The classic DIKW pyramid and it’s problems. This is an intro post for a longer series about my Discipline of Understanding.

In 1989, Russell Ackoff published ”From Data to Wisdom.” While his paper became the canonical reference for the D-I-K-W hierarchy, Ackoff actually proposed five tiers—adding Understanding between Knowledge and Wisdom. The field cited him but quietly dropped his contribution somewhere along the way.

We all stand on the shoulders of giants, and Ackoff was no exception. The four-tier version he was building on didn’t appear from nowhere.

In 1988, Anthony Debons, Esther Horne, and Scott Cronenweth introduced the "Knowledge Spectrum", the first graphical representation of the DIKW hierarchy. The pyramid visualization that everyone pictures? It came from them. But they aren’t responsible for the hierarchy itself.

That would be Milan Zeleny, who formalized the framework in 1987, mapping each layer to a cognitive mode:

1. Data as “know-nothing”
2. Information as “know-what”
3. Knowledge as “know-how”
4. Wisdom as “know-why”

But there’s still someone who predates all these flowchart enthusiasts.

The first recorded mention of this framework comes from the most unlikely creature imaginable: A hobbit. Wait, no, that’s not right.

The first recorded mention of this framework comes from a poet. In 1934, T.S. Eliot wrote in The Rock: “Where is the wisdom we have lost in knowledge? Where is the knowledge we have lost in information?” Eliot is lamenting decay, not mapping ascent. The original articulation of this progression pointed down, not up.

Somewhere between Eliot’s poem and the business school slide decks, the direction got inverted. What started as a meditation on ideatic entropy became a roadmap for accumulation. Eliot warned us we were losing something. The pyramid promised we could gain it back by stacking more.

DIKW Problems

The Accumulation Fallacy: The Knowledge Spectrum visualization implies that accumulation leads to transformation. Stack enough data, information emerges. Stack enough information, knowledge appears. This “automatic progression” interpretation isn’t what Ackoff claimed—he explicitly stated that wisdom cannot be automated. Martin Frické said the model had “a central logical error” and is “based on dated and unsatisfactory philosophical positions”.

The Linear Progression Fallacy: The DIKW Model compresses a complex process into a one-way pipeline. But real understanding doesn’t work that way. David Weinberger picked up Eliot’s thread almost 8 decades later in Too Big to Know (2012), arguing that knowledge is “creative, messier, harder won, and far more discontinuous” than the pyramid suggests.

The Pristine Permanence Fallacy: The feeling of understanding doesn’t require the foundation to be sound. You can build elaborate castles on sand and never notice until the moment you need it most and they collapse. Even sturdy foundations need constant maintenance. Climbing requires effort; falling is free. Wisdom calcifies. Experience atrophies. Information degrades. Eliot’s warning from last century still echoes, but the hum of the data centers is drowning it out.

The Field Is Stuck

These problems aren’t new. They’ve been documented, debated, and restated for decades. The field knows the model is broken. It just doesn’t know what to do about it.

The responses fall into three camps, and none of them fix the problem.

The first camp keeps using it anyway. Terry Dwain Robertson acknowledged in 2013 that the hierarchy “fails under rigor” but defended it as a teaching heuristic — useful for introducing concepts, just don’t examine it too closely. Peter Jackson (not that one, your eyes must be cheated by some spell), writing in 2021, went further: he reframed DIKW as “an ecosystem in which every part supports all others.” Which is an interesting move — if it’s an ecosystem where everything supports everything, it’s not a hierarchy anymore. The defenders are quietly gutting the model while insisting they’re preserving it.

The second camp has pitchforks and is calling to burn it down. Frické called for abandoning the model entirely. Gordon Vala-Webb declared it “must die.” Rafael Capurro dismissed it as “a fairytale.” Strong words, all of them. But none offered a replacement. Critique without construction. You can’t just remove a model that’s embedded in ITIL certifications, DoD knowledge management, and every business school slide deck without putting something in its place.

The third camp documents the damage and moves on. Chaim Zins surveyed 57 scholars and collected 130 definitions of data, information, and knowledge — proving the field can’t even agree on what the layers mean. Jennifer Rowley documented the transformation gap — no one can explain how you actually get from one layer to the next. Rigorous autopsies of a model that’s still walking around.

Robertson’s position is the most revealing. He frames it as a choice: the model either works as theory or it works as heuristic. This is a false dichotomy. “Wrong but useful” and “right” aren’t the only options. There’s a third: not wrong, just incomplete.

The layers are real. The progression is real. What’s missing is the mechanism — how does data actually become information? How does information become knowledge? In 2023, Bratianu and Bejinaru put it most directly: the model “does not explain what the mechanisms and the driving forces which produce each transformation from one given level to the next level of complexity are.”

That’s the gap. Not a flaw in the structure — a hole in the explanation.

The Missing Verb

I think Ackoff was right to include Understanding, but wrong to include it as a component, part of the structure of the DIKW hierarchy. You don’t arrive at understanding and then move on. Understanding is how you move at all. Understanding is the verb, the transformative mechanism within the structure.

Eliot asked where the wisdom went. Maybe it didn’t go anywhere. Maybe we just stopped doing the work, expecting it to emerge from the pile that was supposed to be a pyramid.