Creative Imagination: Inside and Outside the Head

Creative Imagination: Inside and Outside the Head

When we are asked what we mean by “imagination,” what springs immediately to mind may be thoughts such as that of a small child creating vivid imaginary worlds peopled by one or more imaginary playmates, or of someone (young or older) who loves to engage in pretend or role play.  Or we may think of someone we know who can (almost magically) take seemingly unrelated ideas (characters, objects) and creatively interweave them together into a compelling story or picture.

Each of these are, indeed, clear examples of imagination.  But they’re all examples of only one sort.  They’re all bundled together under a more specific heading that a recent process-based exploration of imagination would call “expressive imagination.”  This form of expressive imagination – typified by such creative activities as storytelling, role-playing, and day-dreaming – emerges in a largely bottom-up way from an individual’s personal prior experiences and existing mental representations.  It’s imagination that springs from “inside the head.”

But there is another form of imagination, equally valued and valuable.  

Rather than emerging primarily from within an individual’s internal world of memory and mental concepts, creative imagination can also be focused outward, on the external world.  With an intense outward gaze, it is quietly on the lookout for patterns, relations, or connections in the external world.  Peering outward, this form of imagination – sometimes called “instrumental imagination” – often is purposefully directed toward specific problems.  

Let’s look at a recent research study that exemplifies how we might use both these forms of imagination.

The setting

The story begins in an interactive exhibit at a museum, variously visited by individuals, families, or groups. Set off in its own room is a large multi-touch tabletop, with glowing lights and wooden blocks of various sorts.  As we enter, we’re told to imagine that we are electrical engineers trying to help “fictional scientists in an uncharted aquatic cave teeming with never-before-documented species of bioluminescent fish.”  We can design and build glowing fishing lures using different colored LEDs.  If assembled correctly, a virtual circuit (a circuit with the correct ratio of resistors, batteries, and LEDs) will glow, attracting the fish out of the cave, allowing the scientists to identify and catalog them.  Each museum visitor can choose which of the many fish to try to lure into the light, and, although each visitor can see and talk with the others, each visitor’s block-assembling actions do not affect the actions of other co-visitors at the table. 

Some museum-goers start interacting with the blocks and, through experimentation and trial-and-error or experimentation in combination with prior learning, they find a way to configure the blocks successfully.  There’s a sudden bright glow and a fish emerges from the darkness of the cave toward the light for everyone to see.  But other museum-goers have trouble finding a configuration of the blocks that will work. They try this and that, and that and that, without success.  What happens then?  What happens when it seems that failure is facing us?

This was exactly the question the researchers of the exhibit wanted to answer.  When failure seemed to be looming large and the way forward was not clear, what patterns of interaction –with the blocks or of the museum visitors with each other – could help get over the hurdle that obstructed them?  What actions would impel them forward, enabling them to transition from unproductive, frustrating, and unfruitful attempts, to a productive and successful approach?

To answer this question, the researchers videotaped visitors’ interactions at the tabletop using three unobtrusive cameras and an audio recording.  (A sign outside the room indicated when videotaping was taking place, so participants could choose to enter during recording or enter at a different time.)  The actions of 3,546 participants were recorded, leading to more than 47,000 separate actions.  But that presented its own challenge:  What to do with that massive amount of data?  How could it tell us anything about which actions led from frustratingly unproductive to rewardingly productive search and experimentation?

Finding patterns

And here is where the research team put together some powerful pattern-detecting methods.  First, they developed a systematic way to keep track of all of the circuits that each visitor made.  For example, if a visitor arrived at the table and made a complex circuit with many components that did not work, but it was their first attempt at that type of circuit, and no one else at the table had tried anything like it, it would be coded as “CNUO” (complex, not-working, unique for them, and original to the table). If another visitor arrived, and made a simple 3-component circuit that worked, and it was the first time they had made it, but it followed the same configuration as that of another visitor who was at the table during the same time, this would be coded as “SWUE” (simple, working, unique for them, and an echo of someone else’s circuit).  

This coding scheme allowed the researchers to develop what is called a “Hidden Markov Model” (HMM) to predict when a visitor was likely to move from an unproductive circuit-making state (when they were making a circuit that did not work) to a productive one.  Using this model, they could tell that once a visitor reached a productive state (with one working circuit), they most often continued to generate other circuits that were also working.  But if a visitor instead transitioned from a productive state to an unproductive state, they very rarely returned to a productive state.  That is, if a visitor fell into an unproductive state, they tended to remain there, until leaving the exhibit.  

But still, a few visitors did go back to making productive circuits.  What was different about the visitors who did get over the hurdle, from the many others who never managed to get unstuck?

Getting past the hurdle of failure

To answer this, the researchers first used the Markov Model to create a list of all the participants who moved from a sequence of three or more unproductive circuits – suggesting a sustained and persistent exploration of the problem – to a productive one.  Out of all 3,546 participants, only 204 participants (less than 6% of all participants) showed this pattern of getting across the hurdle from a series of unsuccessful attempts to a successful one.  

Next the researchers zoomed in on 22 such instances, all from one day of the visitors’ interactions.  They now applied another more detailed and contextually-enriched coding scheme to capture exactly what participants were doing at each point.

What they found is that in the great majority of cases, leaps forward came after a “stuck-in-a-rut” visitor stopped to watch how other visitors at the table were configuring their blocks (75% of the instances) or in which the “stuck-in-a-rut” visitor actually interacted with others at the table (53% of the instances).  

That is, the move toward success came when the visitors who were stuck switched, at least temporarily, from simply working in parallel or alongside other visitors on the task to a more mutual or collaborative approach.  These two types of actions (“boundary spanning perception” and “boundary spanning action”) were also often coupled with other forms of interaction, such as asking for clarification or making suggestions. 

So, a key and substantial contributor to the transition from unproductive exploration and tinkering to productive exploration was the spontaneous collaborative interaction that occurred between visitors, who were often strangers to one another.

Creatively finding patterns

Seeing and documenting this across-visitor pattern required the imaginative combination of two externally focused forms of pattern detection.  First, creation and development of the “Hidden Markov Model” enabled the researchers to selectively identify, and flag for further study, those few promising instances – from the millions of events across thousands of visitors – in which museum-goers at the tabletop transitioned from a sustained unproductive state to a productive state.  Second, the researchers needed to create, and apply, a systematic coding system of the types of interactions that visitors could engage in.  And then, the visitors themselves tell us something important about different types of imagination as well.

To creatively understand our world, we clearly need everything that internally generated expressive imagination can give us.  But, equally, we need instrumental or pattern-focused imagination, coupled with collaborative interaction and feedback, to empower us to better chart and comprehend both our world, and each other.  We need creative imagination – inside and outside our heads.  


Feng, Z., Logan, S., Cupchik, G., Ritterfeld, U., & Gaffin, D. (2017). A cross-cultural exploration of imagination as a process-based concept. Imagination, Cognition and Personality: Consciousnesss in Theory, Research, and Clinical Practice, 37, 69–94.

Tissenbaum, M. (2020).  I see what you did there! Divergent collaboration and learner transitions from unproductive to productive states in open-ended inquiry.  Computers & Education, 145, 103739.

Tissenbaum, M., Berland, M., & Lyons, L. (2017). DCLM framework: Understanding collaboration in open-ended tabletop learning environments. International Journal of Computer-Supported Collaborative Learning, 12, 35–64.

Image source: Archivo Agencia Brasil via Wikimedia Commons

Oh no, my cow just fell over—but I can reboot

Suppose you need to translate technological computer terms, such as “browser” or “cache” or even “crash” into another language in which such technological terms are absent? How literal can you be—or is metaphor what is needed?

How do we convey meaning effectively when the cultural building blocks are so different?

Take this imaginative approach:

“Ibrahima Sarr, a Senegalese coder, led the translation of Firefox into Fulah, which is spoken by 20m people from Senegal to Nigeria. ‘Crash’ became hookii (a cow falling over but not dying) . . . In Malawi’s Chichewa language, which has 10m speakers, ‘cached pages’ became mfutso wa tsamba, or bits of leftover food. The windowless houses of the 440,000 speakers of Zapotec, a family of indigenous languages in Mexico, meant that computer ‘windows’ became ‘eyes.’”

The translation project is also a great example of goal synergy: “As well as bringing the linguistically excluded online, localisation may keep small languages alive.”

Analogies and metaphors are part and parcel of our communicative repertoire and we can use them more or less purposively, and more or less creatively. Metaphors and analogies are not curlicues—they are enmeshed in how we think. They are not mere ripples on the surface but currents that move the stream—and us—forward.

As we observe in our book Innovating Minds, analogies:

“are clearly important in the generation of our ideas but they can also serve several other functions in fostering positive creative change and development.  Analogies enable us to use what we already know in order to better understand or grasp something that is novel or less familiar.  In one study of new product development projects, 6 of 16 people interviewed explicitly noted that analogies helped to promote communication between team members, designers, and engineers during new product development.  Two of the interviewees even stated that enhanced communication was the most important aspect of the analogy in the given project.  The communicative and explanatory functions of analogies may prove especially pivotal in bridging between teams and individuals with quite disparate backgrounds, task priorities, and thought processes.”