Human Knowledge Design and Human Knowledge Management| Home / more articles - The author - Contact |
Human Knowledge Design and Human Knowledge Managementby Matthew Leitch, 31 July 2002
Sixteen years ago I wrote a paper on a new approach to teaching and working with knowledge. It was submitted as a project for the third year of the psychology degree I was then studying. The ideas in it are still new and valuable - especially these days when the phrase "knowledge management" has been so devalued by companies eager to sell computer software that it now means databases for documents and contact lists.
The knowledge I wrote about was the knowledge in our heads. The magic stuff that lets us do things we couldn't do before we got it. My thesis was that this knowledge could and often should be designed instead of just being allowed to form haphazardly. It turned out to be a great idea I've applied many times.
Not only was this a very practical idea for anyone interested in better ways to learn or teach, but it was also a proposal for the future of psychology itself. At that time psychology was a failing science (and possibly still is) because it was approaching the human mind as it if was a natural phenomenon that worked just one way, when in fact we can and often do create our minds and make choices about alternative ways of thinking.
But before you click on the title below to read this paper please be warned: although I knew what I was talking about in 1986 I could bore for England. The style is turgid and academic and few of the exciting ideas that lie behind it managed to force their way into the text.
So I've written a modern explanation and update that follows immediately after this link.
Human Knowledge Design (February 1986)
Designing human knowledge is similar in principle to designing computer software. It takes time, careful thought, and creativity to do it well. The biggest differences are:
You can't just load the knowledge into someone's brain like you would load some software onto a computer; you need an implementation procedure to build the knowledge in the learner's mind.
The strengths and weaknesses of the human brain are radically different from a computer. The design needs to play to the strengths of the human brain.
If human knowledge design and management were practiced widely our lives would be better as a result. Here are some applications to show the possiblities.
Well executed Human Knowledge Design, with a good implementation procedure, can produce much more effective learning materials than the conventional approach to "instructional design", which does not recognise clearly that we have choices about how we think and how we structure the knowledge that drives our thinking.
Human Knowledge Design is ideal for teaching materials where:
The learner is happy to sit back and be spoonfed. It is not suitable for promoting controversial ideas.
The knowledge has long term value. It is rarely economic for knowledge that is only useful briefly.
High performance is required. For example, the knowledge of a financial market trader, a rocket scientist, a telesales person, a helpdesk team member, or computer programmer needs to perform well.
Full simulation or real experience is hard to obtain. Normally, the closer practice conditions match actual conditions the better but there are times when simulations or actual experience are hard to achieve. Human Knowledge Design can help in this situation because it is possible to determine exactly what stimuli are relevant to the knowledge used and sometimes this does not require a full simulation of the real environment. Also, it is not necessary for the learner to experiment to find the knowledge he/she needs as it has already been worked out.
Rapid learning is required. Provided the learner doesn't have to wait for a knowledge design to be created (or doesn't mind waiting) well designed knowledge delivered by an effective implementation procedure is easier and quicker to acquire than the torture of a conventional text book or training course.
Individual differences need to be reduced. Research strongly indicates that one of the biggest differences between quick learners and poor learners is that the quick learners are better at organising knowledge for themselves. You can help the weaker learners approach the skills of the faster learners by designing the knowledge for them.
The best knowledge designs will provide learners with great abilities for minimal quantities of knowledge, built into their minds using a procedure that is pleasant and efficient. The quantity of knowledge might well be expressed in "chunks". Imagine searching the internet, shopping for human knowledge packages with advertisements like this:
NEW FROM HKD! Conversational Italian version 4.0 Gives mastery of Italian up to GCSE grade A, with vocabulary of the 1,000 most common words in Italian, plus optional add-ins. Only 2,547 chunks and average build time of 127 hours. Amaze your friends and family. Features authentic accent and lifestyle vocabulary add-ins: Party Animal, Beach Bum, Business Traveller, Culture Vulture.
Often, just sketching out the structure of the knowledge needed helps enormously. I was once able to help a close friend get past a mental block on maths and get through a professional exam retake. Failure would have meant losing his job. After examining his skill in detail I noticed that he was completely unable to tell the difference between problems of different types. Consequently when questions of different types were mixed up in an exam he was stuck. I explained the gap and suggested that his skill be structured as condition -> action pairs, where the condition would identify the elements that distinguished each type of problem. I advised looking at each type of problem on the syllabus and writing down the elements of each problem situation (e.g. a set of securities to value, the value is the missing variable, the securities are bonds) one one side of a card, then some brief notes on the solution method on the other side. The cards could also be used to practice recalling the appropriate solution method given a set of problem elements. (I have since thought of a better way.) This worked dramatically well and a few days later I got a call saying that suddenly maths was making sense. The exam was no problem.
In this case I designed the structure for someone else. Designing the structure for yourself is a way to build better learning and thinking skills.
With practice, knowledge design can be done so quickly and easily that you can do it as you learn materials that haven't be pre-structured for you. This is a route to better learning and thinking. I have described this in detail in A new focus for memory improvement which also gives some details about knowledge structures.
Human Knowledge Management is concerned with the knowledge in peoples' heads, not databases and documents. Organisations such as companies should have a person or team responsible for this knowledge management. They should:
Invest in Human Knowledge Design where it makes the most impact.
Evaluate, measure, track, and monitor the quantity, quality, and importance of knowledge in the organisation. They need to estimate the number of chunks (or use some other measure) of knowledge, the people who have them, the use that is made of them, and the results that flow from them.
Advise on the knowledge implications of changes in the organisation and its activities, the implications of plans under consideration, and strategies to improve the knowledge position of the organisation. This advice should include quantitive estimates.
Factors that will typically have a major effect on the knowledge position of an organisation include:
Joiners and leavers - how many and who they are.
Learning from experience and reflection - opportunities and resources applied. This is most important in developing relevant, important knowledge that supports high performance. It helps to have the right people on it and use Human Knowledge Design methods. Stuff in databases and documents can be useful but too often the quality and accessibility is too poor to translate into knowledge in human heads. Disseminating new ideas is particularly important and can be tracked, modelled, and encouraged if you understand the problem solving people have to do when confronted with a new idea.
Forgetting. A huge factor and ignored unless you try to quantify your knowledge. Can be reduced by knowledge maintenance systems.
Training. Normally rendered almost ineffective by forgetting even when the training itself was good. Training should be just what's required, and be done just in time.
Human Knowledge Management is possible because of the precise undertanding and quantification of knowledge that comes from Human Knowledge Design. There is a vast difference between storing a typical office document in a database and creating a knowledge design and implementation system. The more usual concept of knowledge management as something you do with databases has been a failure in practice and it is time to get back to the true meaning of knowledge. The databases are supposed to help share knowledge. In fact they stand between people and it is much better for people to make contact directly.
A fairly typical way of conducting market research to improve sales effectiveness goes like this:
A sample of customers is analysed.
The information is summarised and conclusions are drawn.
Decisions are made and a strategy is formed for selling the product or products.
The strategy is documented and training materials are developed using generic principles and imaginary case studies are invented.
The strategy is explained to sales people and they get the training.
The sales people try to work out how to approach specific customers.
In cases where the whole point is to get sales people to approach customers in a more effective way this process can be compressed and made far more efficient. The key is to consider what knowledge structures the sales people need.
Sales people need to be able to recognise characteristics of customers that are relevant to deciding the approach, and then recall inferences about the appropriate strategy to use. This requires a set of condition -> action pairs (sometimes called "productions" by psychologists) where the conditions are characteristics of customers. Perceiving these characteristics may require some careful knowledge design, but in practice people pick up this kind of perceptual skill quite quickly given a large number of examples and guidance as to what characteristics they are looking for.
Obviously, there is a close similarity between the information gathered in the initial research and the information needed to train the new skill. We can cut out the information losses and distortions of the conventional method, and save time, by using the detailed information gathered in the research as the training materials.
This involves putting descriptions of customers and their circumstances into a database, then having people (perhaps even the sales force) classify them repeatedly using keywords (so each item could end up in many categories) depending on their characteristics e.g. "big", "international", "family owned", "barely solvent", "newly created", "expanding rapidly", and combinations of these. The implications for sales approach of various characteristics are added to the database against each characteristic and the whole analysis is refined and edited until it is a clean design embodying a sound set of strategies.
The sales people then undergo training using the same database, learning to recognise the relevant customer characteristics by seeing their names (the keywords) and reading multiple examples, and the associated advice, then taking example customers and practicing recognising relevant characteristics and deducing the sales approach.
Why do science? People do it for a lot of reasons but the reason they get funding is the practical value of what might be discovered. "Pure science" is just science with no direct application. Psychology has wasted a lot of effort trying to model human thought as if it was a naturally occurring phenomenon rather than something we humans created. Some aspects of the mind seem to be natural phenomena and progress has been made. Other aspects are flexible and their functioning depends on what we want it to be and the scientific strategy has failed to produce useful discoveries. Instead there has been just a succession of failed theories.
For these hard areas the better strategy is to go directly for applications. In other words, do Human Knowledge Design without waiting for pure science to catch up (which it may never do). The applied research that supports Human Knowledge Design is research to establish what the brain can do well and what it learns well. Alternative strategies for design and implementation of knowledge can just be tried with variations to see what works best and if there are any patterns that make sense.
Knowledge designers need to work from a known base of elementary operations that the mind can perform. As research progresses it is likely that these elementary operations will be broken down into still more elementary operations. There will be some that are more or less hard wired and don't alter much if at all with training after the first few months of life.
Finding out about these is easier if we can control the cognitive strategies people use in experiments so that the noise created by the normal variations is reduced. This is possible if the knowledge has been designed and implemented and cognitive strategies are, therefore, known.
If you're still interested in Human Knowledge Design but have not yet tried my paper from 1986 here's that link again. But be warned, it's slow going!
Human Knowledge Design (February 1986)
Acknowledgements: I would like to thank all those who have read this page and commented. I consider every point carefully and often make improvements as a result.
About the author: Matthew Leitch has been studying the applied psychology of learning and memory since about 1979 and holds a BSc in psychology from University College London. Until recently he worked as a consultant in risk management and systems for a leading professional services firm.
Contact the author at: matthew@learningideas.me.uk
| Home / more articles - The author - Contact |