Many empirical studies have highlighted the importance of freehand sketches in facilitating design ideation. In comparing representation techniques via protocol analysis, Vinod Goel showed more design transformations were achieved with ill-structured representations (freehand ambiguity) than with the well-structured representations (digital precision) during the early design conception.¹

Studies by Gabriela Goldschmidt have also shown how representations ‘defer commitment to a solution’. By calculating ‘lateral transformations’, Goldschmidt’s Linkography deconstructs the creative process by parsing the recorded design protocol into small units called ‘design moves’: ‘a step, an act, an operation, which transforms the design situation relative to the state in which it was prior to that move’. A ‘linkograph’ is built by interpreting the links between the moves, providing graphical representations of design reasoning. The design process can then be viewed in terms of five patterns: a) ‘chunk’ moves, exclusively linked; ‘web’ moves, minimally linked; ‘sawtooth’ moves, uniquely linked; ‘backlinks’, linked to a move’s generation; and ‘forelinks’, linked to the production of further moves.²

For Goldschmidt, design productivity is inextricably connected to the ‘link index’ (the ratio between the number of links and moves) and ‘critical moves’ (forelinks, backlinks or both). Thus, high values of link index and critical moves reveal more creatively productive design processes. For example, in the diagram below Architect (A) was working with Landscaper (L) to design an art gallery on a triangular site with level changes. Each utterance of the session was recorded and tagged sequentially as a ‘design move’ beginning with A01.

Table 1. Extract from the transcript at the early stage of the session

Despite the flexible and scalable nature of Goldschmit’s method, recent studies by Kan & Gero have shown Linkography to be a rather unreliable indicator of ambiguity.³ Indeed, variations in linkographic analyses highlight a model greatly biased by subjective interpretation. As Gazzaniga‘s work on hemispheric specialization has shown, analytical devices like Linkography remain problematic due to an unconscious interpreter attempting to apply linear analytic structures to fundamentally abstract cognitive procedures.

1. Vinod Goel (1992). ‘Ill-Structured Representations’ for Ill-Structured Problems. Proceedings of the Fourteenth Annual Conference of the Cognitive Science Society.

2. Gabriella Goldschmidt (1990). Linkography: Assessing Design Productivity. Paper presented to the Cyberbetics and System, Singapore.

3. Jeff Kan & John Gero (2006). Acquiring Information from Linkography in Protocol Studies of Designing. Key Centre of Design Computing and Cognition, University of Sydney.