Between idea generation and physical embodiment: Exploring the ‘fuzzy middle’ in conceptual design Tomas Hellström 1. Introduction (1 page) The part of the design process that runs from initial ideation or design brief to physical embodiment, has been referred to as the ‘conceptual phase’ (Tovey, 1997). This phase, which precedes the engineering and production stages, typically involves the generation of new ideas, both of a technical and a non-technical nature. Rationally, the conceptual design phase has been said to be represented by a task clarification phase involving problem and market analysis, a conceptual design phase which includes generating and assessing alternative solutions in order to arrive at a design concept, and finally an embodiment phase, where more coherent tests can be made and manufacturing issues can be assessed (Lofthouse, 2004; Roy, 1996). However, while much is known about the initial processes of ideation, as well as about the externalizing processes involved in team work, less is known empirically about the informal and often highly personal process that occurs between the very early stages of design, just when a design brief has been put forward or a general concept has been formulated, and the time when this idea is becoming embodied, for example in a mock-up or some other type of representation. It often happens that the first stage of ideation occur in some social setting, where brainstorming or concept discussions are taking place. The designers then ‘incubate’ these images in different ways, refine them and when ready presents a new version to the group. This may happen slowly or fast, and express itself differently in different contexts. This ‘in between’ phase will be referred to as ‘the fuzzy middle’ of conceptual design and the focus of the present paper is on how this process functions and how it may be structured. The idea of the ‘fuzzy middle’ of conceptual design is that while initial ideation or design brief as well as embodiment/presentation offer social and formal anchoring points for the designer, the intermediate period between these two is more personal and subjective; less amenable to steering and management, and more influenced by the 1 personal characteristics and tacit knowledge of the designer. Consequently, the empirical study on which this paper is built has aimed to capture the specific, qualitative experiences of a number of designers who operated in one project, and their process of passing through the ‘fuzzy middle’, both at a personal specific and a general structural level. To this end a phenomenological deep-interview technique was employed. This type of study typically focuses on a small set of subjects in order to facilitate a rich qualitative account of personal processes. The results suggests a three-stage model of the […] 2. Theory (2 pages) While the conceptual phase may be very socially dynamic and interactive, especially in the early phases of ideation where for example brainstorming sessions are often used to expand concepts, this early stage soon gives way to an incubation period where the individual designer melts the impressions and starts a kind of mental sorting process, only to return to the group with a new concept. This is in many ways a journey from an initial elusive concept, to an externalized conceptual object of some kind, or a first embodiment to be put up for social evaluation. At this last stage of conceptual design, sketches, mockups, models etc. together with preliminary evaluations of aesthetic and use value are addressed in a social setting, before the next step of the product development process begins (Roy, 1996). This ‘in between’ period of mainly individual incubation is here referred to as the ‘fuzzy’ middle of conceptual design, since unlike design brief, idea generation (ideation) and presentation of an embodied concept, which are fairly well anchored in shared practices, this phase is often individual and private. However, it is easy to see how this stage, where the individual designers go off to think and tinker with the concept, is crucial to the way that a tentative set of concepts end up becoming embodied design alternatives. What is going on in this phase of design? We will look at this question by drawing on some of the received views on design thinking. It has been suggested that what signifies the role of the designer in this general phase, usually the domain of the industrial designer rather than the design engineer, is a concern for the expressive aspects of the future product, that is what it says about its user and what values it conveys ‘about itself’ (Svengren, 1997). This type of people-centered or user-centered mentality has been said 2 to exist among industrial designers involved in this phase, as opposed to the technologyorientation typical of later stages of the product development process (Ulrich and Eppinger, 1995). There are however good reasons to believe that these designers too operate with a level of technical involvement in generating and assessing candidate concepts (Lofthouse, 2004). The mix of divergent and convergent thinking discussed by Lawson among others exists also in this phase, as fairly unstructured and expansive idea generation based on aesthetic inspiration, new functionality concepts and imagination of possible users mixed with a technical, logical and evaluative mode of though (Lawson, 1990). Dietrich Dörner (1999) has suggested that this process may proceed from an initial concretization of a design brief on the basis of available analogies and previous solutions from other fields, which produces a vague idea of what the final product may look like. It then enters into sketching or modeling phase which further clarifies the characteristics of the product. This phase may produce a number of dead ends but also a ‘stockpile of ideas’ for future use. A parallel process is that of the ‘picture-world cycle’, where images and words are combined, to the effect that the designer put ideas into words for him/herself as well as for others. According to Dörner, rephrasing might open up new avenues for though and specify a concept, but it may also prematurely destroy the dynamics of thinking about the design. Yet another step is the process of ‘mental simulation’, where the designer tries to elaborate on the design by looking for contradictions and possibilities by mentally running the machine/product and asking questions such as ‘what will happen if…?’ and what side effects will result from this change…?’. All these aspects of conceptual design really combine divergent and convergent thinking in different ways, and thereby exemplify how the two modes need to be entangled with each other to be truly efficient, at least in this ‘in between’ phase of design (Lawson, 1990). The ‘disciplined’ generation and evaluation of concepts throughout the design process was emphasized by Pugh, who also pointed to the fact that as the process evolved the number of solutions gradually decreased, as did the space for ideation (Pugh, 1991). Cross further emphasized that while concept design contained deliberate divergence for the purpose of searching for new ideas, the process as a whole was 3 characterized by convergent thinking by its very nature (Cross, 1994). These two authors have been read as together suggesting two centrally desired features of the conceptual design phase (1) the design process benefits from following a multiple divergent and convergent approach, and (2) the number of concepts generated in this process is gradually decreased to the point where only a few are left (Liu et al., 2003). Liu and Bligh have suggested that this divergent-convergent process can be further characterized by (1) the number of levels of abstraction at which divergenceconvergence take place, (2) the order in which these happen, and (3) the maximum level of divergence (Liu et al., 2003). In the case of the first of these points, the authors refer to Lee et al., who suggest that designers may reduce complexity by bundling design criteria together and considering many of these at once. Bundling several single criteria together in this way creates multiple levels of abstraction, depending on the number and type of criteria one chooses to combine, and this may aid in tackling complex design requirements (Lee et al., 1992). This also appeals to the insight that designers must reduce the number of candidate solutions simply because they are […] 3. Method (0,5-1 pages) 3.1 The methodological approach In order to address this question, this study employed an approach based on interpretative phenomenological analysis (IPA). The typical aim of this method is to explore in detail how a person or group engages in an experiences an activity in the sense of assigning meaning to this activity. The approach is phenomenological in that it emphasizes the meanings of particular events, states and experiences as these hold for the participants (Smith and Osborn, 2003). The researcher works out interpretations to large extent already present in the expressions of the participants, i.e. the subjects own elaborations and sense-making (Smith, 1996). Because of the focus on meaning structure and personal experience, the standard measures of representativeness of sampling is replaced in phenomenological analysis with an attention to detail of expression, and a probing attitude with regard to the interview itself and its resulting material. In addition to this the phenomenological approach usually also a focus on creating a homogeneous sample with regard to the phenomenon under study (Smith and Osborn, 2003). […] 4 4. The empirical study (4-5 pages) The interviews revealed three general categories or phases of the conceptual design phase which were often put, more or less consciously, in a temporal-causal narrative by the designers. Schematically this process tended to proceed as follows: After initially ideating a concept or a set of candidate concepts, mental experimentation and conceptualization where often ‘incubated’ in the form of an internal conceptualization, where images were elaborated, evaluated and refined in the ‘mind’s eye, as thought processes in the individual designer. After that, these internal images became developed into more representational conceptualizations, for example in terms of sketches; however these were still very much in the individual cognitive domain. As a third step, the idea was externalized socially, into what can be referred to as a social conceptualization, where other designers (e.g. team members) were presented with the idea and engaged in its further elaboration. This activity would then ideally launch into a phase where the designer works as part of an organized effort to further elaborate and realize the concept. However, as stated above, this study dealt with the intermediary process of conceptualization between initial idea generation and focused team work. These three stages will now be further elaborated and exemplified with quotes from the interviews. Figure 1 serves to structure these categories, how they were expressed by the interviewees (boxes under each category refer to representative meaning units from the interviews), and their temporal progression. -----------------------Figure 1 about here ------------------------ 4.1 Internal conceptualization The interviews revealed how, following the generation phase, design brief or initial ideation, there was an early, individual phase of internally conceptualizing the design idea. This phase involved a mix of mental manipulations, which seem to have the function of testing the scope, as well as the integrity and the ‘ecological feasibility’, of the idea. In the first place, the commonly known exercise of functional separation or 5 breakdown of concept was combined with concept expansion, where the functions of the future artifact were mentally isolated and then imagined in a variety of alternative ways. One of the designers explained it like this: “I had an idea about a baby chair. Then I though, a chair should be safe. What can safety be? What represents safety? Secondly, I though about the sitting surface. What can a sitting surface be? What is it supposed to do?” The concept breakdown/expansion procedure was also achieved on a higher level, in the sense that the conceptualization of the idea was goal oriented with regard to function as well as to emotional qualities. In the interviews, these two aspects seemed to be conceptualized informally on this stage, as part of the same ‘mental package’, e.g. “I was thinking about the technical function but also about how to communicate this function and how to load it with a value… how to mix the product with an experience or an emotion that makes one want to own it”. These processes in a way represent a hypothesis generating mode, where the different aspect of the idea are conceptualized and reconceptualized in order to arrive at a richer inner picture of what one is going to work with. A part of this process also involved attempting to finalize images of the future artifact, albeit not necessarily realistic such images. One designer reported on the importance of visualizing the perfect artifact; a flawless design devoid of any problems, which directly corresponds to one’s inner image: “It’s a feeling you have about what this thing could ideally be, and it turns into a picture in your head. This picture has no boring or difficult aspects; one chooses oneself what to put in and what to leave out.” The ongoing mental refinement of the concept however continues, partly by imagining alternatives and ideal end-goals but also, significantly, through mentally falsifying alternative solutions: “I constantly tested the idea in my head and searched for limitations and flaws. It was never like: Oaw there it is. More or less I tried to pressure test the idea at this stage.” It seems as if this stage is a very volatile place to be for an idea, and several participants expressed the need to quickly get the concept out and into a more visual form, for example by sketching or making notes. One participant expressed the limit to mental imagery in the following way: “I have to get it out fast, and onto paper or something otherwise I forget it… I can’t develop it in here [pointing to head], I have to 6 get a three-dimensional understanding, so that I can see it properly. After that one may discover that ‘OK this didn’t work’”. This is where the idea goes from an internal to a representational concept, and this is the phase to which we turn next. 4.2 Representational conceptualization The most obvious type of representational conceptualization is experimentation on paper, or sketching. All of the participants referred to sketching as central in this phase of conceptual design, which is not strange since it permeates the whole design process. In this phase the designers typically tried out several representations quickly, e.g. “ideas ‘fly around’ on the paper. They move around in the head, and come out directly through the pen.” In this sense it may be difficult to separate sketching from the previous phase, since the mind and the hand seem so closely connected. Later we will look at the particular role paper experimentation can play in this phase. Several participants referred to alternative ways of creating the first external representations of the concept, which did not involve sketching, but rather the creation of ‘symbolic triggers’ which could serve as external focus points for ideation. For example: “I took several photographs of the interior which I then put together in different ways to see what materials could be used and what forms could work in the context.” The use of photographs as a source of inspiration after the concept has started to take shape may fill a similar but less committal function to that of externalizing by sketching. This function is also related to an aspect complementary to sketching, which also seems important in this phase of externalization namely that, as one participant expressed it: “apart from visualizing shapes, this part of the work is lot about picking out materials and solutions, more or less consciously.” Sketching is here seen to be combined with other visual aids as concept refinement moves from shape and structure towards qualities of look and feel. The interviews illustrate how representational conceptualization through sketching and other aids may generate solutions, more problems or new concepts, by a branching out from the original idea. The generation of solutions is exemplified in the following quote: “I go from two-dimensions to three-dimensions and so on, and new solutions appear the whole way. In all sketching I’m on the look-out for a solution and this solution 7 happens on the paper.” This is a clear illustration of how the material representation affects the ability to generate solutions. It also generates new problems, as is illustrated in the following statement, which depicts the move from internal to external conceptualization: “First I have an idea and I go to sketching. But as I visualize it, new constellations of problems become clear to me. As I continue to sketch more problems occur, especially in the move from two to three dimensions.” These two outcomes of representational conceptualization are related to another typical characteristic of this phase, namely that concept sketches generate new concept sketches, which are not simple specifications of the original idea but rather a branching out into new but related concepts. Most of the respondents touched upon this phenomenon, but there was very little understanding of how this branching came about. One designer said for example that: “I have a picture in my head, I get it down on paper and I’m sketching away on it. All of a sudden I’m sketching on something else, which may be somewhat related. I get more ideas from the sketching. It’s like the sketching creates these new ideas.” However, as these ‘semi-external’ conceptualizations mature, they require new forms of validation, particularly that of the team or the group. The participants all recurrently refer to this stage as a fairly traumatic one, where one has to take a concept which has now been fairly ‘personalized’ to someone else or to the group for further pressure testing and falsification attempts. This step is also part of the externalization process (social externalization), as the concept is now re-represented ‘as communicated’ in a social setting. This stage may be referred to as ‘social conceptualization’. 4.3 Social conceptualization The first steps towards representing the concept socially may be taken informally, after the concept has fermented for a while, and gone through some of the processes elaborated above. One designer described this as the concept gradually passing through a ‘filter’, where it becomes more and more exposed to peer judgment: “I have to give the idea a few days to go through the filter, I talk to people, more or less informally. Little exchanges, just to see if it carries. Sometimes it doesn’t.” As the communication of the concept becomes more and more formalized, e.g. when the designer prepares to ‘pitch’ 8 the idea to his/her colleagues, the verbal externalization becomes more expositional in nature and this creates clarity and overview. One participant explained it like this: “Then I was in a situation when I had to present this to the others, and something happened to the idea. I got a better perspective. Maybe I had been too narrow in my focus, but when I tried to explain I started to see flaws and also new benefits.” This process may be reductive/analytical, going from the general to the specific elements of the design. One participant who had earlier been explaining the role of imagining the ‘perfect design’, described how “as one starts to communicate this image, a lot of noise is appearing. Reality comes into the picture and one has to start wrestling first with the concept as a whole, and then starting going into the details. That’s a special problem.” The initial social process of specifying and concretizing by communicating and articulating the concept seemed to […] 5. Discussion and conclusions (2 pages) The thematic structure outlined above will now be discussed in more general terms, and conclusions will be drawn for each of the identified phases of conceptual design’s fuzzy middle. After initial ideation, the first conceptualization phase appear mainly as an internal process where symbolic aspects of the future artifact are distilled and then emoted and visualized in various ways. This ‘mental molding’ of the concept seems to involve convergent as well as divergent processes. However, it is not easy to separate these. Rather they appear to co-evolve in a few distinct ways, for example through a functional separation of the concept (an analytical and reductive process), which simultaneously leads to a concept expansion (a more open explorative process). This intertwining of divergent an convergent thinking was also apparent in the way that goal orientation emerged as functional and emotional at the same time; achieving functionality criteria, which is more of a traditional problem-solving task, while seeking emotional goals involving for example aesthetic values, which is a more open ended process. The statements about visualizing a ‘perfect artifact’ is also a curious example of this duality, for while the ‘perfect artifact’ is convergently structured in the minds eye, it also 9 represents a form of unconstrained design hypothesis, and must therefore be seen as a divergent exercise. These observations are supported by authors like Lawson, who retain the analytical distinction between convergent and divergent while seeing the two as intertwined in the process of ideation (Liu et al., 2003; Lee et al., 1992). In the present case we can see how the two processes become intertwined as a result of the designer bundling criteria together in the early phases of conceptual design, and actively collapsing criteria of various levels of abstraction in order to evaluate diverse alternatives more economically and flexibly (Lawson, 1990). The representational conceptualization phase illustrates how externalization of concepts, i.e. the first steps towards embodiment of the design, can be diverse as well as creatively generative in their own right. The interviews suggest that this process can take on various modalities. The modalities of representation illustrated in this study involved paper experimentation (i.e. sketching), and the creation of external pictures to capture ideas (through photography). The last category can be viewed as representing ‘symbolic triggers’ which act as focus points for sketching and other representational forms of ideation. They are symbolic in the sense that one picture can trigger many different concrete design alternatives in the sketching process. This can be viewed as a version of the ‘picture-world cycle’ discussed by Dörner (1991), however here ‘images and words’ are not combined, but rather the ‘syntax’ of sketching is being infused with a ‘semantics’ of images, where such images are partially interpreted through sketching. Similarly to what Schön and Wiggins (1992) are describing as the ‘seeing-moving-seeing’ cycle, pictures are symbolic and elusive images of what the designer aims to accomplish; sketches are concretization attempts, which are then tested against the images, in a cyclical movement. Representational conceptualization is clearly also a very focuses generative process. As was evident from the interviews sketching created new problems, solutions and concepts in a branching out movement which was not simply a convergent specification of alternative solutions to the original idea, but one that created new concepts as well. Finally, the first steps towards embodiment proper are taken when the design concept is aired in a social setting. This part of the process implies a testing of the various aspects of the concept against new and possibly unexpected criteria raised by the group. 10 The pressure on the designer to present an emotionally as well as logically appealing narrative about the concept […] Finally, the first steps towards embodiment proper are taken when the design concept is aired in a social setting. This part of the process implies a testing of the various aspects of the concept against new and possibly unexpected criteria raised by the group. The pressure on the designer to present an emotionally as well as logically appealing narrative about the concept forces him/her to higher levels of expositional prominence than was necessary before, and in this sense social conceptualization is a creative if mainly convergent phase. There were indications that this process could be preempted to a certain extent, by exposing the concept to peers gradually, through informal interaction. This supports both the notion of ‘thought synchronization’ among designers (Valkenburg and Dorst, 1998), and illustrates one way in which designers may handle previously observed desires to keep communication targets small (Chiu, 2002). Both these issues refer to an observation made in this study, that confirmation of a concept, as important as it is to the designer, is always accompanied by the risk of falsification and concept elimination. References Chiu, M-L (2002) ‘An organizational view of design communication in design collaboration’ Design Studies Vol 23 pp 187-210 Cross, N (1994) Engineering Design Methods – Strategies for Product Design John Wiley & Sons, Chichester UK Dörner, D (1999) ‘Approaching design thinking research’ Design Studies Vol 20 pp 407415 11 Gero, J S and Tang, H-H (2001) ‘The differences between retrospective and concurrent protocols in revealing the process-oriented aspects of the design process’ Design Studies Vol 22 pp 283-295 Giorgi, A and Giorgi, B (2003) ‘Phenomenology’ in J A Smith, (ed.) Qualitative Psychology Sage Publications, London pp 25-50 Hellström, T (2005) ‘Role-taking, role-breaking and role-shaking among designers: A qualitative study of collaborative design’ The Design Journal, forthcoming Lawson, B (1990) How Designers Think Butterworth Architecture, Oxford Lee, C L, Lyenger G and Kota S (1992) ’Automated configuration design of hydraulic systems’ in Proceedings of the Second International Conference on AI in Design, Boston USA Liu, Y-C, Bligh, T and Chakrabarti A. (2003) ‘Towards an ideal approach for concept generation’ Design Studies Vol 24 pp 341-355 Liu, Y-T (1996) ‘Is designing one search or two? A model of design thinking involving symbolism and connectionism’ Design Studies Vol 17 pp 435-449 12