, which could risk becoming distracting, obscure or noise-like in character. Sonifications with coded cues to provide warning signals for passing vehicles that are possibly too-close to the rider, or real-time navigation that is readable in bright daylight or at night, allow the mind to observe its environment while at the same time, such systems observe the mind’s attentional constraints in this highly active context.
Also desirable would be a system of warning sounds (conveyed through cheek-resting bone conduction headphones, as per above) that could distinguish a passing cyclist from a vehicle, since one’s ‘fellow’ cyclists can also be a source of annoyance and even accidents in the limited space of road shoulders or at intersections.
4.6 Perturbance/Mind
Warning signals are exemplary in the way they intentionally disrupt attention, instantly shifting awareness to a new significant reality that takes over focal awareness. At speed, the sound of wind over the ears becomes a significant roar and it can be difficult to perceive traffic sounds until very close to the moment of passing, which sonification systems could cut through by providing important cues sonically about events in the local environment.
Bike computers typically chirp their warnings when the battery is running low, or a turn in the route is coming up. Another source of mind-perturbance is the consumption of alcohol and recreational drugs, which informatic systems can be made aware of through sensor technologies attuned to inebriated cycling. One can even envision a bike that locks the crankset and refuses motion to a rider under the influence (which is an everyday expression for a mind perturbed!).
4.7 Coupling/Communication
This matrix junction forces one to consider the optimal locations and types of communication channels that can be established in the cycling context. Walkie talkies would be a very poor form factor indeed, since the hands are needed on the handlebars (assuming non-daredevil riders or those not on unicycles).
Communication can be between riders, toward oneself (e.g. a system for taking audio memos along the route), toward other vehicles (“Hey you, you’re driving too fast and too close! I have a camera recording everything and live-streaming to the cloud, so you better not hit me because I’ve got evidence that can convict you!), to anyone on a hands-free mobile call, or even to an AI-infused cycle frame where a conversation may need to be had between a rider and his bike (“Hey, Bike AI, I usually like to inflate my tires to their maximum PSI but it’s a hot day, so please deflate the tires a tad so they don’t warp or burst”).
4.8 Observation/Communication
Communications can be surveilled, though in this context through a desired self-surveillance. Information that is more of a kind that can be described as ‘perception-to-mind’ is perhaps better understood under the Mind/Observation heading above.
In Luhmann’s theory, communication bridges operationally closed minds, and so signals formatted for attention, focus and memory may satisfy the Mind/Observation matrix position, while the observation of communication develops more of an ‘in-between’ space where meaning unfolds between two nodes of interacting systems (e.g. human-information interaction).
A good site for this design space might be the web-based data dashboards, visualizations and tables one can peruse after syncing the bike computer to one’s online cycling profile at home. These online records of one’s rides over time create a scene for navigating data-rich meaning spaces with the ultimate goal of extracting insights from the data set.
4.9 Perturbance/Communication
In a cycling context, processes of noise reduction are needed to separated desired signal from noise. There may also be competing communication needs which requires one communication channel to be overridden by another. We are probably not far from the day when the AI in the bike frame (or in our helmets?) will rudely interrupt our cellphone conversation in order to point out more situationally important information, such as locating (with its embedded telescopic computer vision sensors) a distant Johnny on the Spot at a construction site which can provide a rider with a much-needed bio break.
4.10 Discussion of the Design Approach
It is clear from the above design space exploration that the integrated HCI model is highly flexible and adaptable in its prototyping ramifications. The framework covers all of the considerations and technological enhancements that one would seek from a comprehensive design approach based on high-level systems theoretic concepts. As such, it can be applied to many contexts since its conceptual foundations operate at both conceptual generality and connect profoundly with straightforward practical applications.
A designer of integrated HCI components as envisioned here does not need to become an expert in Luhmann’s systems theory, or the subtheories that it integrates, such as Maturana and Varela’s biological phenomenology, or Brown’s calculus of form, or Parson’s functional differentiation of social systems. The notion of Mind, Body and Communication as distinctly addressable sets of inputs and outputs cross-modulating each other in highly coupled, interactive and technologically enhanced situations and environments is a model that can be taken up directly for prototyping new embodiments of functions, displays, sensors and actuators in sporting contexts and by extension, any suitable context.
Another interesting ‘finding’ (since a conceptual framework imaginatively applied can be considered a thought experiment) is that the matrix challenges a designer to think well beyond initial ideational constraints. In other words, the matrix asks for an articulation of the whole experience of cycling and as such, works in the direction of imagining a suite of system designs, which frees the inventor from early fixation on single prototype concepts.
In approaching cycling as the use case, I originally intended to use the framework to conceptualize a more computationally media savvy riding helmet. However, the matrix pushed me well outside the scope of this one particular imagined object and, when taken up as its own creative spur, opened up a more comprehensive attention to the whole scene and situation of cycling which in turn generated a more artifact-rich design space.
5. Conclusion
It should be acknowledged, of course, that not all, and indeed probably not most, interactive technologies would call for such an integrated HCI approach. Interaction design for mobile apps, for instance, might do well by only taking into account cultural interactions between a select community of users along with currently popular user interface aesthetics. Or, a new vehicle interface might need only to take into account certain dimensions of visual cognition and a user persona, and so on.
However, with computation increasingly embedded into many objects and environments, I suspect that new methodological ‘returns’ to the knowledge base and skills et of first wave HCI may well become increasingly important, while also needing the full conceptual resources of the latest HCI research to produce robust designs informed by contemporary movements and ideas.
References
Brown, G. S. (1969). Laws of Form. London: Allen and Unwin.
Filimowicz, M. and Tzankova, V. (Eds.) (2018a). New Directions in 3rd Wave Human-Computer Interaction Volume 1: Technologies. Springer HCI Series.
Filimowicz, M. and Tzankova, V. (Eds.) (2018b). New Directions in 3rd Wave Human-Computer Interaction Volume 2: Methodologies. Springer HCI Series.
Selinger, E. (2006). Postphenomenology: A Critical Companion to Ihde. SUNY series in the Philosophy of the Social Sciences. Albany, NY: SUNY Press.
