By Gordon Rugg
Over the last year, we’ve blogged about various aspects of knowledge modelling. That’s allowed us to go into depth about specific topics.
We’re now pulling that information together into a structured format, as an online book. This article contains the core structure of the book, with links to our previous blog articles about the topics within the book. Those articles cover about half of the material that the final version of the book will contain.
We’ve gone for this format, rather than a single downloadable document, because it’s more practical at this point. The knowledge modelling book covers a lot of topics, and even the current partial draft would be a very large document, with a lot of illustrations.
We’ll update this draft fairly frequently, via further blog articles. Some of those articles will be case studies showing how concepts from the book can be applied to real examples. Other articles will be about the broader and deeper context of the book; in particular, the introductory sections and the discussion sections for the main sections. At some point, we’ll put a more reader-friendly version onto the Hyde & Rugg website, which we’re currently updating.
We welcome constructive feedback and suggestions.
The Knowledge Modelling Book
By Gordon Rugg
The concept of studying knowledge has been around for a long time. It’s taken various forms. One form is the traditional philosophical approach of epistemology. Another form is mathematical formalisms of various types. More recently, there’s the field of knowledge representation within artificial Intelligence (AI). There is also knowledge management, and knowledge engineering, within the business world and within computing.
Knowledge modelling overlaps with all of these, but it’s also different from all of them. It’s similar in terms of aiming to be rigorous and systematic. It’s different in terms of its scope, and also in terms of being built around an underlying process for handling knowledge. That process involves four main stages. One is elicitation – getting knowledge out of human beings. The second is representation – choosing and using the most appropriate way of representing that knowledge. The third is testing – checking the knowledge for completeness and correctness. The fourth is education – closing the loop by choosing and using the most appropriate way of teaching and learning the knowledge.
This book unpacks all of those concepts, with worked examples, and with links to relevant theory and to practical implications. It’s currently a work in progress; some sections are more complete than others. Much of it has already appeared as blog articles. Pulling those articles together in the format of a book should help to explain the big picture within which those articles fit.
The big picture is pretty broad, so we’ve deliberately not gone into detail about parts that have already been well covered in detail by other people. For some key topics, such as graph theory, we’ve pointed the reader to the literature. I’ve written about other related topics, such as research skills, in my other books, and I’ve linked to those where relevant.
This book is composed of copyleft material; you’re welcome to use any of it for non-commercial purposes, including lectures, as long as you retain the Hyde & Rugg copyleft attribution on the material you use.
The work described within this book is very much a team effort. I’ve written most of the blog articles, but those articles draw heavily on collaboration with others – for example, the numerous discussions and collaborations with Sue Gerrard and Jo Hyde, who have played a major part in the development of this work. All the members of the Knowledge Modelling Group at Keele have been significant contributors to this work, particularly Ed de Quincey. I’m also indebted to many other colleagues, such as Marian Petre and Shailey Minocha, for their help. In addition, I’m deeply grateful to the students past and present whose questions and observations have given me deeper insights.
I hope that you find this book useful, thought-provoking and enjoyable.
Gordon Rugg, January 2015
What is Knowledge Modelling? Combining methods from psychology, Artificial Intelligence and other fields can provide a unified, systematic way of handling knowledge, with powerful, practical implications.
Part 1: Elicitation
What are the best ways to gather information from humans, especially when they can’t put their knowledge into words? Client requirements, opinion surveys, market research, etc. Beyond interviews, questionnaires and focus groups…
Part 2: Representation
What are the best ways to represent data, information and knowledge? Beyond pie charts, graphs, spreadsheets and mind maps…
Part 3: Error
What regularities are there in the mistakes that people make, and what’s the broader context? Decision-making, systems pressures, cognitive biases and aesthetic biases…
Part 4: Education
There are numerous different types of memory, skill, knowledge, etc. What are the implications for teaching and learning? Beyond Visual, Auditory & Kinaesthetic…
An overview of how the previous four parts fit together, and of the bigger picture that emerges from combining them.
What does this give you, and where could this go next?
Detailed table of contents, and links
Note: The sections below mainly consist of brief descriptions of the core concepts with each section, followed by links to articles on the Hyde & Rugg blog. Some sections are fairly complete; others are less complete. I’ve included brief descriptions of sections that still need to be written, so that the overall structure of the book is visible.
In the following sections, I’ve included brief explanations in italics where the title of an article isn’t self-explanatory. Where there’s a series of articles on the same topic, I’ve usually linked to the first in the series, which should contain links to the others. I’ve sometimes linked to the same article more than once, where it’s relevant to two or more topics. I haven’t given some topics, such as design, a section of their own; instead, I’ve mentioned them where they’re relevant within the main structure of this book.
General Introduction (not written yet)
When it’s completed, this chapter will describe the context within which our approach to knowledge modelling was developed, and it will give an overview of our approach.
Part 1: Elicitation: Getting information out of humans
Introduction to Part 1 (not written yet)
What is elicitation? The underlying problems in market research, consumer surveys, client requirements, etc. The limitations of interviews, questionnaires and focus groups. There are numerous different types of memory, skill and communication. You can’t access most of these via interviews, questionnaires or focus groups. The bigger picture; mapping types of memory, skill and communication onto appropriate methods to access the knowledge involved.
The problem in a nutshell: Why people can’t, don’t or won’t give you answers
A framework for choosing the appropriate techniques: The compass rose model
A set of articles, working through various issues in client requirements. I’ve linked to the first in the series, which contains links to the others.
Finding out what clients want: Design rationale
Handling the reasons for specific requirements, in a simple visual way.
Seeing what people actually do, as opposed to what they think they do.
Direct observation, indirect observation, STROBE, task analysis, and more.
Getting people’s explanations in their own words. Short Term Memory; memory biases and limitations.
Scenarios, think-aloud technique, critical incident technique, projective approaches, and more.
What categories and criteria do people use to structure their views of the world? Card sorts are a simple but powerful way of finding out.
An elegant, powerful way of finding out about the structures of people’s goals, values and categorisations; also good for unpacking subjective and technical terms.
A simple, quick and very practical way of generating, improving and recording ideas.
Questionnaires and Likert-style scales
Most questionnaires and Likert-style scales range between bad and worse than useless. Common mistakes, and some better practices.
Other methods (not finished yet)
Timelines, activity sequences, etc.
More (Not written yet)
Part 2: Representation: How best to show information
Introduction to Part 2 (not written yet)
What is representation? Overview and key concepts.
Sets and categorisation
Crisp sets, fuzzy sets, rough sets and their implications; gender theory and categorisation; more.
Serial processing and parallel processing
Two different methods for processing information. Although this distinction is well known in computing, it’s received hardly any attention elsewhere. It’s an extremely important distinction, with far-reaching implications for a wide range of fields.
When a concept can’t be meaningfully applied; range of convenience
Most concepts can only be meaningfully applied within a limited range of settings. This article examines the implications.
Applying parallel processing to online search
The Search Visualizer software presents documents visually in a way that’s designed to make best use of human skills in parallel processing. This enables users to see things in texts that would be missed by textual outputs and by statistical analysis.
Applying parallel processing to nonverbal signage and wayfinding
When bad is not the opposite of good: Using two scales instead of one
A common feature of questionnaires is the scale with an opposite concept at each end. That isn’t always a good idea. This article shows why.
Artificial Intelligence and Knowledge Representation (not written yet)
The field of knowledge representation is well established, and has a rich, systematic set of formalisms for representing knowledge.
Connectionism and neural networks
Most people are familiar with traditional, step-by-step reasoning. That, however, isn’t the only approach to processing information. The human brain mainly operates using a very different system, which is good for problems that traditional reasoning can’t handle.
An elegant, extremely powerful way of modelling connections between concepts.
It’s often possible to structure knowledge about a topic from more than one viewpoint. Facet theory is a way of systematising these multiple viewpoints.
Overview of deep structure versus surface structure (not written yet)
The same surface appearance can be caused by more than one underlying deep structure. This article looks at the differences between deep structure and surface structure, and at the implications.
A case study in deep structure: Why horror is scary and humour is funny
Humour and horror can be simply and elegantly modelled as two faces of the same coin, involving shifts in perception of deep structure. If something as apparently complex and subjective as humour or horror can be explained so simply, what are the implications for other apparently complex, subjective phenomena?
Schema theory and script theory
These concepts involve mental templates for structuring pieces of knowledge together, such as the schema for the key features of a car. They’re invaluable concepts, with major practical implications.
Schema theory in action: Hagiography and management theory case studies
Management theory textbooks and articles draw heavily on case studies. To what extent are these case studies misleading, because they are being framed within an oversimplified schema that can be traced back to stories of early saints?
Chunking, schemata and prototypes
The human brain lumps things into groups. This article looks at some ways in which this happens.
Argumentation and visualisation
It’s often better to show a chain of reasoning using formal diagrams than via words. This article shows a worked example.
More (not written yet; topics to be decided)
Part 3: Error
Introduction to Part 3 (not written yet)
What is error? Overview and key concepts.
Traditional models of reason and unreason
The traditional dichotomy between reason and emotion.
The Verifier approach revisited
Our approach to tracking down hard-to-find errors in reasoning and research.
The nature of real-world problems; the role of heuristics
For most real-world problems, traditional logic doesn’t help, for various reasons; usually, you have to use rules of thumb (heuristics) instead.
Expertise and expert behaviour (not written yet)
Logic, evidence and evidence-based approaches
Systems (whether organisational or physical or software) often behave in unexpected ways. This can lead to human error, but can also explain why some apparent errors are actually the right decision. Systems theory explains what’s happening, and why.
Game theory, risk and decisions
Game theory is a powerful branch of mathematics and decision theory, which makes sense of many phenomena that would otherwise be puzzling. This has significant implications for error.
Sub-system optimisation versus system optimisation
A common misconception: Improving the parts better doesn’t necessarily improve the whole, and can actually make it worse. (And vice versa.)
The limitations of binary categorisation
A common error is the attempt to force reality into just two categories.
The cases that don’t fit into the pigeonholes
If you’re trying to fit reality into the wrong pigeonholes, then this will lead to error.
Attribution errors (not yet written)
People tend to attribute events to intended actions, rather than to chance. This can profoundly skew people’s reasoning. This article looks at some aspects of how people rationalise decisions.
The concept of infinity is often misunderstood. This article uses the worked example of whether or not a client’s requirements for a product are actually infinite, and if so, what the implications are.
Occam’s razor: The limitations of simple-looking explanations
Selective misuse of unrepresentative evidence; a widespread error.
Visualisation, representation and risk (not yet written)
Showing evidence about risk in different formats can greatly improve people’s understanding of the actual size and severity of risks. I’ve written about this in Blind Spot.
Fear of failure causes a lot of needless stress and heartache. This article looks at the concept of failure from a different angle.
What are the regularities in what humans want, and in what they find attractive? Beyond the Golden Ratio…
The Mathematics of Desire: The uncanny valley
A whole category of horror involves things that are nearly, but not quite, human. What’s going on there, and does the same issue have broader implications?
The Mathematics of Desire: What makes a design interesting?
The Mathematics of Desire: Parsing art
The Mathematics of Desire: Skeuomorphs
Products that look like something else.
Articles about factors that nudge movies away from realism. One article is already written; more will follow at some point.
More (not written yet; topics to be decided)
Part 4: Education
Introduction to Part 4 (not written yet)
What is education? Overview and key concepts.
Mapping types of memory, skill, etc onto teaching and learning methods
There are numerous types of memory and skill, including Short Term Memory, incidental learning and “muscle memory”. What’s the best mapping between them and the various methods available for teaching and learning?
Introduction to core concepts: Data, information and knowledge (not finished yet)
Data, information and knowledge: The knowledge pyramid
Assessment methods and the knowledge pyramid
Formalised academic knowledge versus craft skills
A critical examination of misconceptions about facts and evidence in education.
Education in context: Its purpose and lifespan
How complex should education theories be?
False dichotomies in education theory
Ways of visualising the structure of a taught course
It’s highly likely that no education system will every produce much more than 95% literacy rates. This article examines the reasons and the implications. We pick up the same theme in our articles about nonverbal signage and wayfinding.
Technology and education: Insights from sociotechnical approaches
The form of a technology makes some activities easier, and others more difficult. What are the implications for educational technology and educational practices?
Representation and presentation of information (not yet written)
This article will bring together key concepts from the second part of the book, about visualisation and representation, and describe their implications for education.
Are writing skills transferable?
Specialist writing is very different from nonspecialist writing, for very practical reasons. Different fields have different forms of specialist writing, so the concept of teaching generic transferable writing skills is based on a misconception.
Parallel processing and natural learning
The concept of “natural” crops up repeatedly across educational theories; it’s often an attempt to describe concepts that are well understood in other disciplines. This article describes how one concept of “natural learning” is an attempt to describe parallel processing.
Compiled skills and “natural” learning
This article describes how one concept of “natural learning” is an attempt to describe compiled skills.
Some thoughts about the concepts of “nature” and “natural”.
“Natural” and “artificial” learning
An analysis of what these concepts actually refer to.
The perils of premature pigeonholing
Passive ignorance, active ignorance, and why students don’t learn
Visualising themes in the Plowden Report
Using the Search Visualizer software to show the distribution and extent of various topics within the Plowden Report.
There has been comparatively little formal study of craft skills – the practical, hands-on skills that turn a concept into practice. This article looks at the implications.
Academic craft skills: Why is scientific writing so boring?
Short answer: For good reasons, based on bad things that have happened in the past. This article looks at why most academics distrust exciting writing, and the implications for learning academic writing skills.
Academic craft skills: Doing online searches
Academic craft skills: Doing meta-analysis, and Systematic Literature Reviews
Meta-analyses and Systematic Literature Reviews of multiple papers are becoming increasingly important. Done appropriately and correctly, they can be invaluable; done inappropriately and wrongly, they are worse than useless, and can lead future researchers astray for years.
Academic craft skills: Finding the right references
Academic craft skills: Literature reviews
More (not written yet; topics to be decided)
General discussion (not written yet): This will be a big section, containing in-depth discussions of the theoretical and practical implications of the concepts described in the previous sections.
General conclusion (not written yet): This will be a shorter section, which briefly summarises the topics discussed above, and which considers future directions and implications.
References and links (not written yet): This will be an annotated set of references, further reading, and links.
Appendices (not written yet): This will contain assorted useful materials, templates, etc, that you can use on a copyleft basis (for any non-commercial use, including lectures, provided that you include the “Copyleft Hyde & Rugg” attribution on them).
Pingback: When liking and disliking aren’t opposites | hyde and rugg
Pingback: The apparent attraction of average faces | hyde and rugg
Pingback: Beauty, novelty and threat | hyde and rugg
Pingback: Premature closure and authoritarian worldviews | hyde and rugg
Pingback: Liking, disliking, and averaging: Why average things are attractive but very attractive things are not average | hyde and rugg