After my posts about WIKID, Diarmuid Pigott and I had an interesting email exchange in which he mentioned a paper he had written that takes a different view of this question: The Noetic Prism: A New Perspective on the Information, Data, Knowledge Complex (pdf), with Valerie Hobbs from the School of Information Technology, Murdoch University. (Presented at the 4th Western Australian Workshop on Information Systems Research, 2001.)
In it, they propose noetic prisms as a new model of thinking about the combination of data, information and knowledge (DIK) to account for the familiar problems with the common view of a progression from data to knowledge. And they wrap up the paper with several examples of making sense of the model in business.
Having read through the article, I kept thinking about how this applies to personal knowledge management. From the perspective of the model, nearly everything one does adds to their own noetica - body of knoweldge - expanding and contracting, and making it more complex. One of the ideas I've had behind PKM is that there are different ways to navigate my own noetica, or maybe it is that there are ways I can teach myself to think about my noetica that will enable me to find specific items, connect those items, discover connections, jump to new items, etc. The spiral or nautilus shell analogy the authors use seems to make sense.
They do a nice job of reviewing the background, primarily reviewing how computer science has viewed the terms through the last 60 years of development in the discipline. This summary says it all and lays the foundation for their model.
As we can see from this brief review, then, there is no consensus on either the detailed definitions of data, information and knowledge, nor of their relationships. Definitions of 'data' range from contextfree 'facts' to the intelligently-structured material found in databases, 'information' from Shannon's meaning-free signal to a contextualised and processed method of informing a user, and 'knowledge' from a further-refined variety of information to something that is attainable only through personalised experience.
When we look for commonalities in these definitions though, we find general agreement on two points. The first is that there is a process by which something is being transformed into something else that is more useful, either through a physical process of calculation or through a personal act of internalisation or contextualisation. A further point of agreement is that this processing is cyclic, with outputs becoming inputs to another process: this leads to the definitions of data-information-knowledge being fluid, shifting according to the perspective from which they are viewed.
So, instead of the hierarchical model that is so familiar, what to do? Pigott and Hobbs talk about a noetica, the body of all this stuff, no matter if it is data, information or knowldge. The point is the value that it has and the value that it gains as the entities within the noetica are aggregated, transformed and interrelated (their terms). With this way of thinking, the type of entity being acted upon isn't so important as the fact that we are continually growing the noetica via these actions. They don't say it explicitly, but this growth also informs the process of bringing new entities into the noetica -- learning and experiementing outside the existing walls. I suspect things like PMI's Project Management Body of Knowledge (PMBOK) is a representation of a noetica that is continually expanded/revised as the PMI community learns and refines the contents.
The authors describe their new noetic prism model as triangular prism with shape, granularity and scope as the axes of the triangle and complexity in the vertical direction. Working with the path that a noetica is potentially infinite, this vertical axis is essentially infinite.
They wrap up the paper with three examples: 1. Merger of IT systems within a business merger; 2. Busines-to-business technologies; 3. Intelligent search tools. All three examples focus on IT. The first goes into some detail on the value of understanding context before going about "merging" these systems. This applies equally well to management systems. Bringing them together without appreciating their differences, similarities and backgrounds will spell disaster. Merging companies spend all sorts of time establishing how they will merge and they still don't get it right.
I would love to see a discussion of business applications of this model to extend it beyond the world of computer science. Some starting points: a new executive takes over; an employee is promoted to her first management position over poeple she had worked alongside; a consultant comes into an unfamiliar business; a new community develops within a company. Fun...