DIKW Hierarchy
The DIKW hierarchy (Data → Information → Knowledge → Wisdom) is a widely cited model in knowledge management and information science that arranges four cognitive constructs in ascending order. Multiple versions exist with distinct characteristics.
Definitions
- Data: Raw, unprocessed symbols — numbers, characters, signals. No meaning attached.
- Information: Data with context, relevance, purpose, and organization.
- Knowledge: Information applied through synthesis, decision-making, building models.
- Wisdom: Knowledge tempered through time, trial, and error — judgment and perspective.
History
| Year | Contributor(s) | Contribution |
|---|---|---|
| 1927 | Clarence W. Barron | Clarence Barron articulated a knowledge-intelligence-wisdom hierarchy in 1927 |
| 1934 | T.S. Eliot | T.S. Eliot articulated an information-knowledge-wisdom hierarchy in 1934 |
| 1938 | H.G. Wells | H.G. Wells proposed a universal knowledge organization system in 1938 |
| 1945 | Vannevar Bush | Vannevar Bush proposed a system for transforming information access into knowledge creation in 1945 |
| 1948 | Claude Shannon | Claude Shannon defined information as a mathematical quantity in 1948 |
| 1948 | Norbert Wiener | Norbert Wiener explored the relationships between signals, information, and meaning in 1948 |
| 1949 | Sherman Kent | Sherman Kent proposed a data-information-intelligence process in 1949 |
| 1955 | Kenneth Boulding | Kenneth Boulding proposed a signals-messages-information-knowledge hierarchy in 1955 |
| 1956 | Kenneth Boulding | Kenneth Boulding defined information as experience that changes one's subjective "image" of the world in 1956 |
| 1956 | Benjamin Bloom, Max D. Engelhart, Edward J. Furst, Walker H. Hill, David R. Krathwohl | Benjamin Bloom et al. proposed a six-level cognitive processing hierarchy in 1956 |
| 1958 | Michael Polanyi | Michael Polanyi defined tacit knowledge as "We know more than we can tell" in 1966 |
| 1959 | Peter Drucker | Peter Drucker coined 'knowledge work' to frame knowledge as an economic force in 1959 |
| 1962 | Fritz Machlup | Machlup classified five types of knowledge as measurable economic output in 1962 |
| 1962 | Robert S. Taylor | ”The Process of Asking Questions” — four levels of information need (visceral, conscious, formalized, compromised). Hierarchy of question states, not epistemic products. |
| 1966 | Börje Langefors | Börje Langefors proposed The Infological Equation defining information as a function of interpreting data in 1966 |
| 1966 | Michael Polanyi | The Tacit Dimension — expanded tacit knowledge framework. Argued some knowledge is inherently inarticulable, limiting what can be transmitted as information. |
| 1967 | Russell Ackoff | ”Management Misinformation Systems” — argued managers were drowning in irrelevant information. Distinguished relevant from irrelevant information, 22 years before his DIKW formulation. |
| 1969 | Herbert Simon | The Sciences of the Artificial — distinguished data processing, information search, and satisficing decision-making. Treated cognition as computation over symbols without stratifying epistemic products into a hierarchy. |
| 1971 | West Churchman | West Churchman proposed knowledge is subjective and requires inquiry, not just accumulation in 1971 |
| 1972 | Gregory Bateson | Steps to an Ecology of Mind — “levels of learning” describe transformation between epistemic states: Learning 0 is response to signal (data-level), Learning I is change in response (information-level), Learning II is change in the set of alternatives (knowledge-level restructuring), Learning III is paradigmatic transformation. Never used DIKW terminology; structural parallel noted by cybernetics scholars. |
| 1972 | Stafford Beer | Brain of the Firm — Viable System Model uses five recursive organizational levels. System 1 handles operations (data-adjacent), System 5 handles identity and policy (wisdom-adjacent). Beer did not frame this epistemically; structural parallel to DIKW requires interpretation. |
| 1973 | Daniel Bell | Daniel Bell proposed the data-information-knowledge model in 1973 |
| 1974 | Nicholas L. Henry | ”Knowledge Management A New Concern for Public Administration (Henry 1974). |
| 1974 | B.C. Brookes | B.C. Brookes proposed an equation defining the relationship between information and knowledge as receiver-dependent in 1974 |
| 1979 | Frank Zappa | Frank Zappa re-articulated the information-knowledge-wisdom model in 1979 Information is not knowledge, Knowledge is not wisdom, Wisdom is not truth, |
| 1980 | B.C. Brookes | B.C. Brookes integrated his Information-Knowledge equation into Karl Popper's World 3 Framework in 1980 |
| 1981 | Fred Dretske | Fred Dretske applied Shannon's Information Theory to Epistemology, proposing a causal mechanism for the Information to Knowledge transformation in 1981 |
| 1982 | Harlan Cleveland | "Information As A Resource" — proposed information-knowledge-wisdom model. Credited T.S. Eliot. No data layer. |
| 1983 | Fritz Machlup | Fritz Machlup documented pervasive definitional confusion between information and knowledge in 1983 |
| 1983 | Edward Feigenbaum & Pamela McCorduck | The Fifth Generation — distinguished data from knowledge in the context of expert systems. Operationalized a two-tier model (data as unprocessed inputs, knowledge as encoded heuristics and rules) without an information intermediate layer. |
| 1984 | Nicholas Maxwell | Nicholas Maxwell proposed the IKW model as academic "wisdom-inquiry" in 1984 |
| 1984 | David Kolb | Experiential Learning (Prentice Hall) — four-stage cycle: concrete experience, reflective observation, abstract conceptualization, active experimentation. Describes how knowledge is created through transformation of experience, not what epistemic products exist. Parallel process model. |
| 1984 | Knowledge Engineering Review | Cambridge University Press launched the journal. Expert systems literature used “data,” “information,” and “knowledge” frequently but interchangeably, not as a formal hierarchy. The AI/CS community did not cite the IS/KM tradition in this period. |
| 1986 | Mortimer J. Adler | Mortimer Adler proposed the DIKUW hierarchy in 1986 |
| 1986 | Bruce Blum | Bruce Blum proposed the DIK model in 1986 |
| 1987 | Milan Zeleny | Management Support Systems — formalized DIKW with cognitive modes (know-nothing, know-what, know-how, know-why). |
| 1987 | Mike Cooley | Mike Cooley proposed the DIKW hierarchy to argue against the deskilling of workers through automation in 1987 — [[Architect or Bee (Cooley 1987)|Architect or Bee?]] (Hogarth Press revised edition; the 1980 first edition did not contain the DIKW passage). Used the hierarchy as a political argument against the deskilling of workers through automation. Contemporaneous with Zeleny. |
| 1988 | Anthony Debons, Esther Horne, Scott Cronenweth | Knowledge Spectrum — first pyramid visualization; neither Zeleny nor Ackoff used a pyramid. |
| 1989 | Russell Ackoff | Russell Ackoff proposed the DIKUW model in 1989 in From Data to Wisdom |
| 1989 | Robert W. Lucky | Robert Lucky visualized the DIKW model as a pyramid in 1989 |
| 1989 | Richard Saul Wurman | Richard Saul Wurman defined the transformation of data into information through presentation in 1989 |
| 1989 | Judith R. Graves & Sheila Corcoran | Judith Graves and Sheila Corcoran used Bruce Blum's DIK model in their nursing informatics framework in 1989 |
| 1989 | Ramona Nelson & Irene Joos | Ramona Nelson and Irene Joos proposed the DIKW Model based on Judith Graves's and Sheila Corcoran's DIK model in 1989 |
| 1993 | Karl Wiig | published Knowledge Management Foundations where he makes a distinction between information and knowledge |
| 1994 | Nathan Shedroff | Nathan Shedroff framed the DIKW Model in the Design Domain as a Continuum of Understanding in 1994 |
| 1995 | Max Boisot | I-Space model — three-dimensional alternative (codification, abstraction, diffusion). Inspired Snowden‘s Cynefin. |
| 1995 | Ikujiro Nonaka & Hirotaka Takeuchi | Ikujiro Nonaka and Hirotaka Takeuchi popularized knowledge as an organization asset in 1995 |
| 1996 | N. Venkat Venkatraman | N Venkat Venkatraman proposed the DIKAR model in 1996 |
| 1996 | U.S. Army | The US Army formalized a four-tier cognitive hierarchy in FM 100-6 in 1996 — [[Information Operations (FM 100-6 1996)|FM 100-6: Information Operations]] (August 1996), Chapter 2, Figure 2-1: data → information → knowledge → understanding. Earliest formally named cognitive hierarchy in Army doctrine; substituted understanding for wisdom. |
| 1997 | Karl Sveiby | The New Organizational Wealth — knowledge management framework distinguishing individual competence from internal/external structure. Explicitly rejected the idea that knowledge is “refined information,” calling this conflation the central error of KM. Anti-DIKW position. |
| 1998 | Robert Sternberg | ”A Balance Theory of Wisdom” (Review of General Psychology 2(4)) — defined wisdom as application of tacit knowledge mediated by values toward a common good. A K-W model: no data or information layers. |
| 1998 | Thomas Davenport, Laurence Prusak | Thomas Davenport and Laurence Prusak popularized the DIK model for corporate knowledge management in 1998 |
| 1998 | Chun Wei Choo | The Knowing Organization — integrated sensemaking, knowledge creation, and decision-making into a process framework. Published same year as Davenport/Prusak but did not reference the DIKW hierarchy. Treated knowledge as organizational activity rather than static layer. |
| 1999 | Ilkka Tuomi | Inverted Hierarchy Thesis; see Tuomi Inversion Refutation. |
| 2000 | Paul E. Bierly III, Eric H. Kessler, Edward W. Christensen | Paul Bierly, Eric Kessler, and Edward Christensen mapped the DIKW model to Bloom's Taxonomy and organizational learning types in 2000 |
| 2001 | U.S. Army | FM 3-0: Operations — discussed information superiority using a data-information-knowledge framing conceptually. Did not formalize a named hierarchy. |
| 2002 | T.D. Wilson | ”The Nonsense of Knowledge Management” (Information Research 8(1)) — argued only the information-to-knowledge transition was meaningful. Rejected “knowledge management” as a label; only information can be managed. Dissolved the D and W layers as operationally empty. |
| 2002 | Mark Nissen | ”An Extended Model of Knowledge-Flow Dynamics” (Communications of the AIS 8) — data-information-knowledge-action-performance flow model. Distinct from Venkatraman‘s DIKAR. |
| 2003 | U.S. Army | FM 6-0: Mission Command (Appendix A) — carried the Cognitive Hierarchy forward from FM 100-6 (1996) into mission command doctrine. Understanding defined as the ability to apply knowledge to recognize the fundamental nature of a problem. |
| 2004 | Gene Bellinger, Durval Castro, Anthony Mills | Gene Bellinger, Durval Castro, and Anthony Mills reframe Russell Ackoff's understanding layer as a mechanism in 2004 |
| 2004 | Jonathan Hey | Jonathan Hey applied Lakoff & Johnson's conceptual metaphor theory to the DIKW model in 2004 |
| 2004 | Irma Becerra-Fernandez & Rajiv Sabherwal | Knowledge Management: Systems and Processes — KM textbook embedding DIKW in business education curricula. Now in 3rd edition (2024). |
| 2004 | Elias M. Awad & Hassan M. Ghaziri | Knowledge Management (Pearson/Prentice Hall) — six-layer model: data-information-knowledge-expertise-capability-wisdom. Added expertise and capability as distinct layers between K and W. Used in university curricula. |
| 2004 | Milan Zeleny | Zeleny proposed the DIKWE model in 2004 |
| 2005 | Kimiz Dalkir | Knowledge Management in Theory and Practice — canonical KM textbook, now in 4th edition. |
| 2005 | Diarmuid Pigott, Val Hobbs, John Gammack | Diarmuid Pigott, Val Hobbs, and John Gammack proposed the Noetic Prism, DIKW as aspects of a single substrate in 2005 |
| 2006 | U.S. Joint Chiefs | JP 3-13: Information Operations — addressed data-information-knowledge relationships in the context of information advantage at the joint level. Did not adopt FM 6-0’s four-tier “cognitive hierarchy” label. |
| 2007 | Shirley Lacy & Ivor Macfarlane | ITIL v3 Service Transition — embedded DIKW as doctrine for 1M+ IT professionals. |
| 2007 | Anthony Liew | Anthony Liew identified problematic circular definitions in DIKW models in 2007 |
| 2007 | Jennifer Rowley | Jennifer Rowley documented the DIKW transformation gap in 2007 |
| 2007 | Chaim Zins | Chaim Zins surveyed 45 scholars and collected 130 definitions with no consensus in 2007 |
| 2007 | Rafael Capurro | Rafael Capurro dismissed the DIKW hierarchy as a fairytale in 2007 |
| 2008 | Nikhil Sharma | Nikhil Sharma partially traced the DIKW hierarchy's independent convergent genealogy in 2008 |
| 2008 | Jean-Baptiste Faucher, Andre M. Everett, Rob Lawson | Jean-Baptiste Faucher, Andre Everett, and Rob Lawson proposed the EDIKWE continuum with understanding as the differentiating process between constructs in 2008 |
| 2008 | U.S. Army | [[FM 6-01.1 Knowledge Management Operations|FM 6-01.1]] — first dedicated KM field manual; applied the cognitive hierarchy from FM 6-0 to knowledge management practice. Revised in 2012 (see below). |
| 2008 | American Nurses Association | The American Nurses Association adopted DIKW as the profession's knowledge framework in 2008 |
| 2009 | Martin Frické | ”The Knowledge Pyramid: A Critique” — foundational attack on logical errors; called for abandonment. |
| 2009 | Murray Jennex | Bidirectional model revision; proposed top-down and bottom-up flows (HICSS). |
| 2010 | Dave Snowden | KM4Dev listserv intervention — dismissed DIKW as “just plain wrong”; knowledge is a flow, not a thing. |
| 2010 | David Weinberger | Harvard Business Review critique — “a desperate cry for help”; pyramid visualization misleads. |
| 2011 | Susan Matney, Ramona Brewster, Charlene Sward, Kristin Cloyes, Nancy Staggers | Critical review of DIKW in nursing informatics (Advances in Nursing Science). Traced how nursing literature absorbed Ackoff‘s lineage alongside the original Graves/Corcoran origins. |
| 2012 | David Weinberger | Too Big to Know — knowledge is “creative, messier, harder won, and far more discontinuous.” |
| 2012 | Tom Graves | Rethinking the DIKW Hierarchy — proposed four concurrent dimensions rather than a hierarchy. |
| 2012 | Gordon Vala-Webb | ”DIKW Pyramid Must Die” — Knowledge Management World practitioner critique; five practical problems. |
| 2012 | U.S. Army | [[FM 6-01.1 Knowledge Management Operations|FM 6-01.1: Knowledge Management Operations]] — revised edition of the 2008 KM field manual; continues to build on the cognitive hierarchy from FM 6-0. |
| 2013 | Murray Jennex & Sara Bartczak | Murray Jennex and Sara Bartczak proposed a bidirectional DIKW model in 2013 |
| 2013 | Sasa Baskarada | Mapped DIKW to semiotic dimensions with quality constraints between layers. |
| 2013 | Terry Dwain Robertson | Pedagogical defense — fails as ontology, succeeds as teaching heuristic. |
| 2013 | Anthony Liew | Anthony Liew proposed the DIKIW model in 2013 |
| 2014 | U.S. Army | FM 6-0 revised edition — carried forward the Cognitive Hierarchy (Appendix B, Figure B-2) from the 2003 edition. Retitled from Mission Command to Commander and Staff Organization and Operations. |
| 2014 | Ali Intezari & David J. Pauleen | Empirical wisdom study presented at HICSS conference. Examined DIKW and management decision-making. |
| 2015 | Mark Baker | ”Every Page is Page One” — stories precede data; “if there is a pyramid, it is stories at the bottom, middle, and top.” See Baker Recursive Understanding. |
| 2015 | Stan Garfield | Stan Garfield argued the DIKW model "leaves Understanding out entirely" in 2015 |
| 2016 | Charlene Ronquillo | Most comprehensive critical examination of DIKW within nursing informatics. |
| 2017 | Ivo Velitchkov | Challenged DIKW at the definitional level — “facts devoid of meaning” is incoherent because facts require meaning. |
| 2017 | Dyfrig Williams | Named “The Fallacy of the Pipeline” — the assumption that knowledge flows linearly from producers to practitioners. |
| 2018 | Olaf Dammann | Olaf Dammann proposed the DIEK model in 2018 Dammann proposed transformation mechanics in his DIEK model |
| 2018 | Andrei Lobov | Andrei Lobov adopted DIKW as a methodological framework for smart manufacturing system design in 2018 |
| 2019 | ITIL 4 (AXELOS) | ITIL 4 dropped wisdom from the Knowledge Management practice guide in 2019 |
| 2020 | Dave Kukfa | Dave Kukfa identified Understanding as supporting each DIKW transition in 2020 |
| 2020 | Richard Veryard | Richard Veryard proposed the temporal DKJ model in 2020 |
| 2020 | Dennis Pearce | Dennis Pearce proposed an emergent, bidirectional DIKW model in 2021 |
| 2020 | David Gurteen | Six structural critiques of DIKW — model “should not be taken as an account of how knowing works.” |
| 2020 | Dilip V. Jeste et al. | ”Beyond Artificial Intelligence: Exploring Artificial Wisdom” — examined whether AI can achieve wisdom. |
| 2020 | Gideon Rosenblatt | ”When Machines Know” blog — characterized machines as “containers of collective intelligence” within DIKW framing. |
| 2021 | Peter Jackson | The Data Warehousing Institute defense — “every chief data officer should be conversant”; reframed as ecosystem, not hierarchy. |
| 2021 | Adrião & Filho | “The Pyramid of Information: criticism and opportunity” (International Journal of Advanced Engineering Research and Science). Critique later synthesized by Laine/DAMA Finland. |
| 2021 | Brent Davis & Krista Francis | Catalogued DIKW as a learning discourse in education literature. |
| 2021 | Karl Melrose | Karl Melrose proposed the DIKAR Model in 2021 |
| 2021 | Kate McDowell | Kate McDowell proposed the S-DIKW model for narratives in 2021 |
| 2021 | Dennis Pearce | Dennis Pearce proposed an emergent, bidirectional DIKW model in 2021 |
| 2022 | Yucong Duan | Yucong Duan proposed the DIKWP model in 2022 |
| 2023 | Constantin Bratianu & Ruxandra Bejinaru | Constantin Bratianu and Ruxandra Bejinaru restated the transformation mechanism absence in 2023 |
| 2023 | Islam Akef Ebeid | DIKW and AI ethics series on Medium. Applied hierarchy to questions of machine learning ethics. |
| 2023 | Bozena Kostek | Applied DIKW to deep learning audio processing. |
| 2024 | Sami Laine, DAMA Finland | Data management community synthesis — organized criticism into six categories; argued D, I, K are perspectives on the same object rather than different types. |
| 2024 | Michael A. Peters, Petar Jandric, Brianna J. Green | |
| 2024 | Bernard C.K. Choi, Noel C. Barengo, Paula A. Diaz | Bernard Choi, Noel Barengo, and Paula Diaz proposed the DIKIW model for public health surveillance in 2024 |
| 2024 | Michael Grieves | Applied DIKW as an action framework for Digital Twins. |
| 2024 | Fatima Ezzahra El Adraoui et al. | Dynamic knowledge flows in organizational context (Morocco). |
| 2024 | Akhilesh Kumar Mishra & Yifan Zhong | DIPy-AI brain-cognition DIKW architecture mapping hierarchy to neural processing. |
| 2025 | Denis Volkov | Denis Volkov proposed Context as the transition mechanism in the DIKW model in 2025 |
| 2025 | Mike Turner | ”The Wisdom Deficit” — AI ascends three DIKW levels but cannot reach wisdom. UK exam algorithm as illustration. |
| 2025 | Stefano Calzati | Ecosystemic DIKW critique — characterized model as rooted in “positivism and operationalism.” |
| 2025 | B. Schembera | Integrated DIKW with FAIR and XAIR principles for research data management. |
| 2025 | Alexis Jacobo Banon-Gomis | DIKW-based business intelligence architecture. |
| 2025 | Angelica Lo Duca & Kate McDowell | Angelic Lo Duca and Kate McDowell applied the S-DIKW framework to data visualization in 2025 |
| 2025 | Mulugeta Woldetsadik | Applied DIKW to African academic librarianship. |
| 2025 | Matteo Cristofaro & Alexis Jacobo Banon Gomis | HAIL framework derived from 122 interviews. Proposed as DIKW replacement. Matteo Cristofaro and Alexis Jacobo Bañón-Gomis proposed the DIKW model as a recursive circuit with multi-directional flows in 2025 |
| 2026 | Praise J.J. | Genes/Schemes/Memes substrate model — words as pointers triggering neural patterns, not containers delivering meaning. |
| 2026 | Nicolas Michaelsen | Four knowledge modes (propositional, procedural, perspectival, participatory) — DIKW addresses only one. |
Versions
Zeleny’s Framework (1987): Four tiers (DIKW). No visualization. Mapped to cognitive modes (know-nothing → know-why).
The Pyramid (1988): Knowledge Spectrum. Four tiers (DIKW). Pyramid/spectrum visualization. First graphical representation; visual implies accumulation.
Ackoff’s Original (1989): From Data to Wisdom. Five tiers (DIKUW). No visualization. Understanding as distinct layer; explicitly stated wisdom cannot be automated.
Popularized Model (1998): Working Knowledge. Four tiers (DIKW). Pyramid visualization. Business/KM adoption; accumulation-implies-progression became conventional wisdom.
Critiques
Different critiques apply to different versions:
- Linear oversimplification: The pyramid visualization implies a clean, sequential flow. Applies to pyramid versions. (Martin Frické, David Weinberger)
- Automatic progression: The assumption that accumulating enough of one layer produces the next. Applies to the popularized model and pyramid visualization. Common interpretation, not original authors’ claim.
- Definitional vagueness: No consensus on what any of the four constructs actually mean. Applies to all versions. (Jennifer Rowley, Chaim Zins)
- Transformation problem: No agreed mechanism for how one layer becomes the next. Applies to all versions.
- Wisdom unoperationalizable: Wisdom is defined in ways that resist measurement or implementation. Applies to all versions. (Martin Frické)
- Judgment needed at all levels: Ackoff claimed judgment only needed for wisdom; in practice it’s required everywhere. Applies to Ackoff’s original. (Martin Frické)
- Pristine permanence: The model implies that once a layer is achieved, it persists. In practice, wisdom calcifies, knowledge atrophies, information degrades. Applies to all versions. See Entropic Dynamics Of Understanding, False realms.
Dozens of alternatives have been proposed (DIKAR, DIKIW, DIEK, DIKWP, Noetic Prism, E2E continuum, I-Space, dimensional models, bidirectional models) but none achieved consensus adoption. Critics either defend a broken model (“useful as heuristic”), call for abandonment (without replacement), or document problems (without resolution). The DIKW hierarchy is a folk theory arising from a genuine phenomenological pattern in how humans experience epistemic states. It persists because the pattern it describes is real. It fails because no formulation has specified the mechanism producing the transformation between layers.
Traversal Mechanisms
People who have explicitly proposed understanding as the mechanism:
Bellinger, Castro, Mills (2004) proposed that understanding is not a separate tier but supports every transition: understanding relations (data→info), understanding patterns (info→knowledge), understanding principles (knowledge→wisdom). This is your closest prior art. It’s in your doc.
Faucher, Everett, Lawson (2008) went further with their E2E continuum, explicitly calling understanding “the process that defines the differences among these constructs.” Also in your doc.
Dave Kukfa (2020) independently identified understanding as supporting each DIKW transition. In your doc.
Nathan Shedroff (1994) framed the whole thing as a “continuum of understanding.” In your doc.
People who proposed other mechanisms:
Denis Volkov (2025) proposed context as the mechanism at every transition. In your doc.
Baskarada & Koronios (2013) mapped it to semiotic processes: measurement, meaning-making, presentation, understanding. In your doc.
Bratianu & Bejinaru (2023) proposed knowledge fields theory with energy-style transformations between rational, emotional, and spiritual knowledge. In your doc.