DATA, INFORMATION, KNOWLEDGE AND ACTION

This is old ground, but I was thinking again about these definitions.

Data of course is what we get back by the carload from sensors that seem to be everywhere.

Information, according to the old information theory texts, is data which reduces uncertainty. In Baysian statistical terms, one has a defined set of alternatives with a distribution across them, specifying the a priori probability that each is true. Data arrives, and one calculates the a posteriori probability distribution. If the data do in fact change the probability distribution, then they contained information. Indeed, information theorists define the amount of information as exactly the change in the average uncertainty before and after the data arrived.

An interesting aspect of this idea is that the data provides information relevant to a specific question. So the same datum may provide different amounts of data to different people, depending on the use to which they put it. (Say that someone buys a set of books bound in green leather. That datum might well provide critical information the an interior decorator. It would be of very little value in describing what the person would learn from the books.)

In any case, one aspect of the transformation of data to information is that the user has to define the way the data is to be used before it becomes information.

What then is knowledge. I fear this discussion is bound by my language. Were I writing and thinking in Spanish, I would be making distinctions between the verbs “conocer” and “saber”, distinctions that don’t exist in English. But that’s life!

Knowledge seems to me to be embodied information. I feel quite uncomfortable saying that data flowing around in cyberspace is “knowledge”. Long ago in this blog I did a riff on the embodiment of information in people, machines, organizations, and other institutions. I feel comfortable defining information embodied in people as “knowledge”.

Sometimes a small group of people work in such a way that no one person has all the knowledge needed to carry out the work of the group, but at least one person in the group has each piece of that necessary knowledge. When things mesh, the group carries out its work effectively, because each piece of work falls to a person or people with the knowledge required to carry it out. I feel comfortable saying that the group knows something – that is, that a group of people can collectively internalize information and transform it into a body of knowledge that none of its members holds individually.

I am not sure that I want to use the word “knowledge” to describe the information embodied in a machine, or in the processes of a large formal organization, or in a still larger social institution such as a market or a nation. But on the other hand, I don’t have an alternative word either.

How about “understanding”? I guess “understanding” in my mind is more about how things work, and “knowledge” is about what things are. I use the phrase “scientific knowledge and understanding” fairly frequently, and it seems to me that it makes sense to do so. One can know what a star is, and understand the processes that fuel a star. One can know the components of an ecosystem, and begin to understand how it works – how the ecological balance among its components is achieved. One can know people (and address them by name), and one can understand people (and predict how they will act).

But, as James Fallows pointed out in his Atlantic Monthly piece, “Blind Into Baghdad” (not yet available on the web), there was a lot of knowledge and understanding organized by the U.S. Government about Iraq prior to the war. Arguably enough knowledge and understanding was mobilized to have accurately predicted many of the problems that actually arose – problems that could have been avoided with appropriate action. But, and this is my point, there is a gap between knowledge and understanding versus knowledge-based and understanding-based action.

In some fields I think the term “knowledge-based” is gaining currency. Thus we hear of knowledge-based medicine, knowledge-based agriculture, and knowledge-based government services. Interestingly, I think these phrases tend to mean medical, agricultural and governmental interventions based on scientific knowledge and understanding – meaning knowledge and understanding of an especially valued kind.

Wouldn’t it be nice if the knowledge and understanding gained from the most effective processes were more often translated into action?

So much for now.