by D. Cavallo

The empirical basis of this paper is a two-yearproject to bring new learning environments andmethodologies to rural Thailand. Pilot projectswere mounted outside of the education system,with the specific purpose of breaking“educational mind-sets” that have beenidentified as blocks to educational reform. Asalient example of such a mind-set is theassumption that the population and teachers ofrural areas lack the cognitive foundations formodern technological education. The workrequired a flexible approach to the design ofdigitally based educational interventions.Analysis of design issues led to a theoreticalframework, Emergent Design, for investigatinghow choice of design methodology contributesto the success or failure of education reforms.A practice of “applied epistemologicalanthropology,” which consists of probing forskills and knowledge resident in a communityand using these as bridges to new content, wasdeveloped. Analysis of learning behaviors led tothe identification of an “engine culture” in ruralThailand as an unrecognized source of “latentlearning potential.” This discovery has begun tospawn a theoretical inquiry with significantpromise for assessment of the learning potentialof developing countries.

The central thrust of this paper is the presenta-tion of a new strategy, which I call Emergent De-

sign. The paper describes an approach used for ed-ucational intervention; the claim is a more generalone, however, in that the strategy is appropriate insettings for technologically enabled paradigmaticchange. I claim that the more traditional approachesto systems design, implementation, and deploymenthave not produced desired results in situations wherethe goals and needs are for systematic change. When

the desired changes cannot be reliably foreseen, andparticularly when the target domain is computation-ally too complex for automation and thus relies onthe understanding and development of the peopleinvolved, then top-down, preplanned approacheshave intrinsic shortcomings and an emergent ap-proach is required.

Educational environments definitely possess thesecharacteristics. However, in the emerging businessand cultural environment, many other domains doas well. I have utilized this approach previously inthe design and implementation of enterprise archi-tectures and process re-engineering. The most no-table example1 is a health care delivery environmentwhere the Emergent Design of the architecture andapplications of the systems for health care delivery,administration, and patient use enabled a broadchange in medical practice. The approach to the de-sign of the educational intervention I describe hereresembles that of architecture, not only in the diver-sity of the sources of knowledge it uses but in an-other aspect as well—the practice of letting the de-sign emerge from an interaction with the client. Theoutcome is determined by the interplay between theunderstanding and goals of the client, the expertise,experience, and aesthetics of the architect, and theenvironmental and situational constraints of the de-sign space. Unlike architecture, where the outcome

rCopyright 2000 by International Business Machines Corpora-tion. Copying in printed form for private use is permitted with-out payment of royalty provided that (1) each reproduction is donewithout alteration and (2) the Journal reference and IBM copy-right notice are included on the first page. The title and abstract,but no other portions, of this paper may be copied or distributedroyalty free without further permission by computer-based andother information-service systems. Permission to republish anyother portion of this paper must be obtained from the Editor.

CAVALLO 0018-8670/00/$5.00 © 2000 IBM IBM SYSTEMS JOURNAL, VOL 39, NOS 3&4, 2000768

Emergent Design andlearning environments:Building onindigenous knowledge

is complete with the artifact, the design of educa-tional interventions is strengthened when it is ap-plied iteratively. The basis for action and outcomeis through the construction of understanding by theparticipants.

The technological ramifications are immense. We of-ten build inappropriate technology because the do-main changes too quickly, or the designers’ under-standings and aesthetics vary too much from theusers’ understandings, needs, and goals. At othertimes projects fail because, even though the tech-nology might be appropriate, the deployment isflawed. Design cycles that cannot adapt to rapidlychanging conditions miss emergent phenomena thateither need correction because they are undesirable,or need capitalization if desirable. A resultant long-term problem is lack of belief in the true possibil-ities for technology because it did not live up to ex-pectations. This is certainly the case in education,although business uses also share this outlook, as ev-idenced by complaints about lack of productivitygains through technology.

Perhaps more importantly, traditional approachesto learning of and through technology have not mo-bilized the indigenous knowledge and expertiseamong many people. The growing “digital divide”—concerns about the potential of a widening gap be-tween rich and poor in the new, knowledge-basedglobal economy due to a lack of modern, technolog-ical skills among people in lower social-economicstrata, and a growing concern about the potential ofeducational systems to ameliorate this situation—all point to a serious problem becoming seeminglypermanently intractable.

This paper describes an approach to technology de-sign and use that provides hope for a different, morepositive outcome. The same technology that can bea primary factor in widening the divide, may be thebest hope for eliminating the divide. The EmergentDesign approach enabled the discovery and utiliza-tion of latent, engineering expertise and creativityamong people in rural Thailand. Rather than beingbereft of social capital necessary to succeed in thenew economy, these traditionally poor, rural peopleare conceivably better situated for success so long asthe technology and methodology used is expressive, ap-propriable, and constructionist.

While the claims here are broad, I choose to focuson one concrete example, that of an effort in edu-cational reform. Educational institutions, although

relatively young, have proven extremely resistant tochange.2 Moreover, schools for the most part havenot used new computational technology in innova-tive ways. This, despite a lot of hype for the possi-bilities of technology in education, has caused manyto doubt the potential in the technology. The prob-lem, though, is not with the technology per se, butrather with the design, deployment, and uses of thetechnology.

Technology and the reform of educationalenvironments

Educational reform efforts, over a long period oftime, have offered many different blueprints. Yetnone has had the substantial effect for which it wasdesigned. Why is this the case?

We need to look at the way in which education re-forms are usually carried out. Some set of individ-uals decide there is a problem needing addressing(such as low math and science scores) or a changedeserving implementation (such as the introductionof a new item like ethics to the curriculum). A groupconvenes. They call in the various experts, stakehold-ers, practitioners, and other usual suspects. They de-sign a blueprint for their reform. The blueprint con-tains a curriculum, materials, texts, assessment,teacher training, and so on.

This paper presents the view that these blueprintshave failed simply because they are blueprints. Manyanalysts researching this situation, most recentlyDavid Tyack and Larry Cuban,2 have shown how theprocess fails. Whatever blueprint is proposed, it isinevitably going to be transformed in the course ofappropriation, ending more in conformance withwhat the designers originally hoped to reform. Theinstitution tends to reform the reform, perhaps re-taining the rhetoric but rendering it toothless. Ty-ack and Cuban brilliantly term3 the overriding mind-set the “grammar of school.” Like a grammar, theydescribe a deeply held organizing system that allowsonly certain expressions (or actions) as legitimate andrenders some expressions nonsensical if they devi-ate from the underlying system.

Tyack and Cuban made clear that whether reformsare big or small, from the “right” or from the “left,”national or local in scope, they do not work. Somemight deserve to fail because of the nature of theircontent. But while content may or may not be a lim-iting factor, they fail because of the form in whichthey were designed.

IBM SYSTEMS JOURNAL, VOL 39, NOS 3&4, 2000 CAVALLO 769

What is needed is an alternative approach that is nota blueprint. This naturally raises the question ofwhether having no blueprint means the abrogationof all design and planning so that “anything goes.”In the same way that a jazz group can improvisewithin the structure of a piece while remaining co-ordinated and within the theoretical principles of thegenre, so too can an emergent design remain con-sistent within a core set of principles.4

This paper describes a form of intervention in learn-ing that is very different from the model of reformstudied by Tyack and Cuban. It offers hope for ad-dressing the great educational needs created by thedigital age by drawing on two of its important inno-vations: (1) digital technology and (2) the approachto management of organization and of organizationalchange that has come in the wake of the technology.

More precisely, this work draws on the combinationof these two innovations. A distinction must be madebecause, as I show, the temptation to use either ofthem alone has led to failure. It is the combinationthat offers an optimistic vision for the future of learn-ing—the combination of these two products of thedigital age along with a theoretical framework basedon the work of pre-digital-age thinkers who knewwhat to do but did not have the means to do it.Among these the most central is Paulo Freire,5 butalso represented are John Dewey6 and, although hedid not focus on education per se, Jean Piaget.7

Limitations of a single focus. A focus solely on tech-nology leads to technocentrism, that is, a view thatit is the technology and not what we do with it thathas impact.8 Such a focus also leads to a narrownessof vision. In other words, we simply place the tech-nology into the existing structure and thus are notable to see the possibilities that extend beyond theexisting organization. Merely adding technology re-inforces an experimental paradigm out of place. Thisparadigm tries to modify one element at a time, hold-ing the others constant. When using such an ap-proach when introducing technology, what one holdsconstant—rather than maintaining experimental pu-rity—merely serves to neuter the potential for ed-ucational change catalyzed by the technology. Thus,an erroneous view of the technological and learningpotential results.

In his book The Productive Edge, 9 Richard Lesterdescribes mistakes made within conventional mind-sets about business, productivity, and change that res-onate with conventional mind-sets of education and

school reform. Lester describes how many compa-nies, in an attempt to improve productivity, quality,or some other often highly quantifiable attribute,would attempt to apply a seemingly scientific methodby researching a new methodology or so-called bestpractice; they then attempted to test whether add-ing this method to their own operation would gen-

erate positive results, holding all other things con-stant. In a vast majority of cases such applicationsof new methods failed to produce positive results.This not only called into question whether the newmethodologies truly had value, but also led to an ex-perimental fatigue from being repeatedly forced toadapt to the change program of the month.

What Lester demonstrated was that there are nottypically such things as decontextualized best prac-tices that can be grafted onto existing organizationsand thereby produce results. Rather, each companyhas its own complex culture, full of subtleties, andsuccessful companies are the ones that can innovate,cultivate, adapt, and use methods that can thrive intheir particular environment. The successful ap-proaches fit more with the Emergent Design con-cept advocated here than with the more traditionaltop-down, change-one-variable-at-a-time approachesthought to be more in the line of scientific manage-ment. Digital technology enabled the customizationof process to culture rather than forcing culture tobe responsive to management dictates and “the onebest method.”10 The critical point is that adoptionand implementation of new methodologies needs tobe based in, and grow from, the existing culture, andtypically fails when it is merely imposed from abovewithout such cultural considerations. Interestingly,incremental approaches to educational reformclosely resemble the less successful methods thatLester describes.

The reform of educational management, usually inthe form of administrative decentralization, does not

CAVALLO IBM SYSTEMS JOURNAL, VOL 39, NOS 3&4, 2000770

Adoption and implementation of new methodologies needs

to be based in, and grow from,

the existing culture.

break the stranglehold of the grammar and ends upwith reversion to type. By themselves, ideas such asdecentralization of control and decision-making, orintradistrict competition, do not generate new con-tent and methods. On the contrary, they merely pushthe same practices down the hierarchy without fun-damentally changing practice. Thus, the only sub-stantial change is in administration, not innovationin the learning environment.

The need for new principles. Saying one needs tobase new methodologies and the change process toresonate with, and build upon, existing culture doesnot mean that deep change is not intended, nor thatany type of change is desired. In the case of learningenvironments, the primary principles we broughtwere: constructionism, technological fluency, immer-sive environments, long-term projects, applied epis-temological anthropology, critical inquiry, and Emer-gent Design.

Constructionism builds11–13 upon principles in con-structivism. While constructivism holds that thelearner constructs new knowledge based on the ex-isting knowledge he or she has, constructionismbuilds on this idea by maintaining that this processhappens particularly well when the learner is in theprocess of constructing something. For example, inour work with LEGO**-Logo14,15 we witnessed manychildren, including those who had previously doneextremely poorly in school, understand complex ideasin mechanics, physics, and mathematics through con-structing LEGO robots to accomplish various tasks.

The idea of building technological fluency draws onthe image of being fluent in a language.16,17 Whenone is fluent in a natural language one can think,express, communicate, imagine, and create with thatlanguage. In the same way, we like to develop flu-ency through the construction of, and with technol-ogy as a means of, personal and group expression.We try to develop fluency with technology in orderto help people become more eloquent and effectivein their expression. Just as fluency changes the fo-cus to a more holistic use of natural language, thisalso changes the focus of learning with technology.

Just as the idea of fluency is adopted from language,so too is the concept of immersive environments. Be-ing immersed in the culture and environment facil-itates learning a foreign language. So too does work-ing with others in a culture where the knowledge oftechnology and construction is deeply embedded fa-cilitate the development of technological fluency.

Building artifacts of interest to learners aids the con-struction and the development of fluency. In orderto delve deeply enough to unearth the underlyingconcepts and principles, we enable students to workon projects over a long period of time. Rather thanrushing through a broad curriculum in a shallowmanner, we prefer to encourage diving deeply intothe projects. This takes time. This also differentiatesour practice from other project-oriented approachesin education, where the project is preplanned by thecurriculum designers and not emergent from the in-terests of the learners, and where it lasts only a shortperiod of time in order to fit the traditional class-room situation.

Applied epistemological anthropology is a term I havegiven to the practice of unearthing the meaninglearners attribute. This applies on both a cultural andindividual basis. In order to facilitate the construc-tion of new knowledge on the existing knowledge ofthe learners, one must first help discover the exist-ing frameworks as best one can. This practice itselfis facilitated through the construction of objects ofinterest to the learner, where the learner has as muchfreedom of expression as possible. When the free-dom of expression exists, then the learner has thespace in which to express himself or herself in a man-ner faithful to the learner’s thoughts. This is a keyelement in the design of technologies for learning.Through the construction, and mediated by discus-sion, the underlying thoughts become more evident.This enables the teacher or facilitator to better de-sign and implement learning interactions. This leadsto the necessity of a more emergent approach.

Critical inquiry is the process of engaging in a con-versation with one’s world in order to understandand act upon it.5 Through critical inquiry we collabo-ratively determine upon which projects to work. Alsothrough critical inquiry we try to understand the phe-nomena of study in sufficient detail so as to constructartifacts modeling the phenomena or designed toameliorate the situation, as well as to understand anddebug the artifacts of construction.

Emergent Design is what manages the overall pro-cess. Due to the emphasis on approaching learningby building on the existing knowledge of the learn-ers through their expressive construction of projectsof their own choosing, this process by definition hasstrong emergent tendencies. However, design is alsoemphasized as the others in the community who workwith learners—be they teachers, parents, or othercommunity members—also play an active role in as-

IBM SYSTEMS JOURNAL, VOL 39, NOS 3&4, 2000 CAVALLO 771

sisting to assist and guide the learner in the process.The idea of design extended to a “grassroots” levelenters because, just as Emergent Design is practicedto facilitate organizational change, so too is Emer-gent Design practiced in the interaction betweenteachers and learners.

There is no claim that the methodology describedin this paper is the only way to achieve the desiredresults. What this does serve as is an existence proofof a way. That this way at least initially demonstratesinteresting, unexpected, and quite positive resultsamong populations that previously did not exhibitsuch results in traditional settings should serve toquestion the existing school grammar as well as tofacilitate other such experiments. I close this sectionwith a story that illustrates the prevailing educationalmind-set as well as effective informal methods thatcan be leveraged.

DOS commands and flower gardens. On my first visitto Thailand my hosts took me to a nonformal ed-ucation (NFE) site in a Buddhist temple. I saw a com-puter class held at the NFE temple school. A childwas being taught DOS commands. The logic behindsuch an introduction to computers, following the typ-ical school curriculum grammar of using sequentialbuilding blocks of knowledge, is that it provides therequisite basis for later, more difficult learning. How-ever, the useful learning never comes! And in themeantime, the formalistic nature of the beginningwork confuses and frustrates the novice.

The student’s teacher assigned four commands forhim to learn and practice. The first was dir, to geta listing of files in his directory. The second was copy,to copy a file from one location to another. The thirdwas format, to format his A drive (fortunately, it wasnot the C drive). I do not remember the fourth butit was made irrelevant by the reformatting of his disk.

This confounding situation led the student to stopme with a plaintive question, “What is the problemhere? It worked before but now it no longer works.I am following my teacher’s instructions, but this isnot working properly.” On the first iteration of prac-ticing his commands by rote, everything was fine.Subsequently, however, none of the commands wasgiving the specified results. His directory was nowempty. He could not copy his file. I explained to himthat the result of using the format command is thatit reformats the entire disk, meaning it wipes cleanwhat was on it and sets it for the computer’s oper-

ating system. Thus, there were no more files in hisdirectory to list or to copy.

Despite several attempts at various ways of explain-ing it, including recreating the example on the com-puter and showing how dir, copy, and format workwith a newly created set of files, I am not sure heunderstood my explanations. One reason for this is

that my explanations meant that what his teacher hadsaid, done, and assigned no longer made sense, whichwould be quite disorienting. Another possibility isthat no matter what the explanations and examples,learning commands this way is too decontextualizedto make sense. One is merely learning by rote whatsomeone else says is important without any concep-tion of why or how it might be used.

The split between conventional “School thinking”18

and cultural learning was shown vividly in the con-trast between the computer class at the temple andhow the monks themselves teach flower gardening.Beautiful flowers are grown and displayed at all theBuddhist temples in Thailand. They are impressive,colorful, and fragrant. After my visit and the expe-rience with DOS teaching run amok, I inquired abouthow people learned to cultivate such gorgeous gar-dens. A monk explained that when initiate monksenter the temple, they work alongside more expe-rienced ones and learn by demonstration, by askingquestions, in the best sense of learning by doing. Imischievously asked whether any classroom instruc-tion was involved. The monk looked at me askance,but politely answered no, they felt there was no need.I tried to explain that this was the approach we alsopreferred for learning computational ideas. That is,that new learners work on projects of their own, arein an environment with others working on similar,but perhaps more complex, projects, and can observeand ask others questions—in essence they are im-mersed in a culture of computing just as the monksare immersed in their culture.

CAVALLO IBM SYSTEMS JOURNAL, VOL 39, NOS 3&4, 2000772

I suggested that we createenvironments for learning

computational ideas in the same manner

that monks learn gardening.

The depth of resistance to these ideas was illustratedby the way the teacher who was translating my re-marks into Thai misrepresented the explanation, cre-ating an initial misunderstanding between the monkand me. After listening to my translator, the monkpolitely responded that they would never do what Isuggested. Considering that I had just suggested thatwe create environments for learning computationalideas in the same manner that the monks learn gar-dening, I could not understand how he disagreed.So I inquired again about what was said. The teachertold me that she told the monk I had suggested thatthey teach gardening in a classroom just as we teachcomputers in a classroom. When I re-explained whatI had really intended, the teacher could not believeI meant it. Rather than immediately retranslating,she passionately protested. Surely classrooms werethe modern and most effective means of teaching.How could I, from a modern western university, sug-gest that the monk’s method could be better? It tookquite a while to get her to tell the monk what Ithought. In retrospect, this was a powerful learningmoment for my translator, although thoroughly andnecessarily unplanned.

Scenes from Project Lighthouse

The context of this paper is Project Lighthouse, abold intervention to initiate radical change in the ed-ucational processes in Thailand. As its name suggests,Project Lighthouse is not a blueprint for educationor education reform. Rather, it attempts to highlightactual possibilities for powerful learning environ-ments in Thailand, particularly in settings where tra-ditional education has not succeeded. A primary goalis to break mind-sets about what education must beby providing concrete examples. The following aresamples of activities from Project Lighthouse overa 17-month period. The scenes provide a concretebasis for the discussion that follows.

Bangkok, March 1997. In the first scene, SeymourPapert and I, from the MIT Media Lab, were meet-ing with leaders of the Suksapattana Foundation.19

We were designing a proposed intervention intendedto provoke a radical reform of the educational sys-tem in Thailand. The meeting came about becausea group of industry leaders and government officialshad come to believe that, unless they achieved a to-tal transformation of their educational system, Thai-land would not merely stagnate economically, butalso that they would lose all the gains of the previ-ous decade. More critically, the leaders worried thatthere was a growing and more intractable divide be-

tween rich and poor that would destroy the fabricof Thai society. They further believed that in the ab-sence of an educated, thoughtful, literate populace,it would be impossible to support their nascent de-mocracy and prevent a return to autocratic, and cor-rupt, military rule.

The Thai leaders believed that the existing schoolwas not a hospitable medium for developing alter-native forms of learning. Moreover, they felt that tochange it directly would cost too much and take toolong. They believed the existing schools to be toorigid, too reliant on rote instruction, and staffed bytoo many teachers who were barely educated them-selves.

They had set bold and ambitious goals for their ed-ucational system. They had developed a new nationaleducation plan as an essential part of their nationaldevelopment plan. This education plan, combinedwith a special commission from the Office of thePrime Minister devoted to education reform (ONEC),specified the new goals.20 The goals were thought-ful and admirable. They included:

● Becoming learner-centered● Developing critical-thinking ability● Fostering innovation and creativity● Developing collaborative spirit and skills● Learning how to learn● Providing familiarity, ability, and comfort in work-

ing with technology● Developing “happy” learning, that is, a joy for

learning

However, none of the plans specified how to achievesuch a system. They did not discuss how to operatein this new paradigm or how to make the transfor-mation. Thus, while the goals were lofty, the imple-mentation of both the new system and the methodof reforming the current one, were mired in the ex-isting, undesirable paradigm.

The goal behind our endeavor, Project Lighthouse,was to break mind-sets by creating technologicallyrich learning environments that would demonstratethe “out-of-the-box,” yet practical, possibilities forchildren in Thailand.

However, it was not clear what to do. Moreover,there was little agreement on, or acceptance of, ourproposal. Some believed that we should focus ongaining the acceptance of the national curriculumdevelopers as the current system moved only through

IBM SYSTEMS JOURNAL, VOL 39, NOS 3&4, 2000 CAVALLO 773

the curriculum. Others believed we from MIT shouldtrain the trainers who would then train the teacherswho would then work with the students. Others feltwe should place computer labs in more schools andtrain teachers to work there. Finally, there was nearunanimous agreement that the existing teacher corpswas incapable of working in a new, learner-centered,project-based, technologically rich environment. Vir-tually everyone told us that the teachers were barelyeducated themselves and might not be able to learnto use the technology, let alone teach with it.

We proposed creating four pilot projects where wecould quickly demonstrate significant results in someof the most critical areas of need.21 These were:

● Alternative learning environments within nonfor-mal education

● Rural village learning centers● Teacher development● Alternative learning environments for at-risk youth

such as street children in urban areas and girls atrisk for or exiting prostitution

There were two major objections to our proposal.First, commentators argued that our proposal didnot fit the prototype five-year plan, which spells outall activities over that time period. How could peo-ple know what to do if we did not provide such aplan? People wondered if perhaps either we werenot serious or did not know what we were doing. Sec-ond, people told us that the quality of the teachercorps was so low that they would be incapable of car-rying out an ambitious endeavor such as ours.

However, we argued that it would be counter-pro-ductive, if not impossible, to develop any specificplan. It was not merely that we were not familiarenough with Thailand to know what would be theright things to do. More profoundly, what was neededwas a philosophy of design based on recognizing thatno one could know beforehand what would resonate,how people would appropriate new learning tech-nologies and methodologies, what learners wouldchoose as projects, how villagers would react to theintervention, and so on.

At the meetings we tried to show that there is a fun-damental contradiction between having learning en-vironments that function through connecting to,building upon, developing, and deepening the inter-ests of the learners, and planning everything centrallyin a top-down manner where all activities are pre-determined for all learners and all locations. What

is needed is a philosophy of design for educationalinnovation as different from traditional ideas of re-form as the content of the new innovation would befrom traditional educational content. The theoret-ical framework that evolved from this and similarexperiences of Project Lighthouse is Emergent De-sign.

The phrase Emergent Design puts a spotlight on theneed (which has not been recognized by educationpolicymakers) to study the conceptual space wherethe purposeful stance implied by the word “design”mates with the openness implied by the word “emer-gent.” This mating underlies modern approaches toorganizational practice.

The emphasis on emergence as the guiding principledoes not imply that this is an anything-goes environ-ment reacting to the whims of the participant teach-ers and learners. As described above, we brought avery disciplined set of principles, methodologies,tools, activities, models, and exemplars for learningenvironments. However, to deliver a pre-set curric-ulum with pre-chosen problems, explanations, andsequence of events would be not only antithetical tothe underlying learning philosophy, but also it wouldbe incapable of taking advantage of the very ben-efits that the technology affords.

Nong Baot village, BuriRam province, northeasternThailand, January 1998. The second scene fromProject Lighthouse took place in Nong Baot in thenortheast of Thailand, the poorest region of thecountry. It is approximately 100 kilometers from theCambodian border. The New York Times described22

it as having “two harsh seasons, flood and drought.”The economy is based on agriculture but, due to theharsh weather, little can be grown. Nong Baot sur-vives by cultivating one rice crop per season. Thereare some small vegetable plots used primarily for sub-sistence, because there is not enough water to growenough crops to sell. Lately, some groups of villag-ers have tried to cultivate fish farms by creating smallreservoirs during the rainy season. This, too, providesfood for them for only a brief time, because the wa-ter is gone within a few months.

Nong Baot is an area that suffers from logistical prob-lems that have stifled the potential for economic de-velopment.23 It is tropical and does not have readyaccess by water to the rest of the world. These fac-tors inhibit the development of industry. The soil ispoor and there are no mineral deposits. Thus, it hasremained an area of minimal means and wealth.

CAVALLO IBM SYSTEMS JOURNAL, VOL 39, NOS 3&4, 2000774

Education in this area likewise has been minimal.There is little incentive to remain in School. Manypeople leave school as soon as they are legally el-igible, claiming that School has no relevance to theirlives. Children need to work in the fields or in otheroccupations to help their families. Few people go onto attend a university.

Within this scenario, I conducted an introductoryLogo immersion workshop to develop technologi-cal fluency. Unlike most projects that try to bringtechnology to remote or impoverished areas, my goalwas to have the attendees quickly build projects andcreate programs.

The workshop had a mix of participants: villagers,teachers, and a few local economic developmentworkers from the Population and DevelopmentAgency. I begin the workshop by showing what acomputer is, how you turn one on and off, and howyou operate one, because this was the first time thatthe villagers had ever personally seen a computer,except for viewing one on television.

In the evening the MIT participants held discussionswith the villagers to get to know them and their sit-uation. I asked why the villagers said they wantedus to place computers and Internet connections intheir village. They told us that water is very scarcein this region. Worse, there is either too much of itduring the two-month rainy season or there is noneof it during the rest of the year.

It was in the discussion in the evenings after thisworkshop that the village leader expressed the needof the people to gain more control over their livesand the belief that certain uses of the technologycould help them. The people described many of theirproblems as economic, caused by the harsh climatewhere there was either too much or too little water.They wanted access to expert knowledge, but mostimportantly they wanted to be in control of gainingthe access to and making the decisions about whatto do with the knowledge. They felt that the localauthorities did not involve them in the thought pro-cess and decision-making whenever the villagersasked for assistance. This left the villagers feelingdependent and without the hope for their own pro-gress. To make matters worse, due to the appear-ance of new problems with the cattle and the water,the villagers believed the advice and proposed rem-edies they were given to be harmful rather than help-ful.

Introducing the first phase. The villagers wanted toend this cycle of dependency and lack of control bygaining access to information and gaining control ofthe situation via the technology. Even though I hadto introduce the workshop by demonstrating whata computer is, including how to turn one off and on,through the symbolic value of the computer theyviewed competency with the technology as a plau-sible path to this control. Although there could nothave been any real experience with how computa-tional technology could provide this path, they hadheard enough about computers and the Internet tobelieve it had potential for them.24 The computerwas a symbol of modern technology and a connec-tion to the modern world.

In short, the villagers were able to experience whatwe did in the spirit of “cultural leverage.” As a re-sult, the participants were soon building their ownprojects, first in Microworlds Logo, then adding ro-botics with LEGO-Logo. What at first was a foreignand potentially intimidating technology, now becamea source of fun and pride in product. The villagersworked in multigenerational groups, from young chil-dren to the elders in their seventies and eighties. Theteenagers and children did more of the program-ming, being more open to new technologies. Theadults contributed their wisdom, maturity, and expe-rience. They made all of the decisions jointly. Theywere doing programming and engineering, workingon projects of their own design.

Moving to the second phase. When I returned in Au-gust, the situation was quite different. In a brain-storming session about potential projects, we quicklyconverged on the critical need for access to waterfor household and agricultural use. We discussedideas broadly at first, looking for areas they felt weremajor problems or, from a more optimistic point ofview, areas that they thought could provide majorbenefit if we could find means to create solutionswithin these areas.

Naturally, there were many trade-offs in the dimen-sions of each project. The MIT participants wantedto address major problems, but some problems per-haps were extremely difficult to solve. We wanted toachieve some quick successes to help change mind-sets about possibilities, produce real results, and de-velop belief in what we were attempting. However,easy and quick successes are most likely rare or triv-ial. We believed in the potential of the technologyto help think about and design potential beneficialprojects, but the villagers were technical neophytes.

IBM SYSTEMS JOURNAL, VOL 39, NOS 3&4, 2000 CAVALLO 775

We did not want to just design solutions ourselves,because this would neither develop the villagers’ owntechnological fluency and capabilities nor empowerthem in the long term. So we needed to choose ini-tial projects that were approachable by people withtheir scant experience, yet were real enough to ac-tually provide tangible benefit, while simultaneouslyproviding a rich learning experience. Through myexperience both in developing technical solutions toreal-world problems, as well as working with adultsand children learning to engineer and program, weworked to develop a group consensus on the initialset of projects. Knowing that we were committed forthe long term relieved tension from feeling a needto accomplish everything immediately. The initialprojects were to design a dam to create a reservoirfor farming; investigating alternative strategies forrice cultivation; redesigning the irrigation system; de-veloping new means to collect, store, purify, and dis-tribute rain water; and creating new vegetable plots.In this discussion I focus primarily on the dam proj-ect.

A first, important project. We began work calculatingthe potential and the reality of building a dam. Ineach of the past two years the villagers had tried toconstruct a dam to create a reservoir. It was hopedthat the dam would retain water at the end of therainy season that could be used for agriculture in thedry season. In each of the past two years the projecthad failed, since the reservoir did not contain thewater. Now both the villagers and the rural teachersworked to develop the new project together. I tooka supporting, mentoring role rather than a direct rolein the project myself, believing that the only sustain-able benefit would be for them to develop the pack-age of skills themselves.

They had not previously calculated the potential ben-efit from the dam. When we engaged in brainstorm-ing about this topic with them, together we calcu-lated that the villagers would more than double theiryearly income if they could harvest a second vege-table crop. We walked through the flood plain andtook some digital photographs. We measured the dis-tances between relevant objects in the terrain usingthe odometer on a motorcycle. We uploaded thephotographs into Microworlds Logo and the groupsbegan making visual representations of the area.

To my surprise this was a totally new experience, notmerely for the villagers but also for the teachers.While the fact that the villagers could not do this ontheir own might not be surprising, the teachers could

not do so either. They had certainly taken schoolcourses and passed school exams on this type ofknowledge, yet in practice they could not make amap. Together, the teachers and villagers created ac-curate computer representations of the areas, pre-serving distances, maintaining relationships and ra-tios as they created various views at different scalesand calculated the relevant distances between im-portant objects.

Then, a remarkable thing happened! Immediatelyupon creating the maps, we discovered a mistake re-peated each of the previous two years. The villagershad been building the dam in the wrong place! Theoriginal location benefited from natural terrain tocreate the reservoir; however, it was about two kilom-eters from the village water pump used for irriga-tion. Once the villagers constructed their own mapof the area, they realized they could not create a res-ervoir large enough to cover the distance to thepump. Even if the dam had functioned properly, itwould not have provided the expected benefit, be-cause it was prohibitively expensive to relocate thepump and the irrigation hoses.

Discovering an exceptional student. As the design proj-ect continued, we observed how the efforts of oneof the participants was exceptional. He told me thathe had not had any success in school and left as soonas it was legal. He primarily helped his family withthe farming. We had only introduced computers tothe village within the current year. He spent this timeworking on programming—not by taking classes, butby programming his own projects.

What was so striking was that he had quickly becomequite an adept software hacker.25 Atypical of manyof our experiences with more educated people, he,as well as others in other parts of Thailand, divedin and figured out how to build the projects hewanted. If something did not work, he was notdaunted. Rather, he debugged the system andworked until it was satisfactory.

We discovered that he spent considerable time work-ing with engines. By learning how to build and re-pair engines and by working on the farm with fewresources, he had developed a bricolage spirit. Thatis, he would make what they needed with what littlehe had. If something did not work, he fixed it. If hedid not have the right tool or material, he impro-vised. He took this spirit and applied it to compu-tational technology.

CAVALLO IBM SYSTEMS JOURNAL, VOL 39, NOS 3&4, 2000776

As this skill and experience became apparent, he andothers took me to visit their farms. At the farms, ev-eryone who could used a small Kubota diesel engineto power a wide variety of local technological con-traptions. They used the little motors to power ricemills, well-water pumps, irrigation pumps, one-per-son tractors, field vehicles, and even lightweighttrucks. The barns contained little pulley systems forlifting the motor from one device to another. Thelogic of each machine was open and obvious. Theinnovation and creativity were remarkable. The util-ity was tremendous. The people had taken objectsfor other, often quite specific, purposes and com-bined them in a general-purpose melange particu-lar to their needs, resources, and budgets. The expe-rience and expertise of those who worked with theseengines and devices was quite impressive.

Thai combustion-engine culture. Virtually all com-mentators on Thai education and on the ProjectLighthouse proposal believed that the quality of ru-ral teachers was extremely poor and that they wouldbe unable to work successfully in the proposed tech-nologically rich, learner-centered environment.These same commentators bemoan the problemsand capabilities of the overwhelming majority of ru-ral students as well. Lack of faith in the intelligenceand capability of economically disadvantaged chil-dren is an unfortunately widespread belief that is alltoo difficult to dislodge.

Contrary to the perceptions that rural and impov-erished students are not capable learners, ruralteachers are not competent technologists, and Thaiculture is not amenable to innovation, collaboration,deep learning, and technical expertise, we discoveredthat there are deep intellectual roots and significantinnovation practiced and learned over at least manydecades, and presumably much longer. Indeed, al-though not written about in academic circles, thereis a strong tradition of so-called “peasant technol-ogy”26 in Thailand, particularly using and adaptingthe internal-combustion engine to satisfy local con-cerns and constraints. Our final scene in this sectionfocuses on this “engine culture.”

Perhaps the best example of this innovation is thecreation of the long-tailed boat (Figure 1). There aremany areas throughout the country where waterwaysare the principal means of travel. Significantly, thisis also the case on the rivers and canals of Bangkok.In the past, as people desired to transport more andheavier goods, human-powered boats became prob-lematic. In the north, one innovator decided to ex-

periment with placing motors onto the boats. Afterseveral attempts with various types of inboard andoutboard motors, he settled upon using an automo-bile engine with a long driveshaft so that the pro-peller was far from the boat. Typical outboard mo-tors did not work well, because they churned toomuch water into the long canoes that everyone used.The many reeds in the rivers also jammed the pro-pellers too often, negating their benefit. The drive-shaft, or long-tail, not only solved the churning prob-lem, but also served as a rudder for steering andenabled the pilot to lift the shaft from the water toavoid entanglement with the reeds. The use of anauto engine leveraged existing knowledge about re-pairs and benefited from not requiring parts man-ufactured outside of Thailand, which would be dif-ficult and expensive to obtain. People quicklyadopted this technological innovation throughout thecountry.27

Tuk-tuks are another similarly inspired innovation.Small motorcycle engines are placed onto the pedi-cabs, again to alleviate human stress and increasespeed. Other rural innovators have also adapted en-gines to create low-cost, one-person tractors (Fig-ure 2), irrigation pumps (including one ingenious in-vention to pump over roadways, since the native soilhad the tendency to crumble into irrigation tunnels),and devices to help operate wells in drought-strickenareas.

For the most part, not only did these innovations notoccur in universities, research labs, or corporate de-partments, such circles barely took notice of them.Rather, they were a grassroots effort, based in theinterests, needs, and practices of Thai culture. Peo-ple created and adapted new technologies to alle-viate their burdens and to create new opportunities.

These innovations could not have achieved suchwidespread use if a culture of practice and knowl-edge had not also developed to spread and supportthem. In order to use engines widely, a group of peo-ple capable of maintaining them had to exist. Thisgroup did not do well in school and did not receiveits training in school. Rather, almost exclusively theylearned to diagnose and repair engines in informallearning cultures. Making this diagnosis and repairmore difficult is the fact that among this social stra-tum in Thailand, there are not a lot of materials,parts, diagnostic equipment, or written manuals.These mechanics have to become bricoleurs, 28,29 thatis, they must adapt materials at hand to satisfy their

IBM SYSTEMS JOURNAL, VOL 39, NOS 3&4, 2000 CAVALLO 777

goals, even if it is not the accepted way to accom-plish the goal nor the proper materials for the task.

Discussion

What makes this story compelling is that these me-chanics, while respected for their mechanical abil-ities, were not regarded as academically capable.Conventional wisdom stated that people in this groupmay be good with their hands but they were not goodwith their heads. Moreover, the belief in the dichot-omy that different people with different skills are re-quired in order to be good with their heads remains.

However, in the context of Project Lighthouse, thecapability of these motorcycle and engine mechan-ics was immediately evident. Not only did they learn

the new computational technology quickly, they werealso quite adept at adapting it and applying it to solvelocal problems. This was the case with designingdams, improving irrigation, and devising alternativemethods of cultivation of rice and other crops inBuriRam.

Still, moving from one technology, engines, to an-other, computing devices, while impressive, would notnecessarily be remarkable except that in order to ac-complish the tasks with computational technologythey had to competently handle some sophisticatedmathematics, biology, engineering, physics, and com-puter science. What is remarkable then is that:

● They accomplished projects requiring competencein these recognized bodies of knowledge

Figure 1 Engine on a long-tailed boat

CAVALLO IBM SYSTEMS JOURNAL, VOL 39, NOS 3&4, 2000778

● They accomplished this in extremely short time-frames

● They leveraged their mechanical expertise and“hacking” spirit to build a computational techno-logical fluency

● They then utilized the technological fluency to gaincompetence in these bodies of knowledge previ-ously inaccessible to them

To make maps they had to measure distances andperform calculations over these distances. To pro-vide zoom-in and zoom-out views, they had to main-tain proportions and adjust accordingly. To designthe reservoir they had to again measure, calculateareas and volumes, and determine water usage forvarious crops over time while accounting for evap-oration and drainage. To think about placement of

the dam they had to think about how water flowsover terrain. To design the irrigation system they hadto think about networks and shortest paths. To de-termine which project to do or which decision tomake within a project, they had to calculate costsand benefits, factoring in more subjective factors aswell, and create compelling arguments to convinceothers. To test various rice cultivation methods andto create decision-support systems to assist them inthe cultivation and care, they had to delve deeplyinto the supporting science. To create new LEGO ro-botic-controlled apparatuses to assist in farming andenvironmental sensing (or just for play), they had togo deeply into the underlying engineering, control,mechanics, and physics. What unified these variousendeavors was the formal language for description.

Figure 2 Another local innovation, a small diesel-driven one-person tractor

IBM SYSTEMS JOURNAL, VOL 39, NOS 3&4, 2000 CAVALLO 779

Moreover, the dam design is but one example fromone site. Other sites also had similar results. For ex-ample, in the north where water was not a problem,people worked on issues of soil erosion, developingand testing alternatives to slash-and-burn farming,experiments in nutrition, and cultivation of newcrops. They also worked on social issues such as sub-stance abuse, public health monitoring and aware-ness, and creating community on-line magazines. Thepoint is not that everyone should design a dam, butrather that at each site the learners could work withinthe same methodology and same set of tools onprojects of interest and import to them. That eachsite developed uniquely is an important result of thiswork.

The significant accomplishment in this work is dem-onstrating a significant gain in accomplishmentamong a population that had not previously exhib-ited such competence in educational institutions.This work demonstrates how to build on and enhancelocal knowledge. Within the design of this learningenvironment, the learners:

● Work from local knowledge and interests● Bridge to other knowledge domains● Liberate their local knowledge from its specific sit-

uated embodiment

While others have demonstrated the ability of peo-ple to develop technology and use science withoutthe benefit of schooling,29 the key point here is thatthe constructionist use of computational technologyleveraged this ability and helped people apply theirknowledge to new and varied situations in an ex-tremely short period of time. The knowledge did notremain limited to the particular technology such ascombustion engines, but rather they could use themalleable computer technology as a tool for under-standing other domains. Moreover, bodies of knowl-edge such as mathematics and physics were openedto them in new and more accessible ways.

The role of the computer in this process is to drawon a set of skills that can be transferred to some-thing different. Combustion engines provided ameans for developing technical and diagnostic ex-pertise; applications remained mechanical, however.Through computational tools, learners design andconstruct and thereby make the forms of knowledgethey have more general. Developing technologicalfluency enables them to break out of the specific con-text and represent their knowledge in forms they candraw on in many contexts. Neither traditional ed-

ucation nor nonconstructionist use of technology en-abled the recognition and leverage of this indigenousexpertise.

Success was due to the existence of several criticalelements in the design and affordances of the tech-nology. The technology is a malleable, expressive toolfor construction. We do not merely use the computeras a means of delivery of information or as a meansof communication, although both of those uses arebeneficial. Rather, the users program, in languageslocalized to their own language, building their ownidioms, on projects of importance to them. By work-ing on a variety of projects over time, they developa technological fluency. The combination of relativefreedom of expression and self-selected projects ofinterest facilitated the mobilization and leverage ofindigenous knowledge.

The design of technologically rich learning environ-ments and the reform of education. Discovering theengineering expertise and hacking spirit among somany Thai people who had previously not succeededin school is a major benefit from the Emergent De-sign approach utilized in this project. Not only hadThai educators not built on this talent and intelli-gence, they did not even recognize it. The typicalschool reforms, despite their intention to promotecreativity, problem solving, technological capability,and so on, also are generally incapable of discover-ing and leveraging such local knowledge. This is dueto their top-down, preplanned, standardized, curric-ular approach.

There is no way to know beforehand for every sitewhat will resonate and what local concerns and lo-cal knowledge exist. What one can assume is thatthere always is something. Using the Emergent De-sign framework, combined with principles of learn-ing environments and open, programmable, techno-logical tools, this “something” can be built upon andleveraged.

The work suggests a conclusion with a very broadsweep: The latent learning potential of the worldpopulation has been grossly underestimated as a re-sult of prevailing mind-sets that limit the design ofinterventions to improve the evolution of the globallearning environment.

**Trademark or registered trademark of the LEGO Group.

CAVALLO IBM SYSTEMS JOURNAL, VOL 39, NOS 3&4, 2000780

Cited references and notes

1. D. Cavallo, Technological Fluency and the Art of MotorcycleMaintenance: Emergent Design of Learning Environments,Ph.D. thesis, MIT Media Laboratory, Cambridge, MA (2000).

2. D. Tyack and L. Cuban, Tinkering Toward Utopia: A Centuryof Public School Reform, Harvard University Press, Cam-bridge, MA (1995).

3. Ibid., p. 85.4. In the case of this work, the principles of the learning envi-

ronment include constructionism, technological fluency, com-puter immersion, long-term projects, learner-centered activ-ities, and connected projects. Later in this section I willprovide a brief description of each concept. While it is be-yond the scope of this paper to delve into detail for each, thesense of the work will emerge through the description here.

5. P. Freire, Pedagogy of the Oppressed, Herder and Herder, NewYork (1972).

6. J. Dewey, Experience and Education, Collier Books, New York(1938).

7. The Essential Piaget, H. E. Gruber and J. J. Voneche, Ed-itors, Basic Books, New York (1977).

8. S. Papert, Computer Criticism vs. Technocentric Thinking, E&LMemo #1, Epistemology and Learning Group, MIT MediaLaboratory, Cambridge, MA (1990).

9. R. K. Lester, The Productive Edge: How U.S. Industries ArePointing the Way to a New Era of Economic Growth, W. W.Norton & Company, New York (1998).

10. F. W. Taylor, The Principles of Scientific Management, DoverPublications, New York (1998).

11. S. Papert, “Introduction,” Constructionism, S. Papert andI. Harel, Editors, Ablex Publishing Corporation, Norwood,NJ (1991).

12. I. Harel, Children Designers: Interdisciplinary Constructions forLearning and Knowing Mathematics in a Computer-RichSchool, Ablex Publishing Corporation, Norwood, NJ (1991).

13. Constructionism in Practice: Designing Thinking and Learn-ing in a Digital World, Y. Kafai and M. Resnick, Editors, Law-rence Erlbaum Associates, Mahwah, NJ (1996).

14. D. Cavallo, Leveraging Learning Through Technological Flu-ency, master’s thesis, MIT Media Laboratory, Cambridge, MA(1996).

15. M. Resnick, A. Bruckman, and F. Martin, “Pianos, Not Ste-reos: Creating Computational Construction Kits,” Interactions3, No. 6 (1996).

16. D. Cavallo, “New Initiatives in Youth Development: Tech-nology Works Enterprises,” International Conference on theLearning Sciences, Lawrence Erlbaum Associates, Mahwah,NJ (1996), pp. 9–13.

17. S. Papert and M. Resnick, Technological Fluency and the Rep-resentation of Knowledge, proposal to the National ScienceFoundation, MIT Media Laboratory, Cambridge, MA (1995).

18. I am adopting the convention of capitalizing the word“School” when referring to School as an institution contain-ing the prevailing mind-set around organization, process,learning, and teaching.

19. The Suksapattana Foundation was created by Thai MITalumni in honor of the fiftieth anniversary of the King of Thai-land’s ascension to the throne. They procured funding andcoordinated a number of socially beneficial projects in honorof His Majesty the King of Thailand.

20. Office of the National Education Commission, Education inThailand 1998, Office of the Prime Minister, Kingdom of Thai-land, Bangkok, Thailand (1999).

21. This paper will only describe some of the activity within one

village learning center. For more detail, please see Refer-ence 1, above.

22. The New York Times (January 21, 1997).23. J. Sachs, J. L. Gallup, and A. Mellinger, Geography and Eco-

nomic Development, presented at the Annual World BankConference on Development Economics, July 1998, WorldBank Group (1998).

24. While I do not have strong evidence for this belief, I base thisstatement on a number of conversations with villagers andothers who worked with them.

25. I mean this in the original, positive sense of “hacking” wherethe term signifies informal and creative engineering exper-tise, and not someone practicing malicious destruction.

26. This is the term in use in Thailand and so I adopt it, but pre-fer the idea of “indigenous technology” as it is respectful andnot pejorative.

27. S. Phongsupasamit and J. Sakai, “Studies on Engineering De-sign Theories of Hand-Tractor Ploughs,” Proceedings of theEleventh International Congress on Agricultural Engineering,Dublin (1989), pp. 1617–1626.

28. S. Papert, Mindstorms: Children, Computers, and PowerfulIdeas, Basic Books, New York (1980).

29. C. Levi-Strauss, The Savage Mind, University of Chicago Press,Chicago (1966).

Accepted for publication July 10, 2000.

David Cavallo MIT Media Laboratory, 20 Ames Street, Cambridge,Massachusetts 02139-4307 (electronic mail: cavallo@media.mit.edu). Dr. Cavallo is a research scientist in the Epistemology andLearning Group at the MIT Media Laboratory. His work focuseson the design and implementation of new learning environmentsand educational systems, at both the micro and macro levels. Healso works on the design of new technologies for learning andthe role of technology in facilitating the change process. His cur-rent projects are in Thailand, Brazil, Colombia, and a juvenilejail in the United States. Prior to his work at MIT, Dr. Cavalloled the design and implementation of medical informatics as partof a reform of health care delivery and management at the Har-vard University Health Services. In addition, he has served as aprincipal and consulting software engineer at Digital EquipmentCorporation’s Artificial Intelligence Technology Center, wherehe led enterprise architecture and implementation efforts for us-ing technology to produce change in the processes and opera-tions of major companies by focusing on learning. He designedand built numerous knowledge-based systems for industry, mostnotable among them a set of intelligent simulations for trainingair traffic controllers. Dr. Cavallo completed his Ph.D. at the MITMedia Laboratory under the direction of Seymour Papert. Healso holds a Master of Science degree from the MIT Media Lab-oratory and a Bachelor of Science degree in computer sciencefrom Rutgers University.

IBM SYSTEMS JOURNAL, VOL 39, NOS 3&4, 2000 CAVALLO 781