31

Click here to load reader

Adaptation of Sources of Inspiration in Knitwear Design

  • Upload
    martin

  • View
    212

  • Download
    0

Embed Size (px)

Citation preview

Page 1: Adaptation of Sources of Inspiration in Knitwear Design

This article was downloaded by: [Florida International University]On: 22 December 2014, At: 08:52Publisher: RoutledgeInforma Ltd Registered in England and Wales Registered Number: 1072954 Registered office: MortimerHouse, 37-41 Mortimer Street, London W1T 3JH, UK

Creativity Research JournalPublication details, including instructions for authors and subscription information:http://www.tandfonline.com/loi/hcrj20

Adaptation of Sources of Inspiration in KnitwearDesignClaudia Eckert & Martin StaceyPublished online: 08 Jun 2010.

To cite this article: Claudia Eckert & Martin Stacey (2003) Adaptation of Sources of Inspiration in Knitwear Design,Creativity Research Journal, 15:4, 355-384, DOI: 10.1207/S15326934CRJ1504_5

To link to this article: http://dx.doi.org/10.1207/S15326934CRJ1504_5

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) containedin the publications on our platform. However, Taylor & Francis, our agents, and our licensors make norepresentations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose ofthe Content. Any opinions and views expressed in this publication are the opinions and views of the authors,and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be reliedupon and should be independently verified with primary sources of information. Taylor and Francis shallnot be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and otherliabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to orarising out of the use of the Content.

This article may be used for research, teaching, and private study purposes. Any substantial or systematicreproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in anyform to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://www.tandfonline.com/page/terms-and-conditions

Page 2: Adaptation of Sources of Inspiration in Knitwear Design

Adaptation of Sources of Inspiration in Knitwear Design C. M. Eckert and M. K. StaceyAdaptation of Sources of Inspiration

Claudia EckertUniversity of Cambridge

Martin StaceyDe Montfort University, Milton Keynes

ABSTRACT: In an experimental study of designing byadaptation, professional and student knitwear design-ers were videotaped designing sweaters based on a Per-sianrugora19thcentury tapestry.Thedesignersusedarange of source-triggered and goal-directed adapta-tion strategies to create adaptations ranging from theclosest possible translations into the medium to radicaltransformations of abstract characteristics. While eachstrategy sometimes led to each type of adaptation, thesource-triggered strategies were predominant for theeasy-to-adapt source (the rug) and typically led to closeadaptations; while the goal-directed strategies weremore common for the more difficult source (the tapes-try), and more often led to more radical transformationsof the source. The professional designers made moreuse of goal-directed strategies than the student design-ers. The study supports the view that creative behaviorcan usefully be described in terms of consistent patternsresulting from both task demands and from cognitivecapacities and learned skills.

All visuospatial design activities draw ideas from bothprevious designs and other sources of inspiration. Theinfluence of various kinds of ideas on designed prod-ucts is a central theme in cultural studies of design. Buthow designers use these sources of inspiration has beena relatively neglected question. What roles do differenttypes of source play in the design process? What rolesdo sources of inspiration play in design thinking? Whatadaptation strategies do designers use? Do they employa limited number of category prototypes as opposed to amultiplicity of individual category instances? This pa-per reports an experimental study of creating designsby adapting sources of inspiration, in which we observe

knitwear designers employing rational problem-solv-ing strategies for constructing design subproblems andmaking design moves.

The selection and adaptation of sources of inspira-tion isavitally important andopenlyacknowledgedpartof commercial knitwear design. We have conducted anextensive observational study of the knitwear designprocess in industry, combining ethnographic methodsdrawn from the social sciences with knowledge acquisi-tion techniques from artificial intelligence (see Stacey& Eckert, 1999, for our methodological approach). Weobserved and interviewed over 80 designers and techni-cians in 25 companies in Britain, Germany, and Italy;these companies spanned the full range fromdownmarket mass fashion manufacturers to fashionleading designer-label companies, and included a num-ber of direct competitors producing equivalent prod-ucts. We have developed a detailed process model ofknitwear design (Eckert, 1997b) and an analysis of the

Creativity Research Journal 355

Creativity Research Journal Copyright 2003 by2003, Vol. 15, No. 4, 355–384 Lawrence Erlbaum Associates, Inc.

Claudia Eckert’s research on sources of inspiration in knitwear de-sign was supported by ESRC grant L12730100173 for the projectMechanisms of Inspiration in Novel Design (MIND). Her other re-search on knitwear design has been supported by SERC/ACME grantGR/J40331 and Open University Research Development Committeegrant 717. Part of Martin Stacey’s contribution to this work was madewhen he was supported by EPSRC grant GR/J48689. Dr. KenGilhooly of the Department of Psychology at the University ofAberdeen commented helpfully on an earlier draft of this article. Weare grateful to our informants and experimental subjects for the timeand effort they put into participating in our research, especially toMonica Jandrisits, Annabelle Duncan, and Wendy Nicholson.

Correspondence and requests for reprints should be sent to Mar-tin Stacey, School of Computing, De Montfort University, Leicester,LE1 9BH, United Kingdom. E-mail: [email protected]

Dow

nloa

ded

by [

Flor

ida

Inte

rnat

iona

l Uni

vers

ity]

at 0

8:52

22

Dec

embe

r 20

14

Page 3: Adaptation of Sources of Inspiration in Knitwear Design

communication problems from which it suffers (Eckert,1997b, 1999, 2001; Stacey, Eckert, & McFadzean,1999), as well as an analysis of how the design process isinfluenced by the differing business models of differentcompanies (Eckert & Demaid, 2001). We have pro-posed managerial strategies (Eckert, 1999; Eckert &Demaid, 1997) and computer support techniques(Eckert, 2001; Eckert, Cross, & Johnson, 2000) formaking the communication of knitwear designs moreeffective.

The objectives of the Mechanisms of Inspiration inNovel Design (MIND) Project at the Open Universitywere to understand the roles that sources of inspirationof different types play in the design process and howsources of inspiration are transformed into designs, fo-cusing on knitwear design as an example domain withinteresting characteristics. As well as the experimentreported here, the project also included further inter-views and observations of designers working in indus-try, from which we have developed an analysis of howknitwear designers use sources of inspiration to ac-tively construct their design context and formulate de-sign problems, as well as create individual designs(Eckert & Stacey, 2001, 2003).

It is well recognized that designs are not createdout of nothing, and the processes of designing haveattracted more interest in architecture and engineer-ing than in fashion and textiles. Persistent themes inresearch on architectural design have included therole of its environment in guiding the design of abuilding (Darke, 1979), the role of precedents in theapplication of techniques and styles (for instanceGoldschmidt, 1995, 1998), the form and role of archi-tects’ knowledge and experiences (for instance Gero,1990; Oxman, 1990; Schön, 1988), and the part ar-chitects’ own sketches play in stimulating idea gener-ation (notably Goel, 1995; Goldschmidt, 1991, 1992,1994; Schön, 1983; Schön & Wiggins, 1992; seePurcell & Gero, 1998, for a review). Similarly, muchof engineering design is producing variants of exist-ing designs (see for instance Otto & Wood, 2000),and facilitating design reuse is a major research area(see Finger, 1998). In tailoring, developing a garmentdesign by creating new blocks rather than modifyingold ones is now a rare or eccentric activity—com-puter-aided design (CAD) systems for tailoring aredesigned around adaptation as the primary activity.

Although most research has focused on the adap-tation of similar previous designs, designers, espe-

cially in knitwear design, draw ideas from a wide va-riety of other sources, for instance, the rhubarb leafthat inspired Ove Arup’s design of the Kingsgatefootbridge in Durham, England (Walker, 1983). LeCorbusier’s design for the chapel at Ronchampadapted the forms not just of North African windowsand Greek roof turrets, but also the shell of a crab LeCorbusier had collected years before (Broadbent,1973; see Goldschmidt, 1998).

We describe how sources of inspiration are used inthe knitwear design process in industry in the next sec-tion, before discussing the source adaptation experi-ment itself. We report three main sets of results: thetypes of adaptation we have identified, the strategiesdesigners use to adapt a source in creating a single de-sign, and statistical analyses of how the designers inour experiment used these adaptation strategies to pro-duce adaptations of different kinds.

Uses of Sources of Inspirationin Knitwear Design

Knitwear designers use previous garments andother sources of ideas throughout the aesthetic designof a garment, from initial design research and collec-tion planning to the development of the conceptual de-signs of individual garments that are handed over toknitwear technicians for detailed design and imple-mentation (Eckert & Stacey, 2003). (The techniciansalso make use of programs for previous garments andswatches of fabric when developing programs for newgarments.) Sources of inspiration play two major rolesin aesthetic design:

1. Defining the themes, topics, cultural connota-tions, and moods of particular fashions; and delineatingthe spaces of acceptable designs within those fashions.Both designers and ordinary people identify styles andtrends in what they see, abstracting from the similari-ties between individual garments as well as interpretingother information, such as written descriptions and as-sociating styles with times and situations. Our inter-view evidence indicates that designers reason aboutwhether striking features of individual garments areunique or are exemplars of categories forming newtrends. They also reason explicitly about the develop-ment of the forms, cultural associations, and accept-ability of their most general, labeled categories includ-

356 Creativity Research Journal

C. M. Eckert and M. K. Stacey

Dow

nloa

ded

by [

Flor

ida

Inte

rnat

iona

l Uni

vers

ity]

at 0

8:52

22

Dec

embe

r 20

14

Page 4: Adaptation of Sources of Inspiration in Knitwear Design

ing the historical styles that may reappear. Thus,sources of inspiration contribute to the formulation ofdesign problems—intentions to create designs withparticular relationships to the spaces of fashionablegarments (see Eckert & Stacey, 2001).

2. Supplying aspects of individual designs throughprocesses of adaptation, transformation, and combina-tion. The research reported in this article examines thesecond of these uses in isolation, but they cannot be soclearly separated in industrial practice. When they areplanning collections, many designers (who have vividand powerful visual imaginations) think in terms ofconcretely imagined individual garments. These serveas placeholders for categories to be included in a col-lection, though some of these might be developed intofinished designs and manufactured. Sources of inspira-tion used in collection planning may be reused in thedesign of individual garments.

The Design Process

The commercial knitwear design process sharesmany features of engineering design, though on asmaller scale. It involves considering a variety of aes-thetic and technical factors and meeting tight deadlinesand cost targets. It is made complex by the subtle inter-action between the structure of complex knitted fabricsand their appearance: An apparently small change tothe structure may have a radical effect on the appear-ance, whereas an apparently small change to the ap-pearance may make a design much more expensive toproduce or completely infeasible. Knitwear design is ateam activity divided between the knitwear designers,who are responsible for the conceptual design of thegarment, and the knitwear technicians, who do a lot ofdetailed design in the course of using sophisticatedCAD systems to program knitting machines to manu-facture the garment. The interaction between these twois problematic primarily because communicating knit-ted structures is inherently very difficult (Eckert,1997b, 2001; Stacey, Eckert, & McFadzean, 1999) andbecause designers, technicians, and managers in indus-try fail to recognize communication problems or real-ize their significance (Eckert, 1997b, 1999; Eckert &Demaid, 1997). Moreover, designers and techniciansare very different in background, education, and inter-ests as well as expertise (Eckert & Stacey, 1994). Thedesigners hand their designs over to the technicians in

the form of technical sketches. These comprise a briefverbal description, a freehand sketch, and parametervalues for the dimensions of the garment—called mea-surements—that may be inaccurate, incomplete, andinconsistent. Often the only feedback the designers geton their designs are finished sample garments.

Figure 1 shows the major stages in the knitwear de-sign process up to the hand over from the knitwear de-signer to the technician. This is an outline summary ofthe very much more detailed process analysis pre-sented by Eckert (1997b). In terms of the activities andthe connections between them described in that analy-sis, the knitwear design process was identical in all thecompanies we studied throughout the industry. How-ever, companies differed significantly—according totheir market sector—in how much effort was investedin each activity and in how much or how often the pro-cess involved iteration and backtracking to earliersteps—typically the more upscale companies werewilling to invest more effort in individual designs.Companies also differed in the attention paid to partic-ular types of sources of inspiration and other informa-

Creativity Research Journal 357

Adaptation of Sources of Inspiration

Figure 1. Outline of the knitwear design process.

Dow

nloa

ded

by [

Flor

ida

Inte

rnat

iona

l Uni

vers

ity]

at 0

8:52

22

Dec

embe

r 20

14

Page 5: Adaptation of Sources of Inspiration in Knitwear Design

tion inputs. Moreover, designers in Britain, Germany,and Italy differ in their training and attitudes, if not inthe stages of their design processes; we have no dataabout other countries (see Eckert, 2001). The asser-tions made in this article about the behavior of design-ers refer to commercial knitwear designers, and applyas nearly universally as we can judge from ourobservations.

Types of Sources

Any object or image or scene can be a source ofideas for a designer, and any abstract idea might sug-gest a usable image. But some kinds of objects aremore commonly used as sources in knitwear design.

Garments. Aspects of existing garments can beadapted more directly than other types of source.Looking at the real thing provides much more detailedinformation, especially about how a feature isconstructed, while images of garments provide morecontext, especially suggestions of moods and culturalconnotations.

Other textiles. Designers frequently draw motifsfrom printed fabrics and other media such as carpets.These are often translated into knitted form with theleast possible amount of alteration, though other fea-tures or aspects might be passed through more radicaland abstract transformations.

Other designed products. Artifacts such as mo-saics sometimes provide adaptable motifs, while threedimensional artifacts like buildings and stonework canprovide design elements that can be translated into atwo dimensional form.

Works of art. Artworks, mainly paintings, cansupply indications of mood and cultural associations aswell as color schemes, motifs, and suggestions ofshapes.

Natural objects. Objects such as flowers or shellssupply both shapes for motifs and color combinations.

Natural phenomena. Images of scenes such asthunderstorms or tropical beaches are often used assources of color schemes as well as indicators of moodsand cultural associations.

Many designers keep collections of source materi-als including art and craft books, exhibition cata-logues, fashion magazines, postcards, pictures torn outof magazines, and so on. They regard collecting poten-tial sources as an important part of their job, and mostcomplain that their companies do not recognize the im-portance of source collecting or provide enough re-sources for it (Eckert, 1997a; Eckert & Stacey, 1998).

Mental Representationsof Garment Designs

Experienced designers have commented to us thatthe ability to visualize garments is the most importantattribute of a good designer. Most of the knitwear de-signers we have talked to tell us that they see their de-signs mentally as detailed, realistic images ofgarments, similar to photographs. How complete anddetailed these mental images really are is hard to as-sess; research on imagery indicates that details in sub-jectively rich mental images may often not exist untilpeople focus on a particular area or detail (for instanceKosslyn, 1980, 1994; Logie, 1995). But it is quite com-mon for knitwear designers to create, evaluate, and dis-card designs in their heads, sketching only tocommunicate. (This seems to be relatively rare, thoughnot unknown, among architects. Frank Lloyd Wrightmade few freehand sketches and advocated completingdesigns in one’s mind before committing pen to paper;see Goldschmidt, 1995.) We suspect that in much in-dustrial practice the information missing from knit-wear designers’ mental images concerns details andaspects of designs that designers leave to their techni-cians to sort out for them.

While we have only verbal report evidence aboutknitwear designers’ mental representations of indi-vidual garments and garment categories, we canmake inferences from the large body of psychologicalresearch on learning, memory, and expertise. Knit-wear designers possess a technical understanding ofthe structures of garments, enabling them to createhighly structured mental representations; they seehuge numbers of garments, and many report being

358 Creativity Research Journal

C. M. Eckert and M. K. Stacey

Dow

nloa

ded

by [

Flor

ida

Inte

rnat

iona

l Uni

vers

ity]

at 0

8:52

22

Dec

embe

r 20

14

Page 6: Adaptation of Sources of Inspiration in Knitwear Design

able to remember a large number in vivid detail. Forthe knitwear designers we have studied, mental repre-sentations of types, trends, and spaces of acceptablegarments within a fashion are visuospatial. Success-ful designers have both aptitude and skill for under-standing fashion and how their own and others’garments fit into it. While they reason explicitlyabout the development of fashion, their understand-ing is largely tacit and perceptual: Designers recog-nize appropriateness within particular fashions.

Although designers’ memories include details ofboth exact form and context, research on perceptuallearning (see Goldstone, 1998) and expertise in, for in-stance, radiology (for instance Myles-Worsley, John-son, & Simons, 1988), chess (for instance Gobet &Simon, 1998), and electronics (Egan & Schwartz,1979), as well as mental imagery, indicates thatvisuospatial representations are highly structured, in-corporating categorizations of both structural featuresand emergent visual features. Aiken (1978) presentedevidence that architects’ memories for architecturaldrawings depend on schematic encoding of drawingchunks. It is difficult to assess how much of the mentalrepresentation of a garment is unique to it and howmuch is reconstructed from representations of moregeneral categories. The structure and redundancy inmental representations enables details to be recon-structed from sparse mental descriptions. Studies ofmemory for drawings of faces (Wulf, 1922) and forstories (Bartlett, 1934) have shown that unusual fea-tures that are perceived as significant are highlightedand exaggerated, whereas other unusual features aresmoothed toward what is standard for the category (seeKoriat, Goldsmith, & Pansky, 2000). Perceptual rec-ognition of an object or scene as a member of a cate-gory (which involves the use of the categoryrepresentation to construct a representation of the indi-vidual) can distort what people perceive, highlightingsalient unusual features and minimizing others, as wellas enabling them to perceive the object or scene as aconfiguration of particular components (see Gold-stone, 1998).

The sources of the ideas designers use furnish themwith a language for describing their designs to theircolleagues. We have observed that knitwear designerstalking amongst themselves (rather than to techniciansor managers) refer constantly to previous designs, theirown and those they have seen in shops or in fashionmagazines, describing new design ideas in terms of

modifications and combinations of elements ofprevious designs. At least in knitwear design, this is asignificant feature of the design culture (Eckert &Stacey, 2000). Knitwear designers have a vocabularyfor garment features, but the range of available verballabels for garment categories is very much smaller thanthe range of possible categories. Designers’ use of ref-erence-based descriptions, and their reports of havingvivid and detailed memories of large numbers of gar-ments they have seen, indicates that their mental repre-sentations of the space of possible garment designsprimarily comprises very many garment instancesserving as exemplars of subtly differentiated subcate-gories that can only be referred to by their origins.

Designers in other fields make comparable use ofmemories for both generalizations and individual de-signs: Schön (1988) describes functional types and ref-erences as forms of architectural design knowledge.Drawing on the cognitive theory of dynamic memory(Schank, 1982; Schank & Abelson, 1977), Oxman(1990) argues that precedents are used in design as pro-totypes, through a process of typification—in whichindividual designs and problems are used to create andrefine more abstract generalizations—and are indexedin memory by these generalizations.

Sources of Inspiration in CollectionPlanning

Before the knitwear designers we have studiedwork systematically on designing individual garments,they plan the collections their companies will offer fora particular season. Designers begin their work on anew season by researching coming trends. This is doneby looking at materials produced by fashion forecast-ing bureaus, fashion trend publications such as BookModa and Zoom on Fashion Trends, fashion maga-zines such as Vogue, and at garments produced byhaute couture designers and more upscale high streetproducers, as well as those of their direct competitors(see Eckert & Stacey, 2001, 2003). Their objective is tounderstand the envelopes of acceptable garmentswithin particular fashion trends and the cultural conno-tations of garments in particular parts of each space offashionable garments. Designers can then select somefashion trends for their own collections and decidewhat kinds of garments they want within each fashiontrend in relation to the garments that both upscale

Creativity Research Journal 359

Adaptation of Sources of Inspiration

Dow

nloa

ded

by [

Flor

ida

Inte

rnat

iona

l Uni

vers

ity]

at 0

8:52

22

Dec

embe

r 20

14

Page 7: Adaptation of Sources of Inspiration in Knitwear Design

trend-setting companies and their competitors willproduce. Fashion is an emergent consequence of all theworld’s designers performing essentially similar re-search activities, aiming to produce distinctive butsimilar designs (see Eckert & Stacey, 2001).

At the end of the research stage of the design pro-cess in industry, knitwear designers create descriptionsof the ranges they intend to produce. They comprisebrief verbal specifications of the categories the designsshould belong to, often accompanied by sketches ofdesigns that serve as placeholders for categories.Usually very brief and informal written notes orsketches serve as cryptic cues for much richer mentalrepresentations. The categories of garments that de-signers decide to include in ranges define not just thefunctions of garments, but what themes they fit into,and—in qualitative categorical terms—their shapes,the features they include (pockets, collars, buttons andso on), and their decoration. These category descrip-tions include the intended overall visual and tactile ef-fects, and how the types of stitch structures and designelements are to be arranged to achieve them, but usu-ally not any details of the decorative elements. Forsome companies (see Eckert & Demaid, 2001), briefsfrom buyers determine the categories, but these briefsmust be fleshed out by fashion research.

Our interview evidence shows that garment catego-ries are fundamental to design thinking; they cover theentire search space of possible garments, and the moreabstract and general categories provide the only waydesigners have to describe garments verbally. De-

signers think about how decorative elements—mo-tifs—can be placed and combined, in terms of a smallfinite set of options (see Table 1). Thinking of differentarrangement options for a motif is a natural action. Mo-tifs can occur singly, as the horse does in the front viewin Figure 2, or repeated to form overall patterns, or hor-izontal or vertical stripes (called borders) in various lo-cations, as the horse does in the back view in Figure 2.Different motifs can also be combined or juxtaposed toachieve a combined effect. Thus, motifs can take a va-riety of structural roles in a design. In our experiment,many designers used the horse in the rug and the mainleaf swirl in the tapestry as single motifs. (Some simpleborders and overall patterns are designed as borders ofindefinite length or as fill patterns rather than com-posed from individual elements, for instance, the verti-cal borders in Figure 2.)

The ways motifs are combined are governed bychanges in fashion, just as are shapes and colorschemes, and designers often set out to create designsthat have motifs in particular structural roles within apredefined plan. For instance, a very fashionable stylefor men’s sweaters designed in Britain in 1993 andsold in 1994 and 1995 was overall color patterns com-prising several horizontal bands (borders) of relativelycomplex pattern on a uniform background color. Thedesign shown in Figure 2 would have been very fash-ionable in the mid-1980s, but very unusual when theexperiment was conducted in 1997. Large single mo-tifs and figurative designs were out of fashion for win-ter 1998–1999 and summer 1999, which were the

360 Creativity Research Journal

C. M. Eckert and M. K. Stacey

Table 1. Categories of Motif Arrangements

Single motif Location CentralOff centerOn pocket

Border pattern Direction HorizontalVerticalDiagonal

Location CentralLeft/topRight/bottomYoke

Symmetry on garment SymmetricAsymmetric

Connectedness UnconnectedConnected

Overall pattern Arrangement StripesRegularPseudo-random Figure 2. Single motif design with asymmetric shapes.

Dow

nloa

ded

by [

Flor

ida

Inte

rnat

iona

l Uni

vers

ity]

at 0

8:52

22

Dec

embe

r 20

14

Page 8: Adaptation of Sources of Inspiration in Knitwear Design

collections the professional designers were working onwhen they participated in the experiment, but are re-turning early in the next decade.

Yarns and color schemes are selected very early inthe design process (see Eckert & Stacey, 2003). Colorschemes are strongly influenced by the ranges pro-vided for each season by the yarn manufacturers, whoare guided by the fashion forecasts produced by theColor Marketing Council. Designers draw colorschemes from images with appropriate cultural conno-tations, often ones showing colors typical of some cul-ture or geography, for instance, photographs of surfersriding waves for garments with nautical or sporty orsummery associations. Such images are included inmood boards with sketches of garments and swatchesof fabric to convey the cultural associations of in-tended garments and garment categories as well astheir form.

Conceptual Design of IndividualGarments

Commercial knitwear designers focus their creativeefforts almost exclusively on meeting the needs identi-fied by their collection plans for garments of particulartypes within particular themes (the only requirementsspecification they ever have). Other garments are theprimary sources of shapes and structural features, andof emergent aesthetic effects to be achieved, but knit-wear designers frequently look elsewhere for sourcesof decorative motifs (see Eckert & Stacey, 2003). Oncethe themes for a new season have been identified, de-signers look for sources of inspiration within a particu-lar theme. For example, if the theme is “Persia,” theymight look at a book on Persian rugs.

This search can be for a source the designer has seenpreviously and remembers, or for a source resemblingsomething remembered. More typically it will be asearch that is focused by the desire to find a source thatsuggests a garment with some particular characteristic.This characteristic might be a motif of a particular cat-egory, such as a tulip, or it might be an emergent per-ceptual property, such as “aquaticness.” For instance adesigner might look for a particular carpet or for a car-pet containing a rich dark orange with blue. But de-signers often search simply for something that strikestheir imagination (though their imagination is sharplytuned to their brief or collection plan). As we have ob-

served in the experiment reported here, designers areeither driven by a plan—specifying the arrangement ofthe elements of the design, and some of the emergentaesthetic effects of these elements, in selecting sourcesto adapt—or they look for a source that will inspirethem to generate a plan. Designers may make severalrounds of searches for sources of inspiration for differ-ent elements of a design, when partial designs suggestthe need for other design elements of particular types.

Occasionally, designers see a source and, in whatthey experience as a single leap of inspiration, con-ceive a design for a garment that they subjectivelyexperience as complete and detailed. As we note pre-viously, it is not obvious what content a “subjectivelycomplete” mental representation of a design actuallyhas. This is a striking phenomenon that is sufficientlywidespread that designers comment on it frequently,but it is relatively rare in industrial practice, com-pared with more sequential processes of design gen-eration illustrated in Figure 3. Designers typicallybegin with an idea, search for and select a source ofinspiration they can use to realize that idea, adapt thesource, and evaluate the resulting design; however,sources of inspiration frequently trigger the creationof design ideas. If the design is unsatisfactory, the de-signers can go back to the source to create a differentadaptation, or select a new source, or discard the idea

Creativity Research Journal 361

Adaptation of Sources of Inspiration

Figure 3. Adaptation of a source of inspiration for an individualdesign.

Dow

nloa

ded

by [

Flor

ida

Inte

rnat

iona

l Uni

vers

ity]

at 0

8:52

22

Dec

embe

r 20

14

Page 9: Adaptation of Sources of Inspiration in Knitwear Design

and find a new one. Once the design is carried for-ward into detailed design, backtracking will only goas far as changing the adaptation of the source; oncethe design is prototyped, only the detailed design willbe changed to overcome problems (Eckert & Stacey,2003).

The elements that are extracted from sources of in-spiration require translation into the medium in whichthe design is created. Motifs are often used in knittedpatterns with the minimum possible amount of modi-fication, but as knitwear has a low resolution (a sin-gle stitch is quite large) and a limited set of colors,almost all motif sources require some simplification.Creating a gridded pattern from a finer-resolution im-age is not a mechanical process. The results dependon who does it. The knitwear designers we have ob-served seem not to be very conscious of this andoften delegate creating gridded patterns to their tech-nicians. By contrast, carpet designers, who also cre-ate gridded patterns from finer-resolution sources, arevery well aware of how subjective and variable this is(A. Demaid, personal communication, July, 1996).

The Experimental Study

We conducted an experiment to observe the use ofa source of inspiration in the conceptual design of in-dividual knitted garments to supplement the MINDProject’s observational study of the use of sources ofinspiration in industrial practice. The experimentalsubjects, who were either professional knitwear de-signers or experienced students completing knitweardesign degrees, were given a source of inspiration (apicture of a Persian rug or of a 19th century tapestry)and asked to use it to create designs for sweaters.

Our aim was to investigate the actions and reason-ing processes involved in adapting sources to gain amore detailed understanding of part of the process ofdesigning by adaptation. We started from the as-sumption that the adaptation process was rational.The experiment was intended to test the hypothesisthat knitwear designers use a repertoire of identifiablestrategies for creating adaptations and to reveal thesestrategies. We also aimed to assess designers’ use ofgoal-directed reasoning in design by adaptation andto assess the role of the designers’ sketches in the de-velopment of their designs. In the following sections,

we describe the range of adaptations made by design-ers and the strategies designers employed to generatedesigns from the source. The primary objective of theexperiment was to identify the range of different pat-terns of behavior followed by the designers; obtain-ing useful statistics was only a subsidiary objective.We conclude by describing some statistical analysesof the frequency of the different behavior patterns,but the results can only be regarded as tentative.

The Experimental Task

The designers were given several copies of a gar-ment outline printed on sheets of A4 paper and an im-age (either the rug shown in Figure 4 or the tapestryshown in Figure 5) as a source of inspiration. Theywere offered a pencil and eraser and a packet of 12 col-ored pencils; however many of the designers broughtand used their own pens. The designers were given thefollowing brief and were allowed to work uninter-rupted for about 30 min. The experimenter observedthe designing silently, and then conducted a debriefinginterview. The subjects were videotaped throughoutthe experiment and the debriefing.

This is a picture of a rug/tapestry. I would like youto use it as a source of ideas and inspiration for de-signing sweaters. I would like you to design one ormore menswear t-sleeve sweaters with turtle-necks. They can be as close to the source as youlike, or as different as you like. Just draw whateverdesigns you want. You can use the outline shape ora blank piece of paper. Don’t worry too muchabout details, but if you feel you need to worksomething out in detail, you can use the graph pa-per. Just keep going until you’re happy that whatyou’ve drawn expresses what you want it to ex-press. It would be great if you could talk aboutwhat you’re doing as you go along, but don’t letthat interfere with your designing. Don’t worry ifit does not happen. You have got about 30minutes.

We did not explicitly state a topic or fashiontheme in the experimental design, so as not to con-strain the designs more than necessary. However,some of the subjects asked the experimenter what thetopic was. For the rug task the topic was named as

362 Creativity Research Journal

C. M. Eckert and M. K. Stacey

Dow

nloa

ded

by [

Flor

ida

Inte

rnat

iona

l Uni

vers

ity]

at 0

8:52

22

Dec

embe

r 20

14

Page 10: Adaptation of Sources of Inspiration in Knitwear Design

Creativity Research Journal 363

Adaptation of Sources of Inspiration

Figure 4. The rug stimulus given to the subjects. Afshar rug (S. E. Iran). Scanned from Oriental Rugs: A Buyer’s Guide (p. 42), by Lee Allane,London: Thames and Hudson, 1988. Copyright 1988 Thames and Hudson. Reprinted with permission.

Dow

nloa

ded

by [

Flor

ida

Inte

rnat

iona

l Uni

vers

ity]

at 0

8:52

22

Dec

embe

r 20

14

Page 11: Adaptation of Sources of Inspiration in Knitwear Design

364 Creativity Research Journal

C. M. Eckert and M. K. Stacey

Figure 5. The tapestry stimulus given to the subjects. Owl and Pigeon, designed by John Henry Dearle, c. 1895. Scanned from William Morris (p.250), edited by Linda Parry, 1996, London: Philip Wilson Publishers/Victoria and Albert Museum. Copyright 1996 V & A Images/Victoria andAlbert Museum. Reprinted with permission.

Dow

nloa

ded

by [

Flor

ida

Inte

rnat

iona

l Uni

vers

ity]

at 0

8:52

22

Dec

embe

r 20

14

Page 12: Adaptation of Sources of Inspiration in Knitwear Design

“Persia.” For the tapestry, the topic was given as“Morris,” even though the tapestry was not designedby Morris himself. Seven of the nine professional de-signers and three of the student designers were givenboth sources of inspiration; the order was alternatedbetween subjects. Two of the professional designersparticipated when they were too busy to have time forboth tasks.

External Validity of the ExperimentalDesign

All but two of the subjects enjoyed doing the exper-iment and appeared to be motivated to produce designsthat fitted the brief. They were comfortable working tothe brief, and found it a reasonably natural experience.Two of the students did not like the experiment, eventhough they were paid to participate in it; so far as wecould tell this was because they resented being given adesign brief. They were motivated to produce good de-signs, but not restricted by the brief. Working to a briefand designing against one’s personal taste is a normalpart of a professional designer’s experience.

Although the subjects were not given any con-straints on what sort of sweaters they designed otherthan the garment shape (which a few chose to modify),they almost invariably produced color pattern designsusing stripes and motifs drawn in various ways fromthe sources. Creating complex color pattern designsusing elements adapted from sources is a very commonactivity in commercial knitwear design, which all oursubjects were familiar with, even though our profes-sional designers would not normally produce them inthe normal course of their work in 1997, when theyparticipated in our experiment.

Rug designs are a common source of inspiration inindustrial practice, as collections of pictures of rugscontain large numbers of relatively coarse-grainedand easily adaptable motifs. Several of our partici-pants commented on having used rugs as sources inthe past. With the exception of the two participantswho disliked the experiment, our participants thoughtthat the rug was a very good choice of source thatthey liked and found easy to work with, and that theymight choose to use it. The tapestry was chosen as avery different but comparably rich image. It provedvery much less successful as a source. Our subjectsfound it difficult to work with, and most said that

they would not choose it themselves. Only two saidthey preferred the tapestry condition of the experi-ment to the rug condition, as they found it more of achallenge. One reason for the unpopularity of the tap-estry (we believe the major one) is that the smoothcurves of the leaf pattern do not lend themselves tothe construction of coarse-grained grid patterns,though many of our subjects thought the leaf patterntapestry too feminine for menswear. The brief speci-fied a man’s t-sleeve sweater because menswear ismore restricted than ladieswear in shape, patterns,and material, and we wished to focus the designer’sactivities on using design elements from the sourceimages. All our professional designers designed orhad designed men’s sweaters for a living, or hadworked closely with menswear designers and partici-pated in decision making about menswear collec-tions. Shape outlines of the sort supplied in theexperiment are commonly used in industry.

The major difference between our experimentaltask and designing color pattern garments in industrialpractice is that working designers usually choose theirown sources, with practical needs and aesthetic objec-tives clearly in mind; they are free to select or discardsources that meet previously formulated needs. For-mulating the right needs, to create garments that fit intothe context of contemporary fashion and match a com-pany’s target market, is a professional knitwear de-signer’s most important skill, and selecting the rightsources is a crucial part of the creative process (Eckert& Stacey, 2001; Eckert, Stacey, & Clarkson, 2000).Our experiment gave our subjects no need or opportu-nity to exercise this skill. While designers look atsources to see what they suggest, they select only themost suggestive rather than the images other peoplethink would be effective.

Therefore, the constraints and direction imposed bythe source in the experiment are much more arbitrary.We would expect this to cause designers to put moreeffort into thinking “what can I do with that?” than theywould in a comparable situation in normal industrialpractice, when it would normally be more cost-effec-tive to look for another source that could be adaptedmore easily and directly to create the desired effects.This certainly colored our participants’ reactions to thetapestry. However designing garments with brandedcharacters as motifs, with much tighter constraints thanin our scenario, is a major task for a significant numberof knitwear designers producing children’s clothes.

Creativity Research Journal 365

Adaptation of Sources of Inspiration

Dow

nloa

ded

by [

Flor

ida

Inte

rnat

iona

l Uni

vers

ity]

at 0

8:52

22

Dec

embe

r 20

14

Page 13: Adaptation of Sources of Inspiration in Knitwear Design

The Participants

We piloted our experimental procedure with theauthors themselves and a number of colleagues andfriends, most of whom had no experience whatever ofdesigning knitted garments. In the subjective opinionsof the authors, the designs produced by most of theseamateurs were not noticeably inferior to those of theless inspired professionals and students; however theexperiment did not require or enable the experts to usetheir key skill of knowing how to fit their designs intothe context of contemporary fashion. We comparedtwo groups of experimental participants.

Experienced students. The experiment wasconducted with 11 student designers. All had worked asknitwear designers in one-year industrial placementsbetween the second and final years of their degreecourses. The first three students were taking a postgrad-uate MA in Knitwear Design at Nottingham Trent Uni-versity. Two of these were the subjects who disliked theexperiment, one of whom was the only male partici-pant. The second group comprised five apparel designstudents at De Montfort University in Leicester, whoparticipated in the experiment shortly before the end ofthe final year of their degree course. These participantsvolunteered for the experiment when the first authorapproached their whole class after a lecture. Two fur-ther students at Brighton University were approachedat their degree show at the end of their undergraduatecourse. One participant, from the University of Derby,was a placement student at the first company men-tioned in the following section. The Brighton Univer-sity students and the placement student did the experi-ment with both sources of inspiration.

Professional designers. Nine of the ten profes-sional designers who participated in the experimentworked at the two companies that had been most coop-erative in our observational study of design practice. Inone large supplier to a large and well-known British re-tail chain, six designers took part, including the headdesigner, who had volunteered her colleagues for theexperiment. One of these designers was a freelance de-signer working as maternity cover. These designersworked as a close team, and all had at least three yearsprofessional experience. Three designers at a Scottish

manufacturer of own-label golf wear took part. One ofthese had to be excluded from the analysis reportedhere because she drew several design ideas onto onesketch, making it impossible to reconstruct individualdesign ideas from what she produced; she also ap-peared extremely preoccupied during the experiment.A further participant had previously run her own knit-wear company, and now worked in design manage-ment.

The analysis is thus based on the performance of 11student designers and 9 professional designers.

Data Collection

All but one of the professional designers and five ofthe students were asked to think aloud during the ex-periment to generate concurrent verbal protocols oftheir conscious thoughts during designing. All of theprotocol participants were asked to talk aloud whilecounting the windows in their parents’ house as awarm-up verbalization task. All of the designers weredebriefed after the experiment using the same set ofquestions (listed in the Appendix); the debriefinglasted between 15 min and 1 hr, and involved the par-ticipant talking the experimenter through her designs.The analyses reported in this article were drawn as faras possible from the concurrent verbal protocols, sup-plemented by examination of the participants’sketches and their subsequent descriptions of whatthey did. Comparison of the participants’ debriefingdescriptions of what they had done with their think-aloud protocols produced at the time revealed that thedebriefing descriptions were frequently inaccurate.They appeared to be post-hoc rationalizations.

Using Verbal Protocols to StudyDesign Thinking

Studying problem solving behavior by getting par-ticipants to think aloud and analyzing their verbal pro-tocols—to identify the participants’ goals, foci ofattention, mental representations, and the content oftheir mental actions—is a well-established and muchdiscussed technique in psychology (see Ericsson & Si-mon’s, 1993, seminal methodological account). It hasbeen extensively used to study designing (for instanceAkin, 1987; Goel, 1995; Goldschmidt, 1991, 1994)—

366 Creativity Research Journal

C. M. Eckert and M. K. Stacey

Dow

nloa

ded

by [

Flor

ida

Inte

rnat

iona

l Uni

vers

ity]

at 0

8:52

22

Dec

embe

r 20

14

Page 14: Adaptation of Sources of Inspiration in Knitwear Design

the literature on protocol studies of architects and engi-neers creating designs is far too large to survey here(see Purcell & Gero, 1998, for a survey of protocolstudies of sketching in design). Cross, Christiaans, andDorst (1996) present 20 papers on different aspects ofdesigning by different authors, based on the same ver-bal protocols—the experimental task was to design aluggage rack for a bicycle. The use of verbal protocolsto study visuospatial reasoning, especially designing,is problematic and remains controversial. JonathanSchooler and his colleagues have demonstrated thatverbalization does not disrupt activities that can be ver-balized readily, but can interfere with recall of percep-tual memories and other nonreportable nonverbalmental activities. They attribute this to verbal over-shadowing—verbal processing taking away attentionfrom nonverbal configurable processing (for instanceFallshore & Schooler, 1995; Schooler & Engstler-Schooler, 1990; Schooler, Fiore, & Brandimonte,1997). The activities that are degraded by verbalizationinclude insight problem solving, but not noninsightproblem solving (Schooler & Melcher, 1994; School-er, Ohlsson, & Brooks, 1993). Moreover, trying to ver-balize visual stimuli or mental images has been foundto interfere with visuospatial problem solving in situa-tions where verbal labeling leads to counterproductiveeffects of mental set (Brandimonte & Gerbino, 1993;Brandimonte, Hitch, & Bishop, 1992a, 1992b). Visser(1992), in her discussion of the applicability of proto-col analysis methodology to design, notes thatEricsson and Simon (1980) remark that “producingverbal reports of information directly available inpropositional [i.e. language or verbal] form does notchange the course and structure of the cognitive pro-cesses. However instructions that require subjects torecode information in order to report it may affectthese processes” (p. 235) and “Only information in fo-cal attention can be verbalized” (p. 235).

Davies (1995), discussing an experiment on soft-ware design, and Lloyd, Lawson, and Scott (1995),discussing an experiment designing a bicycle rack,present evidence that verbal protocol experimentsdistort normal design behavior. However, their argu-ments have more force against the use of experimen-tal setups that give the designers tasks embodyingunnatural constraints and requirements than againstthe use of protocol methods. Lloyd et al. point outthat experiments force designers to work to shorterdeadlines than normal and to avoid understanding–

enriching activities like looking at similar products,as well as unrelated displacement activities (enablingincubation).

Our observations of our participants’ behavior inthis experiment, and our subjective experience of do-ing it ourselves, suggest that people verbalize the as-pects of visuospatial problem solving that can easily bedescribed in words. When thinking about objects orother visuospatial aspects of designs that defy easyverbal description, such as the shapes of motifs, peopleremain silent or describe their actions in general terms.The research previously cited by Schooler and his col-leagues (cited previously) on the effects of verbaliza-tion on imagery and insight suggests that the effortinvolved in timesharing between verbalization andnonverbal visuospatial thinking might disrupt the non-verbal thinking, reducing designers’ abilities to recog-nize nonobvious perceptual characteristics of thesource materials and their designs, and perform spatialmanipulations. But we were not able to assess thisfrom our experimental data. Of course, in our experi-mental situation, and in much of commercial knitweardesigning, thinking about visuospatial entities thatcannot be named or easily described is an importantpart of the task. Thus, the protocols generated in the ex-periment were a lot more informative about goals andstrategies than about thinking about the subtleties ofshapes and spatial relationships. As many other experi-menters have found, our subjects differed enormouslyin how easily they could verbalize and how happy theywere doing it. For some, generating a protocol in theexperiment was entirely natural thinking aloud. Othersfound verbalizing unnatural and uncomfortable; onedesigner seized the opportunity to observe the 2 min si-lence at 11 o’clock on Armistice Day as an excuse tostop talking.

Concern for the potential distorting effects on de-sign behavior led us to run the experiment initiallywithout concurrent verbalization. But our experienceswith this experiment matched the usual methodologi-cal findings of verbal protocol research. First, that con-current verbalization provides valid information aboutaspects of design thinking that are otherwise inaccessi-ble; and second, that concurrent protocols show thatretrospective accounts of behavior are often inaccu-rate, and so should be treated with great caution. Anydistorting influence of verbalization, though possiblysignificant, is less important than the insight providedby the protocols.

Creativity Research Journal 367

Adaptation of Sources of Inspiration

Dow

nloa

ded

by [

Flor

ida

Inte

rnat

iona

l Uni

vers

ity]

at 0

8:52

22

Dec

embe

r 20

14

Page 15: Adaptation of Sources of Inspiration in Knitwear Design

Types of Transformation

Transforming an element of a source of inspirationinto an element of a design can range from achieving theclosest possible likeness to free association of ideas.Here we divide transformations of source elements intodifferent classes, which generate design elements thatdiffer more and more from the source. The categoriza-tion of adaptations into these classes is not clear-cut; it issometimes difficult to decide which class a design be-longs to. In the analysis of the experiment, classificationhas been based primarily on the designers’ intentionsandtherelationshipof theoutcometo the initial source.

“Literal” Adaptation

When a new design element is kept as close as pos-sible to the source of inspiration, designers call it a “lit-eral adaptation.” When a motif is designed, a knittingpattern is created with a height and width of a certainnumber of stitches and a small number of colors. Un-less the source is already a grid pattern (for example forembroidery) this involves some simplification. How-ever knitwear designers do not regard as simplifica-tions the adaptations required to create a grid patternwhile staying as close as possible to the original. Lit-eral adaptation of images into grid patterns is well sup-ported by current CAD systems, which allow the userto specify the number of colors and the grid size.

Conscious Simplification

Designers consciously simplify designs based onmore complex sources by either selecting parts of thesource (for example, in Figure 2 the designer has pickedout the horse) or discarding parts of the source (for ex-ample, in Figure 6 the designer has left out parts of thecentraldiamond),whileotherwisekeeping thedesignordesign element as close as possible to the source. Thesimplified design element may play a different role inthe structure of the new design (see Figure 20).

Abstraction

Designers aim to capture some essential propertiesof a relatively complex source while eliminating or

changing others to create a relatively simple designelement. These essential properties might be basic pro-portions, geometric shapes, color schemes, or visualemergent properties such as overall busyness or com-plexity. The objective might be to retain the culturalmeaning of the source. In Figure 7, the overall form ofthe rug has been abstracted and combined with a sim-plification of one element of it to create a motif. In Fig-ure 8, the designer has sought to maintain the roundflowing forms of the tapestry.

368 Creativity Research Journal

C. M. Eckert and M. K. Stacey

Figure 6. Simplified design element.

Figure 7. Abstraction of the source design element.

Figure 8. Abstraction from a visual property of the source.

Dow

nloa

ded

by [

Flor

ida

Inte

rnat

iona

l Uni

vers

ity]

at 0

8:52

22

Dec

embe

r 20

14

Page 16: Adaptation of Sources of Inspiration in Knitwear Design

Modifications to Source

Designers can introduce changes when transform-ing a source of inspiration into a design element. Theymight change details within an overall form copiedfrom the source. For instance, many designers used thecentral diamond form of the rug but changed its inter-nal structure. They may regroup elements derived fromthe source. For instance, some designers have placedhorses within the large diamonds. Designers may alsoamalgamate design elements from different sources;this happens when designers have found several goodsources or are planning to create very complex designs.It is difficult to distinguish modification from abstrac-tion and simplification; we have used the criterion thatthe adaptation should add features beyond what is re-quired for any degree of simplification.

Association

A source can also inspire a design by reminding thedesigner of another object or another design they haveseen. Designers may draw an object or design elementthat is visually similar to the source, though not de-rived directly from it. For instance, a designer mightdraw a different rug from memory or imagination as aresult of seeing the rug used in our experiment. A de-signer might use the source of inspiration to define atopic and create another design fitting into that topic,drawing on memory and imagination or another sourceof inspiration. For example, in creating the designshown in Figure 9, the designer defined “Islam” as a

topic within the context created by the rug and itstheme of “Persia.” Designers often exploit associationswithin the natural context of the source. For instance,in the design shown in Figure 10, the designer drewflowers and leaves to complement those drawn fromthe tapestry. Designers also sometimes develop designelements by evolution away from the original source,where an initial adaptation inspires a further transfor-mation. For instance, one designer picked leaves fromthe tapestry; these reminded her of caterpillars (a pro-cess of “seeing as”; cf. Goldschmidt, 1991; Schön &Wiggins, 1992), and she developed them further intocaterpillars to produce the design shown in Figure 11.

Deviation

Some designers in our experiment moved com-pletely away from the source. Designers can take ideasfrom the environment they are designing in. For exam-ple, one designer took a cable pattern from ClaudiaEckert’s sweater and put it into her tapestry-inspireddesign. Such coincidental inspirations can be quitecommon when designers are actively looking for a

Creativity Research Journal 369

Adaptation of Sources of Inspiration

Figure 9. Association to topic.

Figure 10. Association to the natural context of the source.

Figure 11. Evolution of design elements.

Dow

nloa

ded

by [

Flor

ida

Inte

rnat

iona

l Uni

vers

ity]

at 0

8:52

22

Dec

embe

r 20

14

Page 17: Adaptation of Sources of Inspiration in Knitwear Design

source of inspiration within a context. Designers canalso draw on ideas that they already have in mind. Thismight or might not be related to the source. Designerscreating Arran patterns might start by drawing a classicArran for modification. One designer drew a cardigandesign that she had done previously as a starting pointfor her tapestry-inspired design.

Strategies for Adaptation

The designers in our experiment employed a varietyof different strategies for deriving ideas from sourcesof inspiration. The two major approaches are (a) toform a plan for the design and then use elements of thesource to populate the design with the required designelements and (b) to derive a plan for the design fromthe source. This section outlines the range of differentadaptation strategies that were used. This analysis isbased partly on a priori analysis of the adaptation prob-lem, which gave us these two broad categories, andpartly on the experimental evidence about the use ofstrategies to create designs based on the rug and thetapestry, from which we identified the differentsource-driven strategies. In the next section, we dis-cuss the experimental evidence about which strategieswere used, and by whom, in the experiment.

Evidence About Design Thinkingin Our Data

Design strategies defy conventional methods of ob-servation. Our experimental data provided limited evi-dence about reasoning processes in designing byadaptation. However, we were able to identify adapta-tion strategies as well as the types of adaptations thatwere created. During the experiment, our subjects cre-ated designs very quickly—between two and fifteen inhalf an hour—and worked on individual designs for aslittle as 30 sec. Some designs were sketched withoutany hesitation after the end of the previous sketch. Inconcurrent verbalization, the subjects described theirgoals and their actions in terms of what they weredrawing and what they were looking for, but veryrarely mentioned problem solving tasks or other men-tal activities, so we did not observe complex reasoningabout the designs and how they should progress.

We conclude from our data, other research on de-sign thinking, and our own experiences of trying theexperimental task ourselves, that this is because oursubjects’ design thinking was largely visuospatial, andthus not reportable except in broad terms, and too fastto be verbalized or described in real time. The evalua-tion of partial design sketches has been identified bymany design researchers, especially in architecture, asa major component of many designers’ creative pro-cesses (most famously Schön, 1983; and Goldschmidt,1991, 1995; see Purcell & Gero, 1998). We did not findvery clear evidence for designers interactively re-sponding to what they see in their own sketches,though we have had the subjective experienceourselves. We suspect that this is because designers’perception of features of their designs, and of corre-sponding needs or possibilities for further develop-ments, happens too quickly for clear traces to appearfrequently in protocols, given that shapes and spatialcharacteristics are not easily described in words.

Method of Analysis

The strategies identified in this section were derivedfrom a combination of analyses of the nature of the ad-aptation problem, comments made by our experimen-tal subjects in their debriefing interviews, and analysesof their videotaped verbal protocols and actions duringdesign episodes. The holistic “inspiration-strikes”strategy, which designers describe as an instantaneouscreation of a design idea, is included because designersmention it in interviews. It cannot be tested in experi-ments in which designers are given a source. Unless in-spiration strikes, which we believe is much rarer than isoften assumed, designers confronted with our experi-mental task or a comparable situation can proceed ac-cording to one of these logical possibilities: (a)formulate or remember a plan for the form of the de-sign, and use the source to supply design elements thatfit into that plan; (b) derive a plan for the design fromthe source, and then use the source to supply appropri-ate design elements; or (c) derive the plan for the de-sign from an alternative source of ideas, often aprevious design based on the source or rememberedfrom elsewhere.

The source-driven design strategies are subdividedaccording to the degree of conceptual change betweenthe source and the new design. Of course, many de-

370 Creativity Research Journal

C. M. Eckert and M. K. Stacey

Dow

nloa

ded

by [

Flor

ida

Inte

rnat

iona

l Uni

vers

ity]

at 0

8:52

22

Dec

embe

r 20

14

Page 18: Adaptation of Sources of Inspiration in Knitwear Design

signs are created by a combination of strategies: A de-signer begins with one strategy, evaluates anincomplete design, and uses a different strategy to takeit further. In our statistical analyses, we have associ-ated each design with the first strategy used.

Plan Driven Design

The designer can look at a source needing a designelement to fulfil a given structural role in a designwhen the visual effect of the design element and/or thewhole design has been chosen before any details aredecided on (Figure 12). The designer examines thesource with the question: “What can I get from thesource to create this sort of design element?”

Structural role drives source. In industrial prac-tice, knitwear designers often aim to fit garments intoparticular fashions that dictate the use of particulartypes of motifs arranged in particular ways—for in-stance, the complex horizontal band sweaters of themid-1990s. Here the designer begins with a garmentcategory, which requires motifs to fill predeterminedstructural roles.

During the design phase, designers employing thisstrategy look systematically for sources of inspirationthat supply design elements that fit the requirementsimposed by the predefined structure, theme, and de-sired characteristics. In the experiment, designers of-ten decided on categories of designs such as “overallborder pattern,” implicitly specifying an outline planand a set of structural roles for design elements; andlooked for components of the source to fill needs, thatis, for design elements that fit into the vacant slots inthe structure. For example, the designer of Figure 13wanted to create an “overall multi-colored pattern.”Different types of adaptation are used depending onthe visual effect the designer wants to achieve.

Emergent need drives source. In the course ofdesigning, the designer can perceive the need for a de-sign element to serve a particular function, to comple-ment existing design elements to achieve a particularoverall visual effect—this is an example of what Schönand Wiggins (1992) term appreciating the characteris-tics of a design (Figure 14). In our experiment, design-ers frequently adopted this strategy when refining in-complete designs. As this is, by definition, a strategyadopted to extend a partial design, and our statisticalanalyses consider only the first strategy used, we have-n’t included it in our measures of how frequently differ-ent strategies occurred in our experiment.

Source Driven Design

A source of inspiration can trigger the develop-ment of a design, which might be based closely on itor distantly derived from it. The development of thedesign is typically a sequential process involving theapplication of a strategy. Here we outline the qualita-tively distinct source-driven strategies that we haveidentified in our experiment.

Holistic idea. In interviews, designers oftencomment that they come across a source of inspira-

Creativity Research Journal 371

Adaptation of Sources of Inspiration

Figure 12. Structural role drives source strategy.

Figure 13. Structural role drives source design.

Dow

nloa

ded

by [

Flor

ida

Inte

rnat

iona

l Uni

vers

ity]

at 0

8:52

22

Dec

embe

r 20

14

Page 19: Adaptation of Sources of Inspiration in Knitwear Design

tion by chance and are instantly struck by an idea fora design (see Figure 15). For example, one designercommented to us that she saw a piece of tree bark andimmediately translated it mentally into a knitted fab-ric. Many designers have explained to us that duringdesign research they look at many different stimuliand often translate these immediately into designs,many of which are discarded just as quickly.

Our analysis of our experimental data did not revealany holistic idea designs—all the designs appeared tobe incrementally developed from ideas for partial de-signs. But in our experimental setup, it is impossible toclearly distinguish the instantaneous generation of ho-listic ideas from other source-driven strategies. (“A-ha!” statements describing a complete idea before anysketching would be suggestive evidence, but we can-not safely draw the reverse inference from the absenceof such clear statements.) However, it is remarkablethat many designers began drawing almost as soon asthey saw the source of inspiration and did not hesitateduring the sketching of one design idea.

Salience priority. Several designers in our ex-periment began creating designs by placing a direct ad-aptation of the most visually salient feature of the sourceonto the garment design in the most similar availablestructural role. Only then did they appear to reflect onwhat else the design might need. A design created in thisway could be sufficient as it is, be extended using theemergent need drives source strategy, or be extended byrecalling a garment category that fits the initial design el-ement and applying the structural role drives sourcestrategy. For example, the three center diamonds areright in the middle of the rug; several designers placed itin the center of their sweater designs. Some designers did

this in a first design as a starting point for other designs,others did it when they had (temporarily) run out ofideas, whereas others commented that they wanted toavoid doing this because it would be too obvious.

Here again, we encountered serious difficulties withclassifying the designs produced in the experiment, asthe salience priority strategy can be viewed as a specialcase of either of the two following strategies. Either thesalient feature is used in the new design in the samestructural role as it takes in the original source (as in thesweater design shown in Figure 16; see Figure 17), or itis transformed into a design element taking a differentrole. We classified as salience priority designs that ap-peared to be created by the immediate translation of avery salient source element either without reflection orexplicitly as a starting point.

Source drives same structural role. Designersapplying this strategy position features of the source inthe same relationships to the overall form, or other fea-tures of the new design, as they have in the source (seeFigure 18). They have the implicit goal to use parts ofthe source directly without active alteration of their as-sociated structural roles, but with explicit considerationof what’s there and how to use it. This strategy is usedeither to create a new design composed of selected ele-ments of the original source or to supply design ele-ments that complement other parts of the new design

372 Creativity Research Journal

C. M. Eckert and M. K. Stacey

Figure 14. Emergent need drives source strategy.

Figure 15. Holistic idea “strategy.”

Figure 16. Salience priority design.

Figure 17. Salience priority strategy.

Dow

nloa

ded

by [

Flor

ida

Inte

rnat

iona

l Uni

vers

ity]

at 0

8:52

22

Dec

embe

r 20

14

Page 20: Adaptation of Sources of Inspiration in Knitwear Design

that have already been created. For example, in the de-sign shown in Figure 19, a border in the rug is adaptedinto a border in the sweater.

Source drives different structural role. Designersemploying this strategy adapt features in the source tofill structural roles in the new design that are differentfrom the roles they play in their original contexts (seeFigure 20). They have the implicit goal to adapt parts ofthe source by actively considering alternative relation-ships between these parts and other features of the de-sign or the overall form of the design. This may bemotivated by dissatisfaction with the aesthetic effectachieved by a feature in its existing role. However knit-wear designers have only a limited set of ways in whichthey can use motifs (see Table 1), and imagining thefeature in each possible role is a natural activity forthem. In Figure 2, the horse is used as a large single mo-tif, but in Figure 21 it appears as a repeated motif in aborder across the chest.

Analysis of source. Designers employing thisstrategy identify relatively abstract emergent charac-teristics of their sources, and construct traditional pat-

terns embodying those characteristics (see Figure 22).For example, the designer who created the sweater de-sign shown in Figure 8 picked up on the tonal richnessof the tapestry and the flowing layered effect of the ele-ments within it, and translated this into a rounded cablepattern on a background of multicolored wool. Otherdesigners in our experiment reacted to more generalproperties of the sources, for example seeing the rug asrich or the tapestry as flowery, and tried to capture thisin a design, for instance in Figure 10. In these cases, thedesigners fall back on traditional patterns or designsthat they have seen as sources of structures to adapt toimplement the emergent characteristics that they havetaken from the source.

Designing From Other Sources

Designers both in real life and in our experimentsometimes base their designs on sources of ideas otherthan the visual stimuli they are considering. These maybe associations with the theme or recollections of de-signs that provide a starting point for design by modifi-cation, adapting the structural characteristics of thefoundation design to include features from the source,or achieve its emergent aesthetic properties.

Design evolution. Designers often go from onedesign to the next without referring back to the source.They are using their own sketches as sources of inspi-ration for subsequent design ideas that are evolvedfrom their initial ideas rather than to the source (seeFigure 23). This is a phenomenon that has frequentlybeen observed in studies of architectural design(Schön, 1983; Schön & Wiggins, 1992; see Purcell &Gero, 1998).

Creativity Research Journal 373

Adaptation of Sources of Inspiration

Figure 18. Source drives same structural role strategy.

Figure 19. Source drives same structural role design.

Figure 20. Source drives different structural role strategy. Figure 21. Source drives different structural role design.

Dow

nloa

ded

by [

Flor

ida

Inte

rnat

iona

l Uni

vers

ity]

at 0

8:52

22

Dec

embe

r 20

14

Page 21: Adaptation of Sources of Inspiration in Knitwear Design

Legacy design. In our experiment, a few designswere created that had no obvious connection to thesource provided in the experiment. In their debriefinginterviews, the designers commented that they hadthese designs in mind from previous designs of theirown or from designs they had seen elsewhere (see Fig-ure 24)—a somewhat different situation from the ratio-nal adoption of a plan to fit into a particular fashionspace. We view these designs as evidence for fixationon previous designs (sources of inspiration outside thebrief given in the experiment) influencing what oursubjects created. (See Jansson & Smith, 1991, andPurcell & Gero, 1996, for experimental studies of fixa-tion in product design.)

No direct relationship with the source. In theexperiment, some designers created designs without

making any discernible reference to the source, or toany obvious alternative source of ideas. Some very ba-sic designs such as simple stripes were derived from thecategories of the designs the designers wanted to pro-duce without reference to any source, as the form of thedesign elements are determined by and implicit in thegeneral form of the designs.

Combinations of Strategies

Simple designs such as that shown in Figure 19 canbe created by the application of a single strategy. Butmany more complex designs, such as that shown inFigure 2, are created through the application of a com-bination of strategies. In the latter case, the designerwas initially attracted by the horse. (She had picked itout for previous designs but not used it.) She began thenew design by sketching the big horse, applying the sa-lience priority strategy. She then looked at her horseand searched the source for a suitable border pattern,applying the structural role drives source strategy. Shepicked out the step pattern, and looked again andpicked out the diagonal boxes.

Designers in our experiment who used combina-tions of strategies began with either plan or a source-driven strategy, and started sketching. They looked atthe sketch and thought about what might be missing.They then either formed a plan and looked for an ele-ment of the source with which to realize the plan, ordrew an idea from elsewhere. Sometimes they derivednew ideas directly from the sketch.

374 Creativity Research Journal

C. M. Eckert and M. K. Stacey

Figure 22. Analysis of source strategy.

Figure 23. Evolution of source strategy.

Figure 24. Legacy design strategy.

Dow

nloa

ded

by [

Flor

ida

Inte

rnat

iona

l Uni

vers

ity]

at 0

8:52

22

Dec

embe

r 20

14

Page 22: Adaptation of Sources of Inspiration in Knitwear Design

Patterns in Strategy Application

The study was designed to identify the strategiesused by designers, not to provide sufficient quantita-tive data to support strong inferences about relativefrequencies. The difficulty of securely differentiatingbetween categories of strategies means that our numer-ical results should be treated with some caution, but,nonetheless, some interesting patterns emerge fromstatistical analysis.

All the strategies previously described have beenobserved in the rug and tapestry experiment, with theexception of holistic creation. In this section we exam-ine (a) differences between the task conditions—therug is both more typical and more tractable as a sourceof inspiration than the tapestry, and this appears to in-fluence the strategies used to adapt it; (b) the differ-ences between the student designers and theprofessionals, who appear to make more use of plan-driven strategies; and (c) the relationship between theadaptation strategies used and the types of transforma-tions that are produced.

Method of Classification

For the purpose of this analysis, the designs havebeen classified according to the strategy applied for themost salient feature of the design. This was invariablythe first pattern feature drawn by the designers, thoughsome designers changed the given shape before theybegan to draw the pattern. A few designs that had noclear relationship to the source were excluded from theanalysis. Claudia Eckert identified the initial strategyby (a) design intentions stated in verbal protocols, forexample “I will design an overall stripe pattern” wasclassified as structural role drives source, and “I likethe horse” followed by immediate drawing of the horsewas classified as salience priority; (b) comments madein the debriefing in answer to the question, “How didyou come up with this design?” and (c) looking at thedesign itself.

It is extremely difficult to assign a design securelyto one category. This is partly because some categoryboundaries are fuzzy, for instance between saliencepriority and source drives same structural role, and be-tween salience priority and source drives differentstructural role. It is also impossible to ensure equal

treatment of designs produced with and without con-current verbalization.

In the following analysis, we discuss the results interms of how much reasoning about the form of the de-sign different strategies require. We group structuralrole drives source with analysis of source as morecomplex strategies involving more reflection abouthow to use the source to achieve design objectives; andgroup the source-driven strategies (salience priority,source drives same structural role, and source drivesdifferent structural role) as less complex strategies. Itwould be a mistake to regard the former as “better” or“more sophisticated.” In real life, many simple designsinvolve a lot of thought and effort in deciding exactlywhat is required by fashion. If designers know whatthey want, it pays them to search for a source that theycan adapt directly, employing the salience prioritystrategy or the source drives same structural role strat-egy, because this minimizes the effort involved inadapting the source.

Differences Between Sources

A total of 123 designs were created by 20 partici-pants. The number of designs per designer ranged from2 (using one source) to 21 (using both sources). Therug is a typical source of inspiration for knitwear de-sign. Its rich geometric patterns are easy to adapt andcan be used in many different themes. The tapestry wasselected as being equally rich, but its organic roundshapes are very difficult to adapt into knitwear, and theflowery motifs were seen by many designers as beingtoo feminine for menswear. Table 2 shows the numberand relative frequency of the strategies employed bythe professionals and students using the rug. Table 3shows the same information for the tapestry condition.

As Figure 25 suggests, the designers in our experi-ment employed adaptation strategies with differentfrequencies in the rug and tapestry tasks. The signifi-cance of this observed difference was tested by group-ing the strategies used by the professional designers asshown in Table 4. With the strategies grouped intomore complex and reflective (structural role drivessource and analysis of source) and simpler and moresource-driven (salience priority, source drives samestructural role and source drives different structuralrole), as shown in Table 4, a 2 × 2 χ2 test gives p ≤ 0.02.With structural role drives source and analysis of

Creativity Research Journal 375

Adaptation of Sources of Inspiration

Dow

nloa

ded

by [

Flor

ida

Inte

rnat

iona

l Uni

vers

ity]

at 0

8:52

22

Dec

embe

r 20

14

Page 23: Adaptation of Sources of Inspiration in Knitwear Design

source treated as separate groups, a 3 × 2 χ2 test gives p≤ 0.01. In both cases, the more conservative Yates cor-rection to the χ2 analysis gives p ≤ 0.05. (There werenot enough student tapestry designs for the same com-parison to be made for the student subjects.)

These results indicate that designers use source-driven strategies more when they like the source and

find it easy to use, and that they use more reflectivegoal-driven strategies when the source proves rela-tively intractable. This is what one would intuitivelyexpect. The evidence of concurrent verbal protocolsand debriefings indicates that when designers have dif-ficulties with the source and feel that they cannot reuseits components directly (the flowers and leaves of the

376 Creativity Research Journal

C. M. Eckert and M. K. Stacey

Table 3. Number and Frequency of Strategies Used with the Tapestry Source

7 Professionals 3 Students 10 Total

Tapestry Task N % N % N %

Role → source 10 40 1 10 11 31Salience priority 1 4 2 20 3 9Source → same role 1 4 0 0 1 3Source → different role 6 24 4 40 10 29Analysis of source 6 24 1 10 7 20Evolution 0 0 1 10 1 3Legacy 1 4 1 10 2 6Total of designs 25 100 10 100 35 100

Figure 25. Professionals: Strategy frequencies in rug and tapestry tasks.

Table 2. Number and Frequency of Strategies Used with the Rug Source

9 Professionals 10 Students 19 Total

Rug Task N % N % N %

Role → source 13 31 6 13 19 22Salience priority 6 14 7 15 13 15Source → same role 6 14 11 24 17 19Source → different role 12 29 13 28 25 28Analysis of source 1 2 1 2 2 2Evolution 4 10 8 17 12 14Legacy 0 0 0 0 0 0Total of designs 42 100 46 100 88 100

Dow

nloa

ded

by [

Flor

ida

Inte

rnat

iona

l Uni

vers

ity]

at 0

8:52

22

Dec

embe

r 20

14

Page 24: Adaptation of Sources of Inspiration in Knitwear Design

tapestry were seen by many as too feminine, and theyhad shapes that are difficult to translate into grid pat-terns except at a very large scale), they either form aplan within which they can employ elements of thesource or try to capture the characteristics of the sourcein another way.

Differences Between Professionalsand Students

Even though all the student participants had at leastone year of industrial experience, the numerical datashown in Table 5 and illustrated in Figure 26 suggests amarked difference between the students and the pro-fessionals in how frequently they employed the differ-ent strategies.

Table 5 shows the number of times the more com-plex strategies (structural role drives source and anal-ysis of source) and the less complex strategies wereused by professional and student designers, aggregat-ing across the rug and tapestry tasks. A 2 × 2 χ2 testgives p ≤ 0.02, with a Yates adjustment of p ≤ 0.05. A 3× 2 χ2 test separating structural role drives source andanalysis of source gives p ≤ 0.01, with a Yates adjust-

ment of p ≤ 0.02. However, the validity of results ag-gregating over the two tasks is limited by the con-flation of differences in strategies employed by thestudents and the professionals, with the difference inthe total number of designs generated in each conditionby the students and the professionals. For the rug alone(Table 6), a 2 × 2 χ2 test gives p ≤ 0.05, with a Yates ad-justment that is not significant at the p ≤ 0.05 level. The3 × 2 χ2 test is not valid because both groups producedonly one design in the analysis of source category.Only 3 students did the tapestry task, so there were in-sufficient tapestry-based student designs for a valid χ2

test comparing professional and student tapestry-baseddesigns.

These results indicate that the student designersstuck more closely to the source, while the professionaldesigners were more likely to use more elaborate goal-directed strategies. They support the conclusion thatexperienced professional designers are more skilled atbypassing the obvious ways to adapt source materialsand at finding ways to use source materials that do notsuggest designs. Although this can be attributed partlyto the professional designers having more experience,we interpret this as being a consequence of the profes-sionals having developed different skills and habits to

Creativity Research Journal 377

Adaptation of Sources of Inspiration

Table 4. Professionals: Occurrences of Strategies in the Rug and Tapestry Tasks

ProfessionalsRole Drives

Source Analysis

Salience Priority +Source Drives Same

Structural Role +Source Drives Different

Structural Role

Tapestry7 participants 10 6 8

Rug9 participants 13 1 24

Table 5. Occurrences of Strategies Used by Professionals and Students (Excluding Evolution Designs)

Total of Designs Role Drives Source Analysis

Salience Priority +Source Drives Same

Structural Role +Sources Drives Different

Structural Role

Professionals9 participants 23 7 32

Students11 participants 7 2 37

Dow

nloa

ded

by [

Flor

ida

Inte

rnat

iona

l Uni

vers

ity]

at 0

8:52

22

Dec

embe

r 20

14

Page 25: Adaptation of Sources of Inspiration in Knitwear Design

cope with different task demands (see Eckert, Stacey,& Wiley, 1999; Stacey, Eckert, & Wiley, 2002).Throughout their training, British-educated design stu-dents do fairly free projects, within which they are freeto select their own sources of inspiration and interpretbriefs how they wish. Their training is influenced bythe belief, widespread among British design educatorsas well as in the knitwear industry, that creativity isharmed by tight constraints (Eckert & Stacey, 1994).This has the consequence that students gain no experi-ence in designing within tight constraints. Workingprofessional designers learn to design within muchtighter constraints imposed by customers’ briefs andtime pressure. Using the structural role drives sourcestrategy to create designs with a particular predeter-mined form is a normal part of our professional sub-jects’ working lives; they may also have a richer sourceof plans and more strongly learned procedures for us-ing them, and thus be more prone to fixation. More-over, many professional designers are restricted intheir access to source material and have to make dowith what they have, which they experience as a signif-icant limitation (Eckert & Stacey, 1998).

Relationships Between Strategiesand Transformation Types

The designs produced in the experiment wereanalyzed to test the hypothesis that different adap-tation strategies used by the designers in our ex-periment were associated with different types oftransformations.

The designs generated in the experiment wereclassified according to the adaptation strategies theirdesigners used to create them and the types of trans-formations they embodied. Table 7 shows the numberof designs for each combination of adaptation strat-egy and transformation type. The most complex andgoal-directed strategies, structural role drives sourceand analysis of source, frequently led to the mostradical and indirect type of source transformation, as-sociation. The simpler and more direct adaptationstrategies, salience priority, source drives samestructural role and source drives different structuralrole, led to a much higher proportion of straightfor-ward, close transformations. Although there are dif-ferences in frequencies, most of the source-relateddesign strategies led to designs employing most ofthe different types of transformation.

As classifying designs is difficult, and to some extentsubjective, this numerical data should be treated withsome caution; statistical analysis nonetheless supportsthe view that different transformation types are associ-ated with particular strategies. A χ2 analysis of the datainTable7givesp≤0.001,withaYatesadjustmentofp≤0.02. A χ2 analysis of the data for the source-relatedstrategies and transformations (the inner box in Table 7)gives p ≤ 0.001, with a Yates adjustment of p ≤ 0.001.Table 8 shows the number of designs produced with di-rectversuscomplexandplan-drivenstrategies,andwithclose versus complex transformations. A χ2 analysis of

378 Creativity Research Journal

C. M. Eckert and M. K. Stacey

Figure 26. Frequencies of strategies used by professionals andstudents.

Table 6. Occurrences of Strategies Used by Professionals and Students in the Rug Task (ExcludingEvolution Designs)

Total of Rug Designs Role Drives Source Analysis

Salience Priority + SourceDrives Same StructuralRole + Sources Drives

Different Structural Role

Professionals9 participants 13 1 24

Students10 participants 6 1 31

Dow

nloa

ded

by [

Flor

ida

Inte

rnat

iona

l Uni

vers

ity]

at 0

8:52

22

Dec

embe

r 20

14

Page 26: Adaptation of Sources of Inspiration in Knitwear Design

thedata inTable8givesp≤0.02,withaYatescorrectionof p ≤ 0.02.

These results support the intuitive hypothesis thatdesigners are more likely to use elements of the sourcedirectly when applying strategies that are driven by thevisual elements of the source. When designers choosemore complex and less direct strategies, they are morelikely to change the source elements significantly or beguided by associations. Designers choose less directstrategies either when the source material appears un-suitable for direct adaptation—because its elements donot lend themselves to satisfactory close transforma-tions—or because the close transformations they af-ford are rejected as too easy or obvious.

Conclusions

Creating a new design by transforming one or moresources of inspiration is a ubiquitous process in knit-wear design. It has clear parallels to designing by mod-ifying previous designs and other sources of ideas inmany other fields, such as fashion design, architecture,

and engineering (Eckert et al., 2000). This article re-ports a study of designing by transforming a source.Nine professional knitwear designers and 11 experi-enced knitwear design students were given a brief todesign one or more men’s t-sleeve sweaters based on agiven source of inspiration and a picture of a Persianrug or a 19th century tapestry. They were videotaped,in most cases were asked to produce concurrent verbalprotocols, and were asked about their design thinkingin subsequent debriefing. Creating designs for knittedgarments by adapting such a source of inspiration is acommon activity familiar to the participants in the ex-periment, who were also accustomed to designing to abrief specifying particular needs. However, in indus-trial practice designers can normally choose their ownsources; they usually only select those that yield designelements that meet their needs through simple and di-rect adaptations, whereas in this experiment they werepushed harder than usual to think about what theycould do with sources that did not immediately suggestdesigns.

Analysis of the videotape data and the designsthemselves shows that both the designers’ strategiesfor creating a design by adapting a source and the de-signs themselves can be described in terms of clearlyidentifiable patterns.

A source, or one element of a source, can be trans-formed in one of the following ways. (Though somedesigns produced in our experiment bore no relation tothe given source that either Claudia Eckert or the de-signers themselves could identify.)

• Literal adaptation. The source element is trans-lated into a component of a knitwear design asaccurately as the medium will allow (that is,

Creativity Research Journal 379

Adaptation of Sources of Inspiration

Table 7. Relationship Between Adaptation Strategies and Types of Transformation

TransformationStrategy Literal Simp. Abstr. Mod. Assoc. Devi. Total

Role → source 3 5 6 6 9 0 29Salience priority 4 6 7 0 0 0 17Source → same role 6 5 0 5 0 2 18Source → different role 8 13 1 7 4 2 35Analysis of source 0 2 2 0 5 0 9Evolution 1 0 1 1 9 1 13Legacy 0 0 0 0 2 1 3Total 22 31 17 19 29 6 124

Note: Simp. = conscious simplification, Abstr. = abstraction, Mod. = modifications to source, Assoc. = association, Devi. = deviation.

Table 8. More and Less Direct Strategies and Transforma-tions

Literal +Simplification+ Abstraction

Association +Modification

Plan drives sources +analysis

18 20

Salience priority + 50 16Source same structural role+ Sources differentstructural role

Dow

nloa

ded

by [

Flor

ida

Inte

rnat

iona

l Uni

vers

ity]

at 0

8:52

22

Dec

embe

r 20

14

Page 27: Adaptation of Sources of Inspiration in Knitwear Design

with a limited number of stitches and a fewcolors).

• Simplification. The source is consciously sim-plified in the process of translating it into an el-ement of a knitwear design.

• Abstraction. Some but not all of the visualproperties of the source are used to create thecorresponding element of a knitwear design.

• Modification. The source element is altered tocreate an element of a knitwear design that ismore suitable for the medium of knitwear or tothe context of the design, but the design is stillclearly visually related to the source.

• Association. The source element supplies vi-sual properties, emergent features, or culturalcontexts that are used indirectly to create a newdesign.

Transformed design elements can be used with thesame relationship to other elements and the overall de-sign as their originals, or they can be used in differentcombinations and take different structural roles in theoverall form.

When adapting a source of inspiration into a newdesign, designers employ distinct strategies. These canbe described most succinctly in terms of motivatingquestions (implicit in the designers’ actions, ratherthan consciously articulated).

• Structural role drives source. “How can I cre-ate that visual effect that I want with thissource?”

• Salience priority. “This is the most striking fea-ture of this source. How can I make use of it?”

• Source drives same structural role. “How can Iapply this feature of the source in the same wayin my design as it appears in the source?”

• Source drives different structural role. “Whatelse can I do with this feature of the source?”

• Analysis of source. “How else can I create a de-sign that captures the overall visual appearanceof the source?” (Or a particular aspect of the vi-sual appearance of the source.)

• Design evolution. “How can I develop this ideafurther?”

In the experiment, designers also created designsthat were unconnected to the given source by sketchingideas that they already had in their minds. During inter-

views and observations of designers working in indus-try, as well as in the debriefings at the end of our exper-iment, designers often comment that they aresometimes struck by an idea as soon as they see asource. When this happens they usually experience thissubjectively as an entire design appearing fully formedin their minds. Although this is very memorable, it isquite infrequent, so it is unsurprising that it never oc-curred in our experiment.

The study was designed to generate data fromwhich the strategies employed by designers could beidentified and analyzed, not to generate enough de-signs for frequencies of strategies and transformationsto be measured and compared. Nevertheless, someinteresting, if unsurprising, statistically significant pat-terns appeared. The adaptation strategies were classi-fied as direct source-driven strategies (saliencepriority, source drives same structural role, sourcedrives different structural role), or as indirect goal-driven strategies (structural role drives source, analy-sis of source). Statistical tests support the following in-tuitive hypotheses.

• When the source of inspiration is well suited toadaptation into knitwear design elements andwell suited to the brief, and the designers like it,they are more likely to be guided by the visualproperties of the source. When the source is lesssuitable, the more likely designers are to useless direct goal-driven strategies.

• Students are less likely than professional de-signers to use the less direct goal-driven strate-gies. We attribute this to the professionals hav-ing more experience of designing with tightersets of constraints, in particular working withbriefs and sources they did not choose and donot like, as well as having more experience ofdoing plan-driven rather than inspiration-drivendesign.

• Although any adaptation strategy can lead toany type of transformation, direct source-drivenadaptation strategies lead to a higher proportionof close translations, while the less direct goal-driven adaptation strategies lead to a higher pro-portion of radical modifications and abstracttransformations.

This study supports the view that designing andother creative behavior can be analyzed in terms of

380 Creativity Research Journal

C. M. Eckert and M. K. Stacey

Dow

nloa

ded

by [

Flor

ida

Inte

rnat

iona

l Uni

vers

ity]

at 0

8:52

22

Dec

embe

r 20

14

Page 28: Adaptation of Sources of Inspiration in Knitwear Design

consistent patterns that result both from the intrinsicdemands of the creative tasks and the cognitive charac-teristics and learned skills of designers. Both the exper-iment reported in this article and our observations ofthe knitwear industry indicate that much, if not all, ofknitwear design is the rational application of identifi-able strategies both for formulating design problemsand creating visual and mental contexts (Eckert &Stacey, 2001), and for generating appropriate designs.

References

Akin, Ö. (1978). How do architects design? In J.-C. Latombe (Ed.),Artificial intelligence and pattern recognition in computer-aided design (pp. 65–104). New York: North-Holland.

Akin, Ö. (1987). The psychology of architectural design. London:Pion.

Bartlett, F. C. (1934). Remembering. Cambridge, UK: CambridgeUniversity Press.

Brandimonte, M., & Gerbino, W. (1993). Mental image reversal andverbal recoding: When ducks become rabbits. Memory andCognition, 21, 23–33.

Brandimonte, M., Hitch, G. J., & Bishop, D. (1992a). Influence ofshort-term memory codes on visual image processing: Evidencefrom image transformation tasks. Journal of Experimental Psy-chology: Learning, Memory and Cognition, 18, 157–165.

Brandimonte, M., Hitch, G. J., & Bishop, D. (1992b). Verbalrecoding of visual stimuli impairs mental image transforma-tions. Memory and Cognition, 20, 449–455.

Broadbent, G. (1973). Designing in architecture. London: JohnWiley.

Cross, N. G., Christiaans, H. H. C. M., & Dorst, K. (Eds.). (1996).Analysing design activity. Chichester, UK: John Wiley.

Darke, J. (1979). The primary generator and the design process. De-sign Studies, 1, 36–44.

Davies, S. P. (1995). Effects of concurrent verbalisation on designproblem solving. Design Studies, 16, 102–116.

Eckert, C. M. (1997a). Design inspiration and design performance.Proceedings of the 78th World Conference of the Textile Insti-tute (Vol. 1, pp. 369–387). Thessaloniki, Greece: TextileInstitute.

Eckert, C. M. (1997b). Intelligent support for knitwear design. Un-published doctoral thesis, Department of Design and Innova-tion, The Open University, Milton Keynes, UK.

Eckert, C. M. (1999). Managing effective communication in knit-wear design. Design Journal, 2(3), 29–41.

Eckert, C. M. (2001). The communication bottleneck in knitwear de-sign: Analysis and computing solutions. Computer SupportedCooperative Work, 10, 29–74.

Eckert, C. M., Cross, N. G., & Johnson, J. H. (2000). Intelligent sup-port for communication in design teams: Garment shape speci-fications in the knitwear industry. Design Studies, 21, 99–112.

Eckert, C. M., & Demaid, A. (1997). Concurrent design. Proceed-ings of the 78th World Conference of the Textile Institute (Vol.3, pp. 101–122). Thessaloniki, Greece: Textile Institute.

Eckert, C. M., & Demaid, A. (2001). Classifying design and designmanagement in seasonal industries. International Journal of In-novation Management, 5, 401–425.

Eckert, C. M., & Stacey, M. K. (1994). CAD systems and the divi-sion of labour in knitwear design. In A. Adams, J. Emms, E.Green, & J. Owen (Eds.), Women, Work and Computerization:Breaking Old Boundaries—Building New Forms (pp. 409–422). Amsterdam: North-Holland.

Eckert, C. M., & Stacey, M. K. (1998). Fortune favours only the pre-pared mind: Sources of inspiration are essential for continuingcreativity. Creativity and Innovation Management, 7, 9–16.

Eckert, C. M., & Stacey, M. K. (2000). Sources of inspiration: A lan-guage of design. Design Studies, 21, 523–538.

Eckert, C. M., & Stacey, M. K. (2001). Designing in the context offashion—Designing the fashion context. In P. Lloyd & H. H. C.M. Christiaans (Eds.) Designing in Context: Proceedings of the5th Design Thinking Research Symposium (pp. 113–129).Delft, Netherlands: Delft University Press.

Eckert, C. M., & Stacey, M. K. (2003). Sources of inspiration in in-dustrial practice: The case of knitwear design. Journal of De-sign Research, 2. Available at http://jdr.tudelft.nl

Eckert, C. M., Stacey, M. K., & Clarkson, P. J. (2000). Algorithmsand inspiration: Creative reuse of design experience. Proceed-ings of the Greenwich 2000 Symposium: Digital Creativity (pp.1–10). London: University of Greenwich.

Eckert, C. M., Stacey, M. K., & Wiley, J. (1999). Expertise and de-signer burnout. In U. Lindemann, H. Birkhofer, H.Meerkamm, & S. Vajna (Eds.), Proceedings of the 12th Inter-national Conference on Engineering Design (Vol. 1, pp. 195–200). Munich, Germany: Technical University of Munich.

Egan, D. E., & Schwartz, B. J. (1979). Chunking in recall of sym-bolic drawings. Memory and Cognition, 7, 149–158.

Ericsson, K. A., & Simon, H. A. (1980). Verbal reports as data. Psy-chological Review, 87, 215–251.

Ericsson, K. A., & Simon, H. A. (1993). Protocol analysis: Verbalreports as data (Rev. ed.). Cambridge, MA: MIT Press.

Fallshore, M., & Schooler, J. W. (1995). The verbal vulnerability ofperceptual expertise. Journal of Experimental Psychology:Learning, Memory and Cognition, 21, 1608–1623.

Finger, S. (1998). Design reuse and design research [Keynote pa-per]. In S. Sivaloganathan & T. M. M. Shahin (Eds.), Design re-use: Engineering design conference ’98 (pp. 3–10). Bury StEdmonds, UK: Professional Engineering Publishing.

Gero, J. S. (1990). Design prototypes: A knowledge representationschema for design. AI Magazine, 11(4), 26–36.

Gobet, F., & Simon, H. A. (1998). Expert chess memory: Revisitingthe chunking hypothesis. Memory, 6, 225–255.

Goel, V. (1995). Sketches of thought, Cambridge, MA: MIT Press.Goldschmidt, G. (1991). The dialectics of sketching. Creativity Re-

search Journal, 4, 123–143.Goldschmidt, G. (1992). Serial sketching: Visual problem solving in

designing. Cybernetics and Systems: An International Journal,2, 191–219.

Goldschmidt, G. (1994). On visual design thinking: The vis kids ofarchitecture. Design Studies, 15, 158–174.

Goldschmidt, G. (1995) Visual displays for design: Imagery, anal-ogy and databases of visual images. In A. Koutamanis, H.Timmermans & I. Vermeulen (Eds.), Visual databases in archi-

Creativity Research Journal 381

Adaptation of Sources of Inspiration

Dow

nloa

ded

by [

Flor

ida

Inte

rnat

iona

l Uni

vers

ity]

at 0

8:52

22

Dec

embe

r 20

14

Page 29: Adaptation of Sources of Inspiration in Knitwear Design

tecture: Recent advances in design and decision-making (pp.53–76). Aldershot, UK: Avebury Press.

Goldschmidt, G. (1998). Creative architectural design: Referenceversus precedence. Journal of Architectural and Planning Re-search, 15, 258–270.

Goldstone, R. L. (1998). Perceptual learning. Annual Review of Psy-chology, 49, 585–612.

Jansson, D. G., & Smith, S. M. (1991). Design fixation. DesignStudies, 12, 3–11.

Koriat, A., Goldsmith, M., & Pansky, A. (2000). Towards a psychol-ogy of memory accuracy. Annual Review of Psychology, 51,481–537.

Kosslyn, S. M. (1980). Image and mind. Cambridge, MA: HarvardUniversity Press.

Kosslyn, S. M. (1994). Image and brain. Cambridge, MA: MITPress.

Lloyd, P., Lawson, B., & Scott, P. (1995). Can concurrent verbaliza-tion reveal design cognition? Design Studies, 16, 237–259.

Logie, R. H. (1995). Visuo-spatial working memory. Hove, UK: Psy-chology Press.

Myles-Worsley, M., Johnson, W. A., & Simons M. A. (1988). Theinfluence of expertise on X-ray image processing. Journal ofExperimental Psychology: Learning, Memory and Cognition,14, 553–557.

Otto, K. N., & Wood, K. L. (2000). Product design. New York:Prentice-Hall.

Oxman, R. (1990). Prior knowledge in design: A dynamic knowl-edge-based model of design and creativity. Design Studies, 11,17–28.

Purcell, A. T., & Gero, J. S. (1996). Design and other types of fixa-tion. Design Studies, 17, 363–383.

Purcell, A. T., & Gero, J. S. (1998). Drawings and the design pro-cess. Design Studies, 19, 389–430.

Schank, R. C. (1982). Dynamic memory: A theory of reminding andlearning in computers and people. Cambridge, UK: CambridgeUniversity Press.

Schank, R. C., & Abelson, R. (1977). Scripts, plans, goals and un-derstanding. Hillsdale, NJ: Lawrence Erlbaum Associates, Inc.

Schön, D. A. (1983). The reflective practitioner: How professionalsthink in action. New York: Basic Books.

Schön, D. A. (1988). Designing: Rules, types and worlds, DesignStudies, 9, 181–190.

Schön, D. A., & Wiggins, G. (1992). Kinds of seeing and their func-tion in designing. Design Studies, 13, 135–156.

Schooler, J. W., & Engstler-Schooler, T. Y. (1990). Verbal over-shadowing of visual memories: Some things are better left un-said. Cognitive Psychology, 22, 36–71.

Schooler, J. W., Fiore, S. M., & Brandimonte, M. A. (1997). At a lossfrom words: Verbal overshadowing of perceptual memories. InD. L. Medin (Ed.), The psychology of learning and motivation(pp. 293–334). San Diego, CA: Academic Press.

Schooler, J. W., & Melcher, J. (1994). The ineffability of insight. InS. M. Smith, T. B. Ward, & R. A. Finke (Eds.), The creativecognition approach (pp. 97–133). Cambridge, MA: MIT Press.

Schooler, J. W., Ohlsson, S., & Brooks, K. (1993). Thoughts beyondwords: When verbalization overshadows insight. Journal ofExperimental Psychology: General, 122, 166–183.

Stacey, M. K., & Eckert, C. M. (1999). An ethnographic metho-dology for design process analysis. In U. Lindemann,H. Birkhofer, H. Meerkamm, & S. Vajna (Eds.), Proceedings ofthe 12th International Conference on Engineering Design (Vol.3, pp. 1565–1570). Munich, Germany: Technical University ofMunich.

Stacey, M. K., Eckert, C. M., & McFadzean, J. (1999). Sketchinterpretation in design communication. In U. Lindemann,H. Birkhofer, H. Meerkamm, & S. Vajna (Eds.), Proceedings ofthe 12th International Conference on Engineering Design (Vol.2, pp. 923–928). Munich, Germany: Technical University ofMunich.

Stacey, M. K., Eckert, C. M., & Wiley, J. (2002). Expertise and cre-ativity in knitwear design. International Journal of New Prod-uct Development and Innovation Management, 4, 51–66.

Visser, W. (1992). Designers’ activities examined at three levels:Organisation, strategies and problem-solving processes.Knowledge-Based Systems, 5, 92–104.

Walker, D. J. (with Cross, N. G.). (1983). An introduction to design[Open University Course T263 Design: Processes and Prod-ucts, Unit 1]. The Open University, Milton Keynes, UK.

Wulf, F. (1922). Beitrage zur Psychologie der Gestalt: VI. Über dieVeränderung von Vorstellungen (Gedächtniss und Gestalt)[Contributions to Gestalt psychology: About change of visual-izations (memory and form)]. Psychologische Forschung, 1,333–375.

Appendix: Debriefing Questions

Personal data• Name• Age• Courses taken• Placement• Career ambition• What do you like most in knitwear?• What skill do you think is most important in a

knitwear designer?

Unstructured information• Would you like to talk me through your

sketches?• Please tell me about the decisions you made

when you were designing and what you saw inyour head

What is the agenda?• What effect do you want to achieve?• How does this relate to the source?• How much are you influenced by your own

taste?

382 Creativity Research Journal

C. M. Eckert and M. K. Stacey

Dow

nloa

ded

by [

Flor

ida

Inte

rnat

iona

l Uni

vers

ity]

at 0

8:52

22

Dec

embe

r 20

14

Page 30: Adaptation of Sources of Inspiration in Knitwear Design

Sketches• Do you normally produce sketches when you

design?• Could you do the same designs without

sketching?• Do you sketch each design that comes to mind?• What percentage do you sketch?• How do you evaluate which ones to sketch?• Do you think you are a good sketcher?• Does this matter?• Do you think sketching is important to your

thought process?

Design Steps and Problem SpacesWhat were the steps?Single?Multiple?What problem spaces?What goals have they got?How is problem behavior influenced by goals?• Do you set yourself a detailed problem specifi-

cation before you begin a design?• How do these change?What transformations did they do?What do they think is a complete solution? Might vi-

sualize complete garment, but not see complexity, i.e.Subjective feeling of complete solution is not equal tocompletely worked out solution

Visualization• What do you see for each sketch?• Do you see garment on person?• Can you focus in on design detail?• Can you rotate the image in your mind?• Can you see a person moving in the garment?

Schön Questions• How does your sketch influence your design?• Do you change the plan or your mental image

of the design as you go along?• What makes you change it?• Do you draw a design that you imagine?• Do you decide what to do next by responding to

your sketches?

Vividness of image• Do you see color?• Can you vary the colors?

• Can you see knitted structures?• What does the mental model look like?

Design decisions• Why did you select … ?• Which decision did you make?

Evaluation• When did you do evaluation?• How did you evaluate?• Do you also evaluate a mental image?• Do your evaluations coincide with your deci-

sion points?

Design progression• How do you get from one design to the next?• How do you know you have arrived at a satis-

factory design?

Concurrent Verbalization• What do you think about the concurrent

verbalization?• Did it change the design process?• Did it change the way you think?• What did you talk and what did you not talk

about?• What happened in your mind while you did not

talk?

Thinking style questions• Do you talk to yourself when you are

designing?• Do you talk to yourself when you are thinking

normally?• Do you see pictures in your mind when you are

thinking?• When do you see pictures?• For example when you are planning what to

cook for dinner, how do you think about it? Doyou see images of the food or ingredients?

Sources of inspiration• How typical is the task I have given you?• How do you normally use sources of

inspiration?• How do you select them?• What are you taught about the use of sources of

inspiration?

Creativity Research Journal 383

Adaptation of Sources of Inspiration

Dow

nloa

ded

by [

Flor

ida

Inte

rnat

iona

l Uni

vers

ity]

at 0

8:52

22

Dec

embe

r 20

14

Page 31: Adaptation of Sources of Inspiration in Knitwear Design

Creativity• What do you consider creative behavior?• Which type of people do you consider creative?• How could creativity be taught?

Design Culture• How typical is this use of sources of inspiration

for work in industry?• How relevant was what you learned in college

in your placement?

• How typical was your placement experiencecompared to your classmates’?

Collection• What will you do for your collection?• How are you going to go about planning it?• What inspiration will you use?

384 Creativity Research Journal

C. M. Eckert and M. K. Stacey

Dow

nloa

ded

by [

Flor

ida

Inte

rnat

iona

l Uni

vers

ity]

at 0

8:52

22

Dec

embe

r 20

14