As I begin the concept modelling work for the ADAPT project - attempting to map what "data science" means to different stakeholders, and to synthesise models of vocational rehabilitation from the literature - I find myself returning to a framework I first encountered during my postgraduate studies at Brunel around 2009-2010.
Hugh Dubberly's work on concept mapping drew explicitly on Joseph Novak and D. Bob Gowin's educational research. At the time, I was struck by how their framework - developed for helping students learn science - seemed to articulate something fundamental about design practice. Now, entering a complex multi-stakeholder project where different parties appear to mean radically different things by the same words, the framework feels more relevant than ever.
This post articulates the epistemological foundations I'm drawing on. The "Vee heuristic" makes explicit the structure of knowledge construction. Understanding this structure helps explain what I'm trying to do with concept mapping - and why it matters that different stakeholders bring different conceptual frameworks to the same events.
The Problem of Meaningful Learning
Gowin and Novak were concerned with a practical problem: why do students so often fail to learn meaningfully? Students could memorise definitions, pass tests, and still not understand - they couldn't apply concepts to new situations or connect ideas together.
Their answer drew on cognitive psychologist David Ausubel's theory of meaningful learning. Ausubel argued that:
"The most important single factor influencing learning is what the learner already knows. Ascertain this and teach accordingly".
Meaningful learning occurs when new information is linked to concepts already in the learner's cognitive structure. If there's nothing to connect to, the new information can only be memorised by rote - held temporarily but not integrated.
This has implications beyond classrooms. Any situation where people need to construct understanding - which includes design situations - faces the same challenge: new observations must connect to existing concepts, or they remain disconnected facts rather than meaningful knowledge.
Gowin's Vee: A Heuristic for Knowledge Construction
Gowin developed the "Vee heuristic" to make explicit the elements involved in constructing knowledge. The Vee is shaped like its name suggests - a V - with two sides meeting at a point.
The left side represents the conceptual/theoretical:
- World views and philosophies
- Theories
- Principles
- Concepts
The right side represents the methodological:
- Value claims
- Knowledge claims
- Transformations
- Records
At the point of the Vee - where the two sides meet - are events and objects. This is where inquiry begins (and data is derived): observing something happening in the world.
The structure makes an important claim: knowledge construction requires both sides working together. You observe events and objects, make records, transform those records - but all of this is guided by the concepts and theories you bring to the observation. And the knowledge claims you construct feed back into your conceptual framework.
Events and Objects: Where Inquiry Begins
At the point of the Vee are events and objects - the things we observe.
This is deceptively simple. What counts as an "event"? What do we choose to observe? These aren't neutral decisions. They're shaped by the concepts we already have:
"The kind of records we make is also guided by one or more focus questions: Different focus questions lead us to focus on different aspects of the events or objects we are observing".
Consider a simple example from Novak and Gowin: heating ice water. You could observe temperature changes over time. You could observe the appearance of the water. You could observe the behaviour of the ice. The "same" event generates different observations depending on what questions guide your attention.
This matters for design. When we observe a service - through research, through journey mapping, through ethnography - what we see depends on the concepts we bring. Different conceptual frameworks generate different observations of the "same" service encounter.
In the current project, when I attend a meeting about "the Pathway Generator pilot", different participants are observing different events. The project manager sees milestone progress. The researcher sees data requirements. The municipal partner sees resource constraints. The same meeting generates different records because different conceptual frameworks direct attention to different aspects.
Records and Transformations
Moving up the right side of the Vee, we encounter records and transformations.
Records are what we capture from our observations. They're not the events themselves but representations of events - notes, measurements, transcripts, photographs.
Transformations are how we organise and process records to make them useful for answering our focus questions. A table of temperature readings. A graph. A categorisation scheme. An affinity diagram.
Novak and Gowin emphasise that transformation is creative work:
"Some of the creativity needed to construct new knowledge must be applied to finding the best way to organize observations. It should also become evident... that the combinations of concepts and principles we know influences how we design record transformations".
Journey maps, service blueprints, personas - these are all transformations. They take records of observations and organise them into forms that (hopefully) help answer focus questions about the service. But the forms we choose, and the categories we use, are shaped by our conceptual frameworks.
The concept maps I'm creating for this project are transformations. They take records (from documents, meetings, interviews) and organise them into visual structures. But the structure I impose - what I choose to connect to what, what I treat as a category, what relationships I make explicit - reflects my own conceptual framework. Someone else, with different concepts and principles, would produce different maps from the same records.
Knowledge Claims and Value Claims
From transformed records, we construct knowledge claims - answers to our focus questions. These are the products of inquiry.
But Novak and Gowin insist that knowledge claims are not "the truth". They're constructed answers that depend on:
- What events and objects were observed
- What records were made
- How records were transformed
- What concepts and principles guided the process
"Knowledge claims are the products of an inquiry... It should be made evident to the students that constructing knowledge requires that we apply concepts and principles we already know".
Alongside knowledge claims come value claims - judgments about worth, goodness, rightness. Is this service good? Should we choose this approach? What would make it better?
Gowin suggests that "knowledge claims and value claims ride in the same boat, but they are not the same passenger". They're interrelated but distinct. Design involves both: claims about how things are, and claims about how they should be.
The Conceptual Side: Theories, Principles, Concepts
The left side of the Vee represents what we bring to inquiry: our existing conceptual framework.
Concepts are the most granular level - the categories we use to classify and relate things. Water, temperature, melting, solid, liquid. Or in our context: patient, intervention, outcome, caseworker, service.
Principles are relationships between concepts that guide understanding. "Pure water boils at 100°C at sea level". Or: "Early intervention improves return-to-work outcomes".
Theories are broader explanatory frameworks that organise concepts and principles. Kinetic molecular theory. The germ theory of disease. The biopsychosocial model of disability. Service-dominant logic.
The Vee shows these as hierarchical - theories subsume principles, which connect concepts. But the relationship with the methodological side is dynamic:
"There is an active interplay between what we know and our new observations and knowledge claims. And this is how human cultures expand their understanding of both natural and people-made events or objects".
This is crucial. We don't just apply existing concepts to make observations. Our observations can also change our concepts. New knowledge claims can alter our principles. Over time, even theories can be revised. The Vee is not a one-way flow but a cycle.
Why This Matters for the Current Work
How does this relate to what I'm trying to do with concept modelling?
First, concept modelling is knowledge construction. When I synthesise models from the vocational rehabilitation literature, or map how stakeholders understand "data science", I am constructing knowledge. The Vee provides a framework for understanding what I'm doing - and for being rigorous about it.
Second, the conceptual side shapes what we can see. The concepts, principles, and theories I bring to this work determine what aspects of the project I notice, what records I make, how I transform them. If my conceptual framework lacks the vocabulary for certain phenomena, I may not observe them at all.
Third, transformations are design artefacts. The concept maps, hierarchical diagrams, and synthesis documents I'm producing are transformations. They organise observations according to particular conceptual frameworks. Different frameworks would produce different artefacts from the same observations. This is not a weakness - it's inherent to knowledge construction. But it means my maps are not "the truth" - they're constructed representations that reflect both what I observed and what concepts I brought to the observation.
Fourth, different stakeholders bring different conceptual frameworks. When the project manager, the researcher, the municipal partner, and I observe "the same" project, we're constructing different knowledge because we bring different concepts. Surfacing these differences - making explicit the conceptual frameworks in play - is part of what concept mapping can do.
The Vee Applied to Design Artefacts
We can "lay the Vee" on any claim about events or objects - including design artefacts and research reports.
Novak and Gowin suggest questions like:
- What objects and/or events were being observed?
- What records or record transformations were made?
- What was the focus question?
- What relevant concepts or principles were cited or implied?
- Do the records made validly record the main aspects of the events and/or objects observed?
- Are relevant principles stated, implied, or ignored?
- What theory was stated or implied, if any?
- Is there a conscious, deliberate effort to tie concepts and principles to observations, records, transformations, and knowledge claims?
- Were any value claims made, and if so, are they congruent with the knowledge claims?
These questions could be applied to my own concept maps. They expose whether the artefact rests on solid epistemological foundations or merely assumes its conclusions.
Concept Mapping: Making Structure Visible
Alongside the Vee, Novak developed concept mapping as a tool for representing conceptual structures.
A concept map represents concepts as nodes and relationships as labelled links. "Water" connects to "ice" via "freezes to form". "Service encounter" connects to "touchpoint" via "occurs at". "Patient" connects to "intervention" via "receives".
Concept maps make explicit the propositions that constitute someone's understanding of a domain. They reveal:
- What concepts are present (and absent)
- How concepts are related
- Whether the structure is hierarchical or networked
- Where there are gaps or inconsistencies
This matters because:
"Concept maps are intended to represent meaningful relationships between concepts in the form of propositions. A proposition is two or more concept labels linked by words in a semantic unit".
Design work - and multi-stakeholder projects like this one - often operates with implicit conceptual structures. Different stakeholders may use the same words but with different underlying structures. "Data science" might be a concept that connects to "algorithm" and "prediction" for one stakeholder, but to "spreadsheet" and "reporting" for another. Making these structures explicit, through concept mapping, can surface misalignments that otherwise remain hidden.
This is what I'm attempting in the conceptual modelling work I'm undertaking. The hope is that by making conceptual structures visible, we can identify where stakeholders are aligned and where they diverge - and perhaps find productive ways to bridge those divergences.
Meaningful Learning in Organisations
The meaningful learning framework extends beyond individuals. Organisations also construct knowledge. And organisations also face the challenge of connecting new observations to existing conceptual structures.
When new information doesn't connect to existing concepts, organisations - like individuals - may:
- Memorise it by rote (file the report, check the box)
- Distort it to fit existing concepts (interpret contradictory evidence as confirming existing beliefs)
- Reject it as noise (this doesn't make sense, so we'll ignore it)
Meaningful organisational learning requires:
- Surfacing the existing conceptual structures (what do we already believe?)
- Making explicit how new information relates (does this confirm, extend, or challenge what we know?)
- Revising structures when evidence warrants (updating concepts, principles, theories)
Design work that produces observations without attending to organisational conceptual structures may fail to generate meaningful learning. The observations remain disconnected - recorded but not integrated. This is a risk I'm aware of in the current project: that the concept maps I produce might be filed rather than engaged with, if they don't connect to the conceptual structures that stakeholders already hold.
What Comes Next
This post has articulated the epistemological framework underlying my concept modelling work. The Vee heuristic provides a structure for understanding knowledge construction:
- Knowledge is constructed, not discovered
- Construction requires both conceptual frameworks and empirical observations
- Events and objects are where inquiry begins
- Records and transformations mediate between observations and claims
- Knowledge claims and value claims are interrelated but distinct
- The process is cyclical - observations can revise concepts
In the next post, I'll describe the method I've developed for conceptual modelling in design, drawing on this framework and on other influences including Ecological Interface Design and Jon Kolko's work on synthesis.
References
Novak, J.D. and Gowin, D.B. (1984). Learning How to Learn. Cambridge University Press.
Novak, J.D. (2010). Learning, Creating, and Using Knowledge: Concept Maps as Facilitative Tools in Schools and Corporations (2nd ed.). Routledge.
Novak, J.D. (2007). Human Constructivism: A Unification of Psychological and Epistemological Phenomena in Meaning Making. In Bentley, M.L. (ed.), Noddings Handbook of Constructivism. Springer.
Ausubel, D.P. (1963). The Psychology of Meaningful Verbal Learning. Grune & Stratton.
Ausubel, D.P. (1968). Educational Psychology: A Cognitive View. Holt, Rinehart and Winston.
Gowin, D.B. (1981). Educating. Cornell University Press.
Dubberly, H. (2010). Creating Concept Maps. Dubberly Design Office.