Energy Babble: Mixing Environmentally-Oriented Internet Content to Engage Community Groups

Energy Babble: Mixing Environmentally-OrientedInternet Content to Engage Community GroupsWilliam Gaver1, Mike Michael2, Tobie Kerridge1, Alex Wilkie1, Andy

Boucher1, Liliana Ovalle1, Matthew Plummer-Fernandez1

1Interaction Research Studio 2Department of Sociology and Social PolicyGoldsmiths, University of London University of Sydney

[email protected] [email protected]

ABSTRACTThe Energy Babble is a kind of automatedtalk-radio that is obsessed with energyand the environment. We developed itwith, and deployed it to, a number ofexisting ‘energy communities’ in the UK.The system gathers content from avariety of online sources, includingTwitter™ feeds from the communities,from governmental departments, and fromthe National Grid, and chats about itcontinually using a number ofsynthesised voices interspersed with avariety of jingles and sound effects.Designed to playfully reflect andcomment on the existing state ofdiscourse and reports of practice in theUK, the Babble can be considered both asa product and as a research tool, in whichrole it worked to highlight issues,understandings, practices anddifficulties in the communities withwhom we worked.

Author KeywordsResearch through design; environmentalHCI; ludic design; STS; energycommunitiesACM Classification KeywordsH.5.m. Information interfaces andpresentation (e.g., HCI): Miscellaneous.General TermsDesignINTRODUCTIONIn the UK, the government has committedto a 34% cut in carbon emissions,compared to 1990 levels, by 2020, and an80% cut by 2050, despite a predictionthat global energy demand will double by2030 [15]. This ambitious target isbeing approached in a number of ways,including the pursuit of low-carbonenergy production and new approaches tomanufacturing. In particular, asignificant proportion of the saving isexpected to come from reductions indomestic energy use, and thus ‘energydemand reduction’ is a strong theme incurrent government policy, includingpolicies around research.This stance is congruent with recenttrends in ‘environmental HCI’. On theface of it, the rationale for HCI’sinvolvement in energy demand reductionis simple and convincing: Demandreduction can be supported by realtimefeedback about actual energy use,according to this logic, coupled with

Permission to make digital or hard copies of allor part of this work for personal or classroomuse is granted without fee provided that copiesare not made or distributed for profit orcommercial advantage and that copies bear thisnotice and the full citation on the first page.Copyrights for components of this work owned byothers than ACM must be honored. Abstracting withcredit is permitted. To copy otherwise, orrepublish, to post on servers or to redistributeto lists, requires prior specific permissionand/or a fee. Request permissions from

persuasive computing approaches toencourage people to reduce their demand.Since sensing, feedback, informationvisualisation, persuasive computing anddomestic applications are all coreissues for the HCI community, it seemsapparent that it could play a leadingrole in supporting energy demandreduction.Over the last several years, however,the potential for simple energy demandfeedback to reduce consumption has beenchallenged in a number of ways. First,evidence suggests that interventionsproduce only small and short-livedeffects [1], and are prone to ‘boomerangeffects’ in which above-averageconsumers may reduce their energy use,while those below-average ones increasetheirs [10]. Broader concerns are raisedin critiques from the social sciences.For instance, focusing on reducingresource consumption (including energy)tends to overlook the culturallyembedded practices within whichconsumption takes place [12] andportrays the consumer as ‘resource man’[14] capable of and concerned withreducing consumption as an end initself. Dourish [5] suggests thatfocusing on individual behaviour changecasts people as individual rationalactors in a classic new-liberal way,while Brynjarsdóttir et al. [2] linkpersuasive approaches to modernism andits limitations and also review andidentify the shortcomings of manyenvironmental HCI projects. In sum, despite the apparent promise ofresource demand monitoring, the approachseems constrained by several factors.First, addressing individual orhousehold consumption overlooks largersocial realities that can promote orinhibit sustainable practices. Second, anarrowly instrumental approach towardsreducing energy consumption isinsensitive to how energy use is part of

situated practices that people areunwilling or unable to change. Finally,insensitivity to diverse settings tendsto lead to one-size-fits-all designsthat may not fit local circumstances.Energy, Design, STS and CommunitiesHow can we go beyond individualistic,utilitarian and modernist approaches todesign-led research on energysustainability? The approach wedescribe here was conditioned by threefactors that were significant in theproject’s framing and pursuit. First,the design work was influenced by ourwork on ludic design [6], which promotesplayful engagement with ideas to offer awide scope for interpretation [11], aswell as our previous research usingspeculative design as a form of publicengagement with science, in whichdesigners engage with ‘expertpractitioners’ to explore possiblefutures for emerging technologies [8].Second, the project was undertakenthrough collaboration between Design andScience and Technology Studies (STS),with leaders representing eachperspective, and team members havingvariable expertise in either one or bothdomains. Explicitly linking design workto STS emphasised an appreciation of,and engagement with, a heterogeneous mixof actors at variable scales; moregenerally it reflected and resulted in ashared concern with how technologies areshaped by social actors, and converselyhow social realities are shaped bytechnologies. A third factor important in the framingof the project came from the fundingprogramme for which we shaped a proposaland which ultimately funded theresearch. Jointly supported by the UK’stechnology and social research agencies,the ‘Energy and Communities’ programmeemphasised the potential importance of‘community initiatives’, suggesting that

‘change in energy consumption is… notsimply dependent on behavioural changeat the individual level but on the co-evolution of communities and theemerging technology that is available’[15]. In addition, the programme wasarranged to partially and semiformallycoincide and communicate with anotherinitiative, the Low Carbon CommunitiesChallenge (LCCC), which funded the work of22 ‘test communities’ to reduce carbonuse in various ways. This meant projectsfunded by the ‘Energy and Communities’programme were supported and encouragedto engage with a variety of ‘energycommunities’ spread across the UK. The Energy and Communities programmethus provided a timely opportunity forus to explore environmental issues froma mixed STS and Design approach that welabelled ‘Ludic Action Research’ [17].We approached the project with acommitment to exploring and queryingcertain assumptions and reified notionsbuilt into the call e.g. what counts asenergy-practice, what are the issues andactors at play, indeed what counts ascommunity. This approach also sensitisedthe team to an expanded view of‘involvement’ in the design process(e.g. [3], [16]). In the rest of this paper, we describethe trajectory the research took inengaging with communities via a numberof design-led methods, conceptualisingand implementing the Energy Babble,deploying it into the communities, andassessing their responses. Finally, wereflect on the results of our study tounderstand the nature of the Babble bothas a product and a research tool, onethat illuminated the conditions anddifficulties faced by the energycommunities with whom we worked.PROCESS: FROM THE COMMUNITIES TO THE BABBLEAt the outset of the project we had noplans for the sort of device we might

build; instead we wanted this to emergefrom our interactions with thecommunities. In this section we describehow we engaged with members of a numberof different energy communities from theUK, the design-led methods we used toshape our relationship with them, thedesign explorations that eventually ledto the Energy Babble concept, and how werefined the concept through softwaredevelopment and product design.Forming a Network of Communities Soon after the project was launched, webegan to form a network with members ofseveral energy communities around theUK. We met most of these at an eventarranged jointly by our sponsoringresearch councils and the LCCC tointroduce the initiatives and peopleinvolved in them to each other. Also atthis event, we met representatives ofthe Department of Energy and ClimateChange (DECC) who later took part insome of the project’s activities. Duringthe meeting, a number of LCCCparticipants expressed an interest inour project. We followed up theseinitial encounters with short visits toeach site, during which we met with keymembers of the groups and toured theareas to see some of the projects theywere pursuing. The seven ‘communities’ with whom weworked1 were by no means coextensivewith their geographical settings.Typically involving small numbers ofcommitted members who see themselves aspractitioners, innovators and transitionadvocates, they form and reform inresponse to policy and funding changes,seek to enrol the communities of theirvillages or towns or cities, and agonise

1 The groups were: Energise Hastings;Greening Goldsmiths; Low Carbon LivingLaddock; Meadows Partnership Trust;Reepham Green Team; Sid Valley EnergyAction Group; Transition New Cross.

over how they will involve more peoplein their plans. Some were extremely wellorganised and funded, while others werevery ad hoc and quiet. Their engagementwith and characterisation ofenvironmental issues varied as well,with key issues for them ranging fromglobal warming to fuel poverty, and fromrenewable energy to conservation.Design-Led Community BuildingThe initial networking meeting waseffective at introducing communitygroups and academic researchers to eachother. It also revealed potentialtensions between the community groupsand academic researchers, however. Indiscussions, it became clear that theenergy communities questioned, and mightresent, funding given to research thatwas not primarily aimed at producingdirect environmental benefits,particularly given the government’schanging commitments to renewable energyand sustainability. In addition, most ofthe community groups were time andresource poor, and had already beensubject to a great deal of scrutiny bothfrom researchers and policy makers overthe course of their operation. Now wewere asking them to volunteer more timeand energy to working with us. Not onlydid they appear somewhat jaded by theattention they had already received, itseemed that they had formed well-developed personas in presenting theirwork and beliefs.

These considerations influencedactivities we designed to get to knowmore about the communities and to bringthem together with each other and ourresearch team. Most significantly, weheld an ‘engagement workshop’ at aLondon museum, with participantsincluding members of each community, andalso other interested people includingrepresentatives from the fundingagencies and the Department of Energyand Climate Change. Engagement WorkshopFor the workshop, we designed severalplayful and engaging activities,inspired by Cultural Probes, for groupwork. For instance, to exploreparticipants’ expectations and fearsabout the future, we asked them to worktogether to complete partially-designedfront pages for imaginary periodicals,offering templates for genres such as atabloid newspaper, a more seriouslooking broadsheet newspaper, and apopular science magazine (Figure 1a). Toinvestigate people’s emotionalengagement with energy practices, weasked them to draw and use rubber stampsto describe the trajectory of a recentevent, which resulted in annotateddiagrams describing activities rangingfrom getting up in the morning tokilling a chicken to serve at a dinnerparty (Figure 1b). To investigatecommunities both as they currentlyexperienced them and as they imaginedthey might be, we gave them a

Figure 1. Engagement workshop tasks: a) imaginary periodicals; b) emotional trajectories;

c) mapping communities.

diagrammatic map and a large number ofstickers, asking participants to worktogether to create a fictionalamalgamation of their communities(Figure 1c).These activities occasioned a great dealof discussion that was informative aboutthe participant’s communities andenvironmental work. The primaryachievement of the exercises, however,was to configure the participants’relationship with each other, and withus as researchers. By inviting theparticipants to engage with tasksintended to provoke creative explorationrather than to return clear socialscientific data, we seemed successfullyto encourage participants to forgo moreformalised forms of self-presentationand to engage with us in a more personaland relaxed manner. In addition, thetasks worked to expose them to ourmethods and approaches as designers, andto set the tone for the designactivities we pursued later.Cultural ProbesAt the end of the engagement event, wegave each participant a number ofCultural Probes kits to work withthemselves and to distribute to othermembers of their communities. Each kitcontained several items, e.g.: A poster for writing ‘energy rules’,

with prompts (“It’s strictly…”, “Never…”, “Always…”)

A form for writing an obituary for anappliance

An unfolded envelope with a request to “confess a guilty energy use secret” before sealing it for privacy

Cards with two devices (e.g., a clothes iron and a wind turbine) withblank speech bubbles to be filled

Participants returned the probes afterseveral weeks. Again, part of thepurpose of the probes was to encouragean informal and playful relationship

with the participants. This seemedeffective in encouraging returns thatwere revealing of the participants’attitudes and conflicts around energyuse. Some (e.g. “it’s strictly energy use thatdefines us”) indicated that participantswere zealous about sustainability.Others (e.g. confessing to “leaving the dooropen and keeping the heating on”) suggested acapacity for self-deprecating humourabout the contradictions in their actualpractices. In general, the Probes seemedto allow us, and perhaps ourparticipants, glimpses of thecomplications involved in balancingeveryday life with a commitment towardsenvironmentalism. Our view of these complications wasreinforced by our research into currentenergy policies and the sustainabletechnologies that the communities wereusing. We also attended energy events inthe communities, where we were struck bythe many different voices, affiliations,concerns and agendas in which thecommunity actions were entangled. Thesediscussions, the workshop, probes andobservations all set the scene for ourdesign work.

From a Design Space to the BabbleAs our perceptions of the energycommunities and their engagements beganto settle, we started to work on ideasfor the systems and artefacts we mightmake for them. As with many of ourprojects, we developed numerous ideasthat we captured in the form of simpleevocative proposals captured in aseries of design workbooks [7]. Here weoutline a few to indicate avenues weexplored (Figure 2):• Energy tourism. Increasing concernsabout sustainability brings greatervisibility to previously overlookedenergy infrastructures (e.g. powerstations, transmission networks). Couldnew or obsolete energy infrastructurebecome sites for new forms ofsightseeing?• Insistent activism. Discourse aboutsustainability is often channelled intopredictable forms (e.g. communitymeetings, energy exhibitions, etc.).This may make such concerns easier tohold separate from day-to-day life.Could we design systems that wouldinsist on such issues even in seeminglyinappropriate settings?• Energy awareness The desirability ofsupporting energy communities to sharetheir expertise and successes wascentral to the funding programme.Several proposals explored ideas bothin terms of functions and in terms ofthe socioaesthetic experiences theywould afford.

Our proposals evolved over a period ofmonths, and both reflected and spurreddiscussions amongst us about which bestreflected our insights into thecommunities, what might engage thecommunities, and what we wanted toachieve. These culminated in the firstproposal for an ‘Eco Babble’ (Figure3). This was conceived as a system thatwould gather input from a combinationof web scraping and participant inputto a central server, and organise thecontents in multiple ways for replayfrom a variety of public and domesticdevices.The Two Faces of the BabbleFrom the very beginning, and throughoutits development and deployment, theBabble was motivated by a mixed andsomewhat incommensurable set ofintentions. At the time it was firstsuggested, our discussions did not tendto centre on what we wanted the systemto achieve or how the participantsmight engage with it. Instead, wedecided to develop it because we felt akind of satisfaction in being able toreflect back to the participants, in asomewhat mischievous way, thecomplicated mix of discourse,priorities and suggestions with whichwe saw them contending. As we developed the system, however, amore utilitarian perspective on itspossibilities grew, as reflected by awritten brief we developed to guide thework:

Figure 2. Proposals: a) energy tourism; b) activist meter reading; c) energy awareness

Design an Energy-Babble system thatdisplays material, collected fromsome combination of individual,community and public sources, toopen and promote constructive affectand involvement in energy reductionissues and orientations. Morespecifically, the system shouldsupport understandings of, andpractices related to, energy demandreduction.

Thus on the one hand, the Babble wastalked about as a tool for informationand communication about practices andunderstandings related to energyproduction and conservation. On theother, it was a system for portrayingand even parodying the complexities ofenergy-related discourse. Though seldomvoiced, these mixed perspectives on theEnergy Babble coloured its development,surfaced more explicitly later as wedescribed it to the community members,and ultimately inform our discussionhere of how volunteers related to thesystem and how we understand theproject. To foreshadow our discussion, wesuggest that insofar as the Babble waspresented and viewed as an informationand communication device, it tended toserve and be assessed as a product. As areflection of energy-related discourse,in contrast, it served as a research toolthat helped us, and ideally theparticipants, better understand theexisting situations within which energycommunities work. The correlationbetween utility/products andreflection/research is neither perfectnor necessary; nonetheless,distinguishing two ‘faces’ of theBabble—one an instrumental product, theother a reflective research tool—isuseful in orienting to our intentionsand results.

THE ENERGY BABBLEMonths of work separated the decisionto pursue the Babble and its eventualproduction as 28 fully workingprototypes. We do not focus on thedetailed design process here bothbecause of space constraints andbecause we want to explore thecommunities’ engagement with thefinished system and what this revealsboth about the design and thecommunities (though see [18]. Instead,we fast-forward here to a descriptionof the completed design, discussingonly a few relevant aspects of ourprocess.The DevicesThe most visible manifestation of theEnergy Babble is the devices that wedistributed to participants (Figure 4).These comprise a combination of twocustom blown-glass components set ontop of an injection-moulded base. Thebase houses the device’s electroniccomponents, including a Raspberry Pimicrocomputer, various actuators, and acompact loudspeaker. The central glasspiece serves to amplify the output ofthe speaker located beneath it—glass’sacoustic properties make it optimal forthis function—while second glass piecesupports it and holds a microphoneattached to the base by a coiled cable.A prominent yellow knob rotates tocontrol volume; if turned beyond itslowest setting it plays a recordeddescription of the system and itsoperation, using the same synthesisedvoices as the rest of the system.The product design of the devices wasthe result of a great deal ofexploration and development. Startingwith explorations of birdhouses andnests (to develop the theme shown inFigure 3c), attention turned to avariety of containers, and ultimatelylabware such as test tubes, beakers,and retorts. The final result expresses

the enclosure and release of data viathe glass cone in a way that webelieved to be both intriguing andharmonious for everyday settings suchas homes or public spaces. The audio design of what is primarily asound-based device also received agreat deal of attention. Much time wasspent reviewing off-the-shelf voicesynthesisers for understandability andaesthetic acceptability, crafting aseries of jingles meant to evokecontemporary music as well as newsbroadcasts, and designing the audiointroduction to reduce the need forwritten documentation. In operation,the Babble plays a varying stream ofmessages spoken by a number ofsynthetic voices, and punctuated byshort musical jingles reminiscent ofthose used in news broadcasts. Theseare sent to the devices via theinternet from a server located in ourstudio. The SystemThe audio stream played by the Babbledevices is algorithmically constructedand represents the results of a numberof modular processes. The first set ofthese seek new information from theweb:• Twitter Scrape The system scrapesmessages from a variety of Twitter™feeds, including those from thecommunities, from DECC, and from theNational Grid.• Twitter Search The system searchesTwitter™ using a number of terms (e.g.'climate change', ‘'energy bills','renewable energy') to find relevantcontent more broadly.• URL Following If returned tweetscontain ‘http’ and have been retweeted,the system follows the link and scrapesthe first paragraph of text for output.

• Energy Reports The system also tracksthe National Grid, broadcastinginformation about current energy demandand carbon intensity, as well aschanges in the amounts of energygenerated by coal, gas, and wind.• Switched Off The… reports hourly a listof things people tweet about havingswitched off (lights, heat, theinternet).A second set of modules accepts inputfrom participants, or manipulates theexisting content.• User Input Users can input their ownstatements via the microphones on theBabble devices. Voice recordings aresent to the server, which passes themto a commercial service for speech-to-text translation, allowing them to beresynthesised both to anonymise inputsand make them congruent with theoverall audio ecology. Text can alsobe entered directly via SMS textmessaging.• Prompt Bot We developed a string ofgeneric questions (e.g. ‘how do yousave energy?’) that were read outoccasionally by the system to promptparticipant responses.• The Voice of Babble Markov algorithms arerun on the content scraped from URLs,and separately on user inputs, toproduce new text strings. Markov

Figure 4. The Energy Babble device

algorithms find word transitionprobabilities from a corpus; theseprobabilities can be used to generatenew strings of words. The results tendto be locally sensible, but drift overtime in a way we hoped would producenew ideas or at least be amusing. The system orchestrated the content ina variety of ways. For instance userinput was played as soon as possible,while energy reports only appearedhourly. Different voices were assignedto different streams of content, andcontent was repeated periodically.Distinct jingles signalled differentstreams of content. Finally, the outputstream was arranged to be continuousbut uneven, with occasional gapsfollowed by bursts of several messages.The Babble ExperienceThe experience offered by the BabbleDevices, once they are properlyconnected to the server andoperational, is something likelistening to a continuous talk radioprogramme. A series of voices makestatements, read news items, askquestions and report energy use, allinterspersed with occasional musicalinterludes and lapses into nonsense.The majority of content is related toenergy and the environment and thus thedevices present themselves as stronglyfocused on sustainability, though afair amount of ‘off-topic’ content alsocreeps in from Twitter™, from followinglinks, or from participant inputs. Ourquestion was how our participants, allcommitted to environmental concernsthemselves, would engage the EnergyBabble.LIVING WITH THE ENERGY BABBLEWe deployed a total of 21 EnergyBabbles to members of the communitiesin a series of meetings at theirlocations (Figure 5). Each communityreceived 3 or 4 devices, which wereusually given to volunteers present at

the meetings, though in a few casesextras were left for laterdistribution. The remaining 5 Babbleswere distributed to team members, with2 going to people more looselyconnected to the project. Volunteerslived with the Babble for varyingperiods averaging about six months.In the rest of this section, we brieflydescribe what our participants told usabout their experiences with theBabbles. The majority of reports comefrom discussions when we deployed thedevices, or several months later whenwe revisited the communities to pick upthe devices. Others come fromdocumentaries by an independentfilmmaker hired to help us assess thefield trial. Initial Expectations and ImpressionsWe packaged the Energy Babbles,associated documents and equipment incustom-made cardboard boxes fortransport. During deployment events,these were usually positioned visibly,but unopened, during an introductorypresentation in which we reviewed theproject. Then we would unpack a Babbledevice and describe how it worked.Because it took some time to set up oneof the devices for demonstration,during these initial descriptions thegroup had not yet heard the system.Typically, then, initial comments andquestions revealed a mix ofassumptions, expectations and responsesto the devices.

Initially, many participants expectedus to produce a tool that woulddirectly help them reduce energyconsumption — or as G, from theMeadows, put it: ‘We thought we were going toget a gizmo to save energy’. When it becameclear the Babble did not serve thispurpose, they looked for otherutilitarian pay-offs. In Hastings, forinstance, an engineer asked ‘How does thisimprove the social operational wellbeing of thepeople who use it? If I make an investment how doget a payback?’ and explained ‘I wanted it tosolve a problem’. These discussions tendedto encourage the news/communicationinterpretation of the device, whichmollified many skeptics. For instance,the engineer realised it could be usedto broadcast the energy output of theirrenewables: ‘Babble could bring thisinformation to people’. Similarly, inLaddock a group member championed theBabble for using the British RaspberryPi technology, and because he sawpotential for it to broadcast his carbattery operated DIY domesticelectricity system.Building on this, in several of thegroups volunteers saw potential valuein the Babbles as a kind of marketingtool for promoting their groups andenvironmental concerns more generally.The Hastings engineer, for instance,described broadcasting energygeneration figures as ‘a very powerful salestool’. In Sidmouth, the groupspeculated about deploying the Babblein a local energy shop, or using it asa recruitment platform at anAlternative Energy Vehicle show. InReepham, the group decided that one ofthe devices should be free to roam,initially to the Post Office and laterto a variety of environmental events. Some people were happier to relinquisha utilitarian interpretation of theBabbles during the deployment events.For instance, after listening to the

device during the Meadows deployment, Ddecided that they would name theirBabble ‘Finnegan’, in a reference toJames Joyce’s Finnegans Wake. Sheexplained that this was because theoutput is like ‘a stream of consciousness’.In New Cross, J sent an SMS messageafter her Babble started working: ‘It’samazing! I love it so much already. The messagingsystem reminds me of the barbed wire telephonesystem in Wild West. Seriously - Google it. Thanksguys. :) ‘Installation and AccommodationInstalling the Babbles involvedconfiguring the devices to local routersettings, dealing with security, andsetting it up to communicate using therouter’s wireless network. In manyinstances this proved unproblematic,but in some cases, including deploymentevents, it proved more difficult. Whilenone of the problems we encounteredwere insurmountable, they seemed todemonstrate to potential volunteers thepossible inconveniences of borrowing aBabble. More serious problems arosewith some of the devices we leftbehind. For instance, in Reephamproblems with a local firewallprevented the Babble from beinginstalled in a local primary school.Other devices were borrowed but neverinstalled, possibly because of theperceived difficulty of set-up.Pragmatic issues were salient even for

Figure 5. A deployment event.

imagined deployments: for instance, inSidmouth ideas for showing the deviceat the Alternative Energy Vehicle showinvolved thinking about powering it viaa car with solar panels, and achievingmobile internet access.Visual and Auditory AestheticsThe Babble has an idiosyncraticaesthetic that extends from itsphysical design to its auditory output.Most of the volunteers found thisappealing. For instance, in Ladock, Jsaid she and her husband appreciatedthe Babble because it was ‘nice, funkylooking thing’. In the Meadows, P, an arttutor, said it was a ‘really nice object’ andthat the microphone was a ‘lovely visualelement’. This appreciation was mixedwith some bemusement, however. Severalpeople remarked on it having a ‘retro’appearance, or as J from New Cross putit, the Babble looked ‘like my gran couldhave had one’. It also was compared tokitchen appliances. For instance G fromLadock, told us 'it looked like a food processor…a bit quirky', while G in the Meadows toldus that visitors to his home usuallyasked jokingly why he had a blender inhis living room. The glass elements,too, attracted a mixed reception: Rfrom Reepham described the Babble as‘beautifully made in hand blown glass', but itsfragility was a worry for the librarianof a secondary school where it wasinstalled.Despite some initial concerns, wereceived no complaints about thesynthesised voices used by the Babble,and several people remarked favourablyabout their clarity. On the other hand,the fact of it being an audio devicecould be disruptive. G, in Ladock, toldus he had moved it from the kitchen tothe living room because it was a'conversation stopper’ for the family. L,in Reepham, the mother of a young baby,found it ‘annoying and noisy… not really for ahome’. R, her husband, told us he had

relocated it to his office because ofthis. In the Meadows, G said the Babble‘annoys everyone’ so he turned it on whenhe was alone.The intermittent nature of the outputcould also be unsettling. ‘A few times itfrightened the living crap out of me!’, J in theMeadows told us. She elaborated thatonce when the office was completelysilent at around 9pm, the Babble hadgiven her a fright when it unexpectedlycame on. She also complained that it‘didn’t talk on cue’ when she showed it tovisitors, and (like other volunteers)would have liked to be able to replayinteresting outputs.Babble as a Source of InformationVolunteers often oriented to theBabbles as a potential source ofinformation. This is not surprisinggiven that the audio was designed inthe style of an automated newsbroadcast, that many of the volunteersshowed a propensity to seek utilitarianexplanations for the device, and thatwe tended to encourage theseexplanations to reassure them about thesystem, rather than foregrounding adescription and defence of the Babbleas a playful, reflective system. By the end of the field trial, however,many volunteers expresseddisappointment in the Babble as asource of environmental news. InLadock, G told us that he did hearinformation ‘which was interesting’, butexplained that he did not follow manyalternative sources of environmentalnews. In contrast, his colleague J toldus that she welcomed the Babble as asource of new information, but‘disappointingly, not as much as I hoped’. ‘Itseemed a bit sparse,’ she explained, and‘very repetitious’. In Reepham, Rsubscribes to DEC emails that he looksat in the morning: ‘if they’re about somethingI’m interested in I read them’. He told us that

the Babble never provided relevantinformation of which he hadn’t beenaware.A recurring theme in discussing theBabble was that too much of its outputwas irrelevant. J, in Ladock, forinstance complained that there was ‘a lotmore of the jumbly stuff and less of the straightstuff’. ‘I tend to be on the serious side', sheexplained, and 'definitely the balance waswrong'. Considerations such as these ledto suggestions for filtering theoutput. For instance, R in Reephamspeculated that the Babble could bemore like a radio: ‘you might have onestream about transport, another about food, aboutheating the home…’, describing the resultas ‘far more relevant’. To our suggestionthat mixing streams might supportserendipitous connections he wasdoubtful: ‘people’s attention spans are gettingshorter’, he explained, so they would getbored before putting things togetherthemselves. J, in the Meadows, alsosaid that the Babble gathered too muchirrelevant information and that itneeded a ‘filter’ to focus on reportingnews about communities and government,‘rather than oil and gas’. G, also from theMeadows, echoed this, suggesting theBabble could be an app with buttons toswitch on and off channels ofinformation—a ‘filtering mechanism’.There was no clear consensus acrossvolunteers about which streams ofinformation were worth hearing,suggesting that the ability to selectamong them would lead to a moreutilitarian design. In the Meadows, forinstance, P found the energy reportsfrustrating as they didn’t meananything to him, while G, interviewedseparately, said they were ‘really reallygood’. In New Cross, P reported that shecouldn’t follow the ‘technical information’,referring both to the energy reportsand the information on renewable

systems. For her, ‘you can connect more topersonal comments, to the emotional side of energy’

Babble as a Medium for Communication Volunteers had mixed reaction to theability to input and hear commentsusing the Babble’s microphone and SMSfacilities, and this was reflected inthe relatively few messages they lefton the system (about 35 over 5 months).There was an evident reluctance toenter messages. During the first weeksof living with Babble, for instance, Jin New Cross made a few contributionsusing the microphone. Eventually herreaction became more of an ‘internalconversation’. She reported that when shereached for the microphone she feltnervous about saying somethingimportant to the system. Similarly, Rfrom Reepham told us that he didn’tinput much because he has ‘controversialviews’ and didn’t want to ‘upset anyone’.He recounted how he had heard somethingon the Babble that he disagreed with,but refrained from expressing his viewsbecause he considers them quitecontroversial. ‘I was aware that DECC might belistening, I want to come across as quite conservative,you don’t know who is listening’. In theMeadows, G was concerned to prevent hisdaughter from saying silly things intothe microphone (when asked what thosemight be he responded ‘are there any fithotties out there?’). Nonetheless, manycomments diverged from clear relevancefor environmental concerns, and G wasannoyed by messages he thought trivial.The reluctance to contribute to thesystem ran counter to appreciation forthe content that did appear. Forexample, J from the Meadows said shewould have liked to hear more from theother communities, especially ‘stories andtips on how they’re dealing with these issues’. InNew Cross, J enjoyed the comments: ‘youcan connect more to personal comments, to theemotional side of energy’. Conversely, G inthe Meadows would have liked to read

out his household energy use, while Pwould have liked to broadcast his solarenergy production; however there waslittle reflection about who theaudience for these figures might be.Finally, we had some indications thatthe lack of user inputs into the Babblereflected a lack of interest incommunicating with other groups moregenerally. R, in Reepham, was clearestabout this: he told us that while heoccasionally kept track of what othercommunities are doing, differingcircumstances meant that ‘what might not beright for them, might be right for us’. He mightcheck for good ideas but unlesssomething was ‘revolutionary’ there wasn’tmuch use in this. Equally, he likedtelling people what worked in Reepham,but described this as ‘reactive notproactive’—his group doesn’t proselytise‘the way Transition Towns do’.Appreciation for the BabbleDespite the lack of clear success forthe Babble as a utilitarian informationor communications product, all thevolunteers we spoke with were largelypositive about it. In part thisreflected appreciation for it as awell-finished, device that could fitthe home (Figure 6). In part, itstemmed from admiration for the Babbleas a novel technical device. InReepham, for instance, R found‘stimulating’ the way it uses audio ratherthan visual/text as a way ofencountering social media. In theMeadows, P speculated about extendingthe Babble’s technology, for instanceto automatically tweet about his solarpanels, or to nag him about his badenergy habits. Admiration for the Babble as a noveltechnical and aesthetic device blendedwith its value as something to show toother people. C from Hastings, forexample, was effusive about the Babble,

describing the novelty of the deviceand the attention it had garnered atwork, where she originally installedthe device, and at home, where she tookit later. R from Reepham described itas ‘a curiosity for visitors’ that he enjoyed tomembers of a number of otherenvironmental organisations with whomhe worked.Finally, several volunteers expressedappreciation for the Babble as a sourceof ambient awareness of environmentalaction. In New Cross, J told us it wasreassuring to hear evidence ofexpertise: ‘Thank God for people who know thetechnical bits, it’s strengthening to hear that there arepeople out there in charge’, and moregenerally that the Babble gave her asense of a larger community concernedwith environmental issues: ‘it makes youthink that you are not alone in thinking about savingthe world’. In the Meadows, J expressed amore abiding affection for the presenceof the Babble: ‘aw, I’ll miss him actually. It wasnice to have him on in the background, I’m used to itnow. Its quite aptly named, Babble’.Babble and Wider ConversationsThe accounts above all reflectdiscussions centred fairly closely onthe Babble system as a product. Whatbecame striking to us, however, was theway that our conversations with thevolunteers frequently opened from aninitial concern with the Babble’susability, functionality and aesthetics

Figure 6. Babble in a volunteer’s home

to encompass the broader and moreparticular issues, practices andcontroversies with which our volunteerswere living. Though these discussionsmay be of questionable relevance forassessing the Babble as a product, wesuggest that these conversations andthe insights they revealed can beviewed as an outcome of the Babble as aresearch tool.For instance, at the Meadows, during asuggestion that the Babble contentshould be filtered to focus oncommunities and government ‘rather than oiland gas’, J suddenly exclaimed ‘except thatBritish Gas are bastards!’, and conversationwith her diverged into lengthycomplaints about DECC’s lack ofsupport, British Gas call centres, andpigeon droppings building up undersolar panels. In Ladock, ourconversation with J about the Babblesoon expanded to include her complaintsabout the hurdles involved in securinggovernment funding for environmentalwork (‘we think they’re rubbish’), and thefrustrations of not being able to giveaway radiator backdrops, energymonitors and LED down lighters at anEnergy Fair she organized (‘it was a totalfailure’ that ‘didn't engage the people we set out toengage’). Also in Ladock, G described theirattempts to put up a new wind turbinethat was rejected by the council ‘onspurious grounds’. He attributed this to'about half a dozen' residents who spread'a lot of misinformation' about how the LowCarbon Living group were out to ‘line theirown pockets', culminating in ‘a minor punch-up'. Like J, he expressed frustrationat the difficulty in reaching out todissenters within the community(there’s ‘no forum to talk to those people’),and also with the government: ‘lots ofbusinesses are starting up then going to the wallbecause the government keeps changing the rules'.These complaints were mixed with pride

in the group’s achievements. Forinstance, he referred to a 'story' he puton the Babble about how on a sunny dayhe used his PV to charge his car andheat water: people were impressed thathe could 'drive 75 miles and have hot water forabsolutely nothing'. He concluded that ‘youcan't depend on the government to do things, youcan depend on the community to do things'.A notable theme that emerged fromseveral volunteers had to do with theentanglement of energy concerns withother issues. For instance, R inReepham told us he would be going toBuckingham Palace to be honoured forhis contributions to energy efficiency,but said that he’d like to berecognised for the work he does thatgoes beyond that. The Babble should gobeyond energy, he told us, to addressfuel poverty, transport poverty, andtake a ‘holistic’ view. “Energy is a key part ofit but the stories are stories about many otherthings’ he said, ‘It’s too sterile if you look atonly energy’. Similarly, in Ladock J told us that herhusband refuses to be involved with LowCarbon Living because he sees theirefforts as futile. 'We should be lobbying'she said, and mentioned the socialnetworking activist groups Avaaz and38Degrees as effective ('though Iunderstand their limits'). She also does workwith Christian Aid. 'You tend to see how it allfits together - the international aspects of climatechange'. For instance, when ChristianAid pointed out that climate changeharms the poorest first, she thoughtthey were off-topic, but then realisedit was true. This led her to realisethat 'how we treat the world and how we treatother people, they're all linked'. DISCUSSION: UNDERSTANDING THE BABBLETaking seriously the idea that theBabble played an important role insparking the intense discussions we hadwith our volunteers suggests that we

move beyond assessing the systemaccording to the utilitariancharacterisation of it as aninformation and communication product.Turning to the reflectiveinterpretation of the Babble instead,as a system that gathers and‘intensifies’ the existing state ofdiscourse around energy practices, maygive us another perspective on how thesystem worked as a research tool in ourmeetings with the participants, byserving as an independent actor thathelped shape conversations leading tobetter understandings of thecommunities and their concerns. A simple version of this account wouldsuggest that the Babble should beunderstood as a research tool that wassuccessful, rather than simply as autilitarian information/communicationproduct that was less so. Thedistinction between these roles is notclear-cut, however. The Babble wasnever seen purely as a prototypeproduct, either by the volunteers orourselves: we never planned to produceit commercially, and they were alwaysaware of it as part of a researchproject. The Babble was never solely aresearch tool either: it was offeredseriously for long-term use, andparticipants engaged with it not onlyto further their discussion with us butto engage with the material it offeredin its own right. The product andresearch-tool faces of the Babble areinterdependent. Here we discuss severalconceptual handles on how this might beunderstood.To start with, it is helpful toconsider the Babble in terms of theconceptual character of the ‘idiot’,who, in Stengers’ [13] account:

resists the consensual way in whichthe situation is presented and inwhich emergencies mobilize thought

or action. This is not because thepresentation would be false orbecause emergencies are believed tobe lies, but because "there issomething more important". Don’task him why, the idiot will neitherreply nor discuss the issue.... theidiot demands that we slow down,that we don’t consider ourselvesauthorized to believe we possessthe meaning of what we know (p.994)

From this point of view, the Babble canbe seen to act as an idiot within theenergy communities who used it (see[17] and Michael [9]), by confoundingexpectations of how technologies shouldcontribute to the communities’ work.This was evident both during thedeployments, when the Babble surprisedand confused volunteers who wereexpecting some sort of demand reductionmeter, or at least a clearlyutilitarian design (‘I wanted to solve aproblem’), and throughout the project, asvolunteers struggled to make sense ofwhat it was doing. Instead ofacquiescing to ‘the consensual way inwhich the situation is presented’, theBabble implicitly suggested that in theconfused flow of messages about energyuse, policy shifts, new technologies,and seeming irrelevancies “there issomething more important”. But what is that ‘something that ismore important’? The Babble never says,but given its output this might includekeeping in touch with emerging policy,sharing best practice, being aware ofenergy sources and demand, and joiningwith other communities—the veryconcerns identified as important by thefunding programme that supported theproject. But the Babble does this inthe most literal, even stupid, way, andthe volunteers resist it. They counterby insisting that policies are ever-changing and wilfully made difficult,

that what works for one community maynot work for another, that it isdifficult to find meaning in statisticsabout energy, and that there is limitedvalue in further contact with othercommunities. From this point of view,the roles are reversed: it is theBabble that presents the ‘consensualway in which the situation ispresented’, and the communityvolunteers who are cast as idiots,asking the Babble, and us, and thepolicy-makers, to slow down, because wedo not ‘possess the meaning of what weknow.’Our conversations with them at the endof the field trial, then, can be seenas reflecting their pent-up responsesto the obduracy of the Babble. Yes, theBabble may be right in saying thatthere are larger concerns at play thancan be addressed by energy demandmeters, but what is needed is notsimply more policy, more news, and morecommunications. On the contrary, theytold us, we need better filtering,better ways to talk about energy,better situated ways to communicate,and recognition that energy use issituated in a wider landscape of localand global issues such as inequalityand sustainability. And through this,they revealed their realities, helpingus to understand that these‘communities’ are shifting collectionsof people who constantly reconfigurethemselves, and who do extraordinarywork to negotiate changing policyopportunities and obstacles, to filterinformation about new technologies, toreach out within their own communities,and to understand when it is worthcommunicating more closely with others.In the end, the Babble might beunderstood in terms of DiSalvo’s [4]account of how design can play a partin constructing publics. FollowingDewey, DiSalvo suggests that publics

form around issues, and that design canparticipate in this by bringing issuesto prominence. He suggests two primarytactics for this: projection, in whichdesigns suggest possible futuremanifestations of current trends, andtracing, in which design is used to makeclear the history of currentsituations. To this, the Babble mightadd a third tactic: concentration, inwhich current accounts and discoursesabout an issue—in this case energy—arebrought together to form, not just aneutral representation, but a focusedstream that inundates listeners withthe many different and potentiallyincompatible ways that that issue isdiscussed, legislated for, measured andworried about. From this perspective, the Babble mightform ‘a public’ not just via the issuesthat comprise it, but the issues raisedby the incoherence of the babbleitself. Thus the Babble begins to pointtoward a public that emerges out of anoscillation between different local andcollective communities, variously incompetition and united, informed andfrustrated. Moreover, the discussionsoccasioned by the Babble suggests thatsupport for situated, local communitiesrequires better appreciation of themorass that the publics/communitiesmust negotiate. This includes thecompetition/lack of communicationbetween communities, as well as thecommonalities of being placed in arelation of competition by thestructure of government projectfunding. In highlighting these issues,the Babble may also help (re)configurea public of HCI researchers, fundersand policy makers to concern itselfwith these realities of energycommunities rather than, simply,technologies focused directly on energydemand reduction.

ACKNOWLEDGMENTSWe are deeply grateful to the energycommunities for their participation inthis research.This project wassupported by the Economic and SocialResearch Council’s award no.ES/I007318/1 and by the EuropeanResearch Council's AdvancedInvestigator Award no. 22652. REFERENCES1. Abrahamse, W., Steg, L., Vlek, C., &

Rothengatter, T. (2005). A review ofintervention studies aimed athousehold energy conservation. Journalof environmental psychology, 25(3), 273-291.

2. Brynjarsdottir, H., Hakansson, M.,Pierce, J., Baumer, E., DiSalvo, C.,and Sengers, P. 2012. Sustainablyunpersuaded. CHI 2012, 947-956.

3. Callon, M. (2004) The Role of HybridCommunities and Socio-TechnicalArrangements in the ParticipatoryDesign. Journal of the Center forInformation Studies 5(3): 3–10.

4. DiSalvo, C. (2009). Design and theConstruction of Publics. DesignIssues, 25(1), 48-63.

5. Dourish, P. 2010. HCI andenvironmental sustainability: Thepolitics of design and the design ofpolitics. DIS 2010, 1-10.

6. Gaver, W. 2009. Designing for HomoLudens, Still. In (Re)searching theDigital Bauhaus. Binder, T., Löwgren,J., and Malmborg, L. (eds.). London:Springer, pp. 163-178.

7. Gaver, W. 2011. Making spaces.CHI’11, 1551-1560.

8. Kerridge, T. (Forthcoming). DesigningDebate: The Entanglement ofSpeculative Design and UpstreamPublic Engagement with Science andTechnology. PhD thesis, Goldsmiths,University of London, London.   

9. Michael, M. (2012). “What Are We BusyDoing?” Engaging the Idiot. Science,

Technology & Human Values, 37(5),528-554.

10. Schultz, P. W., Nolan, J. M.,Cialdini, R. B., Goldstein, N. J., &Griskevicius, V. (2007). Theconstructive, destructive, andreconstructive power of social norms,Psychological Science, 18(5), 429–434.

11. Sengers, P. and Gaver, W. (2006).Staying Open to interpretation.Proc. DIS 2006.

12. Shove, E. (2004). Comfort,cleanliness and convenience. Berg.

13. Stengers, I. (2005). Thecosmopolitical proposal. Makingthings public, 994-1003.

14. Strengers, Y. (2013) Smart EnergyTechnologies in Everyday Life: SmartUtopia? Palgrave MacMillan.

15. UK Research Council (2010),Energy and Communities call forproposals.http://web.archive.org/web/20100813103416/http://www.esrc.ac.uk/ESRCInfoCentre/opportunities/current_funding_opportunities/Energy_and_Communities_Collaborative_Venture.aspx (accessed 22.09.14).

16. Wilkie, A. (2013) Prototyping asEvent: Designing the Future ofObesity, Journal of Cultural Economy:1–17.

17. Wilkie, A., Michael, M., &Plummer‐Fernandez, M. (2014).Speculative method and Twitter. TheSociological Review.

18. Wilkie, A., Michael, M.,Kerridge, T., Gaver, W., Ovalle, L,.DiSalvo, C., Gabrys, J., (2012).Design, STS and cosmopolitics: Fromintervention to emergence inparticipation and sustainability.Proc. EASST 2012.