This is an excerpt from my Master’s Thesis: The Relevance of Emergence in Human-centered Design.
To download the entire Thesis, please go to: My Master’s Thesis

The Underlying Framework
In order to better understand systems and emergent phenomena, researching in the real world is tremendously helpful. The initial studies conducted were short and simple. There were four preliminary explorations (PE). The first exploration (PE1) identified a system of rules in open accessible places. The second exploration (PE2) looked at a simple game and then modified the rules to study the emergent behavioral change. The third exploration (PE3) attempted build a digital game, in which a number of parameters was established, but the rules of interactions were left open.

…

An underlying framework was derived from studying those particular examples. The overall participatory design approach in this thesis was divided into three core levels (Figure 4.1). Level 1 is the fundamental structure/scaffolding for people to work with. In design language, this could be compared to the grid system, in which designers could place and organize different elements on the page. Depending on the differences of the grid systems, the organization of the design elements could vary. Level 2 is the interactions between of the components or agents. Level 2 depends on the rules that give guidance to the components or agents on how to interact. Level 3 is realized when a distinct and new high-level property, such as behavior or function, emerges from the self-organizing aggregation and interaction of the lower-level components.

 

 
Figure 4.1. A review of the overall breakdown of the levels consisted of: Level 1 = Basic structure/scaffolding, Level 2 = The interactions of the components following certain rules, Level 3 = A novel system emerges. Thus, the underlying framework can be simply expressed as: Level 1 → Level 2 → Level 3.

Exploration 1 (PE1): Rules & Behaviors in Public Spaces
The tasks of this initial exploration consisted of mapping the location of people and noise concentration (whenever possible) and measuring overall levels of noise over time. The purpose of this study, however, was not to look at noise level as the emergent phenomena. Instead, it was an attempt to sketch out what possible rules and behaviors may influence noise in public spaces. The observational research was conducted at both Heirloom Cafe at Wexner Center and KSA Cafe at Knowlton School of Architecture at The Ohio State University.

First, an environmental audit was done by looking at the design elements and organization of the space and creating outlines of the floorplans for both cafes. The floorplans served as an overhead diagram to map where people were during the period of the observations (5-minute intervals for an hour and a half). Second, while conducting the observational research, audio of the environment was recorded. The audio was later imported to an audio editor — Wavepad. The audio editor created the visualization of the sound. By looking at the sound waves and comparing them with the plotting of the floorplans, areas that seemed interesting could be isolated. Quantitative data were analyzed and a set of basic rules and behaviors was constructed to help make sense of the patterns of the noise pollution.

The observation started at 11.00 AM and ended at 12.30 PM at Wexner’s Heirloom Cafe (Figure 4.2). From the observer’s perspective, the areas of noise concentrations were on W1, W2 and W3. W1 is where the location of the register. It was dynamic because many people came and went as seen from the map, and the area around the register was busy because the servers were interacting and communicating with each other. W2 was an area where the people chatted quite loudly, because it was somewhat enclosed. Therefore, it was a huge contributing factor for noise concentration. W3 was closest to the observer. At one point, W4, where people went to see the gallery at Wexner, got quite rowdy and may have contributed to the noise, however, it only occurred in a small window of time.

 

 
Figure 4.2. A floorplan of Wexner Heirloom Cafe with observational notes that span for over a period of an hour and a half. The different values and the overlaps of the red dots are the results of overlaying all the 5-minutes interval recording on top of each other.

The observation at KSA Cafe also started at 11.00 AM and ended at 12.30 PM (Figure 4.3). From the observer’s point of view, K1, K3, K4, and K5 were the areas of noise concentration. While in K1, there were people who stayed and talked, K3 & K4 were areas where people queued and were more dynamic because some people chatted and some people just waited for their orders. K5, the register and kitchen, was noisy depending on the amount of people being served. K2 at one point also got quite loud, but only at certain window of time. K6 at one point got quite a few people sitting in groups, however, it was too far away from the observer’s seat to be able to discern the noise made.

 

 
Figure 4.3. A floorplan of KSA Cafe with observational notes that span for over a period of an hour and a half. The different values and the overlaps of the red dots are the results of overlaying all the 5-minutes interval recording on top of each other.

From the observations, in general, there were two rudimentary, low-level behaviors (represented with LL-B). The low-level behaviors are active and passive (Figure 4.4):
LL-B1. People were interacting with each other in a group — therefore, Active (A).
LL-B2. Individuals were not interacting with other people — therefore, Passive (P).

 

 
Figure 4.4. Two types of low-level behaviors: Active and Passive.

From the interaction of those low-level behaviors, complex high-level behaviors (represented with HL-B) emerged. For example, in HL-B1 (Figure 4.5), space and proximity influenced the magnification or containment of effects, but ONLY if the sphere of influence of the behavior was the same on those groups. That means, both active groups would magnify and both passive groups would contain.

 

 
Figure 4.5. High-level behavior Type 1.

A different high-level behavior also formed (HL-B2), when the sphere of influence of the behavior of one group was significantly stronger than the others, then space and proximity became secondary as illustrated in Figure 4.6. In that case, the strong influencer affected the group further away than the one closer with weaker influence.

 

 
Figure 4.6. High-level behavior Type 2.

In Figure 4.7, when two or more groups that were active met, the effect was magnified (HL-B3.1). However, there was a certain level of threshold, in that they did not magnify infinitely — only up to the point where both groups reached the level where they did not need (or were unable) to magnify their effect any longer to sustain their own activities. Finally, when an active group met a passive one (HL-B3.2.), the effect was contained, because passive behavior only contains, and the active group did not need to amplify further than necessary.

 

 
Figure 4.7. High-level behavior Type 3.

PE1: Discussion of Exploration 1
Even in a small scale system, like a cafeteria, emergent behaviors could arise. Human’s sphere of influence not only relies on the quantity of the interacting agents, but also on the quality of the interactions. An individual with a strong presence (like loudness) could have different kinds of influence to other groups. Interactions and encounters of the crowd raised or lowered the noise concentration. And through those interactions, complex behaviors emerged.

Figure 4.8 below shows that the rules and behaviors in the system of noise concentration at the cafes conform with the underlying three-level framework provided earlier in this chapter. The interactions of the LL behaviors based in the structure of the environment gave rise to the HL complex behaviors. And based on the dynamic environmental change, the LL interaction changes also influenced the changes on HL behaviors.

 

Figure 4.8. Human-centered Design and Emergence framework for PE 1.

The observational study was quite holistic and encapsulated quite a wide range of different parameters and data in order to study emergent systems in cafeterias. Breaking the duration of the observation into 5-minute increments helped with simplifying data, thus making recording the observation more manageable. While the method was quite holistic, it was limited to the point of view of the observer. The areas further away from the observer’s seat were difficult to observe. The further away the area from the observer’s point of view, the less accurate the observation would be.