Last changed 24 May 2020 ............... Length about 3,000 words (23,000 bytes).
(Document started on 15 Feb 2005.) This is a WWW document maintained by Steve Draper, installed at You may copy it. How to refer to it.

Web site logical path: [] [~steve] [best] [this page]

Data, information, knowledge, wisdom

By Steve Draper,   Department of Psychology,   University of Glasgow.

(See also the wikipedia entries on: the DIKW model and wisdom.)

Where is the wisdom we have lost in knowledge?
Where is the knowledge we have lost in information?
  T. S. Eliot, The Rock (1934)

The sequence in brief:
  1. Features
  2. Data
  3. Information
  4. Knowledge
  5. Understanding
  6. Wisdom
  7. Being, transformation, enlightenment
Another sequence, by Frank Zappa:
  1. Information
  2. Knowledge
  3. Wisdom
  4. Truth
  5. Beauty
  6. Love
  7. Music

There is a sequence or hierarchy:

  1. Differences, features. These can be detected (sensed) but it is not known if they have regularity, let alone what they mean. These may be either noise or data; it is not yet determined.

  2. Data. Defined differences to which meaning may be attached. E.g. raw sensor data, before correction and transformation. These are created systematically, but we don't know enough about their generation to draw conclusions about their meaning. (Alternatively we could say that data are measures created for a particular purpose, but we have to consider whether and how they could serve our current purpose.)

  3. Information. This takes data, applies knowledge (e.g. knowledge of a previously established code), and delivers meaning. This is the information of information theory, of Dretske. It is essentially defined as what cuts down the choices from a pre-existing set of possible meanings. It is that which delivers us conclusions about the world.

  4. Knowledge.

  5. Understanding. This essentially, while never complete, means knowing the connections of a given item to numerous other perspectives, ideas, facts. Knowing not an isolated item, but its relationships to other things we know.

    Professor M.E. McIntyre: [something like] "I always tell my students that understanding means seeing it from more than one viewpoint, and making it all consistent: in words, equations, diagrams, pictures." (See also his ideas on lucidity.)

    Malcolm Gladwell: "The key to good decision making is not knowledge. It is understanding. We are swimming in the former. We are desperately lacking in the latter."

    But von Neumann: "Young man, in mathematics you don't understand things, you just get used to them."

  6. Wisdom. The best place to start reading is Lester (2000) which, although short, covers a lot of history and references, and makes some good points.

    These days our culture is dominated by powerful technical knowledge with highly productive applications, which none of the sages of the past had a clue about. So either 'wisdom' is not needed as an idea or category, or else it is likely to be used for kinds of knowing other than science, and more generally, than the powerful technical knowledge which our lives mostly depend on today. What follows is my own selection of suggestions about what these additional kinds might be.

    Aristotle's Metaphysics defined wisdom as the understanding of causes. This would simply make it close to what educationalists would now call deep learning; and certainly what we call knowledge (not wisdom). At this point, we might suggest that wisdom is being right without comprehension; or equally is a name for something that isn't knowledge or understanding but which leads to successful choice and action. It is the opposite of Aristotle, but in line with the phrase "the wisdom of crowds", which is about how accurate the average of guesses can be even when no-one understands it.

    William James (1890 ch.22 p.369ff.) in effect develops this view by pointing out that the mark of greater intelligence and knowledge is "condensation" or "the principle of selection" by which intermediate steps and reasons are no longer given nor available to consciousness even in the sage. This then appears behaviourally as being right without understanding, certainly without talking or thinking about justifications and reasons. (Aphoristically, but perhaps misleadingly, he summarised this as "As the art of reading ... is the art of skipping, so the art of being wise is the art of knowing what to overlook.)

    Consider that again, and some more:   Wisdom is the art of knowing what to overlook.

    We might say that this could be rational, if you know enough to be confident of the answer, but not enought to be able to rebut some bits of data or reasoning that tend to go against it.

    Other notes / sayings on Wisdom
    "Wisdom is one of the few things that look bigger the further away it is." (Terry Pratchett, Witches abroad)

    Associative memory. A lot of work has shown how both humans and neural networks can learn languages (for instance) entirely from exposure to large quantities of examples, so that they behave as if they had learned complex rules (including exceptions) but without being able to articulate or be aware of those rules. As with James' view, this is expertise without understanding; but in this case, not from having acquired and then forgotten it, but without ever having it. This fits with the association of wisdom with great experience.

    Metacognition, as in knowing when to apply a given skill. Larry Niven suggests that wisdom is knowledge plus the skill to use what you know. Taken literally that would just mean having both the theory and practice, both the declarative and the procedural aspects of something. But he probably meant (in the story context in which it appeared) something like the higher level strategic skill of knowing when to apply something, rather than how. For instance, almost all of us know how to keep quiet, and how to explain ourselves: but it's much more difficult to choose which of these is best to act on in each situation.

    This is exactly what Anderson et al. (2001) mean by "metacognition" in their revision of Bloom's taxonomy. Their idea of what metacognitive knowledge is, is not just knowing a rule (having a skill) but knowing when and when not to apply it. Perhaps this is wisdom. The difference between the written law, and the decisions of a judge.

    Metacognition is treated at greater length on this page: see sections 8 and 9. Section 8 is about an example that might well be called "wisdom".

    Metacognition doesn't fit into the sequence from features to wisdom -- it is a kind of side-branch that is a type of knowledge (knowledge about knowledge); but because it is about controlling and directing other knowledge, it seems "above" knowledge, and occasionally seems to be a significant kind of wisdom.

    Other, probably less insightful, uses of "wisdom" include:

  7. Being, transformation, enlightenment. Possible extensions of the sequence.

    Being: the thought here is that beyond knowing skills (knowledge), and knowing when and when not to apply them (wisdom = metacognition), is the integration or incorporation of all that into a way of being that is also an identity. Lee Shulman speaks of this as a component of what is learned in a professional training or education (for law, medicine, etc.). This may mean an integration of one's values and impulses in line with one's thoughts, beliefs, and actions. Your way of being / way of life; where your desires, goals, impulses are fully integrated and aligned with your actions and beliefs. This is an existential statement or view. And it seems to mark the end of learning, or of the need for significant further development.


BUT note that it is possible to make sense of aphorisms that imply a reverse direction e.g. "we understand more than we know". E.g. nazi citizens were not told, and did not know, about concentration camps; but may have understood something about people disappearing, .... Children may understand they shouldn't talk about some things, .... We may understand exactly what to say to calm a family member down, without knowing that we know that...

If this means anything, it might be about: (1) (not) knowing what we know or remember cf. being surprised we can remember something. (Issues of metamemory.) (2) A facet or consequence of the way we "know" things at two or three levels, which are usually but not entirely aligned in our minds. Examples of this include a) procedural (behavioural) knowledge vs. declarative (conceptual) knowledge; b) inconsistencies (recorded in studies of science concepts in school children) between behaviour, predictions, explanations; c) Activity Theory's distinction between Activities, Actions, Operations.

"Education makes us more stupid than the brutes. A thousand voices call to us on every hand, but our ears are stopped with wisdom." - Jean Giraudoux
Perhaps this uses 'wisdom' in the metacognitive sense, but refers to cases where it is inaccurate.

Belief, knowledge, thought

The word "belief" doesn't belong in the above sequence, but is part of another set that is not really about truth in the world, but about whether the speaker is assuming or is drawing attention to how questionable an assertion is. Here I'll call the putative fact the "proposition", and the person talking or writing about it the "speaker". There is a set of cases depending on the combinations of whether the proposition is presupposed true or is in question; and whether who believes it is in question or not:

Knowledge has been defined as true, justified, belief. But in actual English usage, from a Socratic perspective, and from a child development viewpoint, this is back to front. We say we know something when we hold that a proposition we believe is true but are not thinking of any grounding or warrant for it. When its truth is in doubt, then we mark this by saying "believe": in ordinary discourse, we only say we "believe" something if we want to draw attention to the idea that it may be false, whereas if we are just taking something as given, we say "know". A very young child cannot grasp that other people may not know the same things as they do: when they start to be able to handle this, then we talk of them having acquired a "theory of mind"; i.e. of tagging things they know according to who else knows or believes them. It is sometimes claimed that those suffering from autism cannot do this (they can only know, not believe). In fact, to say we believe something is to say we are holding in mind a proposition whose truth or falsity we are able to reason about. In other words, we are able to think critically about it. This is knowledge plus doubt; and is more advanced than simply relying on propositions we are unable to question.

In general, then, "belief" marks going one step beyond knowledge, to a "theory of mind", "truth maintainence system", "reason maintainence system", "critical thinking". That is, not simply remembering facts i.e. what is true, but the reasons for believing it and/or who believes it.

In what way(s) does this matter?

At the simplest level, this topic is just about pondering distinctions between similar words, and wondering whether or not there are important conceptual distinctions hiding there. Whether there are different kinds of knowledge or knowing hiding here.

However for me, this began with reading Dretske's book on what information was. The enduring point this left me with, was that the technical definitions of information (important in computing and in physics), particularly in communication, are dependent on a pre-existing knowledge. Dots and dashes only mean something when sender and receiver have pre-agreed that they are using Morse code, and so on. So information is useless, or rather non-existent, without prior knowledge of the alternatives and of how these differences appear in the data.

If we consider the progression of stages data → information → knowledge → ... from left to right, then the right hand stage at each step extracts new value from the left, but only by virtue of assuming another kind of thing in advance. Essentially, then, the flavour of all the earlier steps in this progression is of building certainty from empirical sense data; a model of perception; "bottom up" construction. Yet it depends on pre-existing certainties, presupposed true but/and not tested or learned by the left to right flow. In this way the assumptions are like the knowledge vs. belief distinction: stances about what is being assumed at a given moment, rather than any absolute status.

There is a converse to this: whenever you have to learn not a new item, but a new field e.g. when you switch research fields or start learning a completely new subject, then what you as a learner most need is a working set (however simplified) of these assumptions without which little can be done. This is a top down direction of travel or priority; and it is one way of explaining the importance of teachers: not to communicate lots of data, but to install assumptions that allow learning to begin, and to progress efficiently. E.g. "don't run before you can walk", "don't bother looking at X it's beside the point", ... Installing these assumptions in a learner equips them, not with conclusions, but with the means to interpret and so self-teach the area.

Some references / links to other views on "wisdom"

Web site logical path: [] [~steve] [best] [this page]
[Top of this page]